0.9 beta 2

Minor refactor for rewrite.

.editorconfig

@@ -1,5 +0,0 @@
-[*]
-end_of_line = lf
-
-[caddytest/integration/caddyfile_adapt/*.txt]
-indent_style = tab

.gitattributes

@@ -1 +1,14 @@
-*.go text eol=lf
+# shell scripts should not use tabs to indent!
+*.bash    text eol=lf core.whitespace whitespace=tab-in-indent,trailing-space,tabwidth=2
+*.sh      text eol=lf core.whitespace whitespace=tab-in-indent,trailing-space,tabwidth=2
+
+# files for systemd (shell-similar)
+*.path    text eol=lf core.whitespace whitespace=tab-in-indent,trailing-space,tabwidth=2
+*.service text eol=lf core.whitespace whitespace=tab-in-indent,trailing-space,tabwidth=2
+*.timer   text eol=lf core.whitespace whitespace=tab-in-indent,trailing-space,tabwidth=2
+
+# go fmt will enforce this, but in case a user has not called "go fmt" allow GIT to catch this:
+*.go      text eol=lf core.whitespace whitespace=indent-with-non-tab,trailing-space,tabwidth=4
+
+*.yml     text eol=lf core.whitespace whitespace=tab-in-indent,trailing-space,tabwidth=2
+.git*     text eol=auto core.whitespace whitespace=trailing-space

.github/CONTRIBUTING.md

@@ -1,187 +0,0 @@
-Contributing to Caddy
-=====================
-
-Welcome! Thank you for choosing to be a part of our community. Caddy wouldn't be great without your involvement!
-
-For starters, we invite you to join [the Caddy forum](https://caddy.community) where you can hang out with other Caddy users and developers.
-
-## Common Tasks
-
-- [Contributing code](#contributing-code)
-- [Writing a Caddy module](#writing-a-caddy-module)
-- [Asking or answering questions for help using Caddy](#getting-help-using-caddy)
-- [Reporting a bug](#reporting-bugs)
-- [Suggesting an enhancement or a new feature](#suggesting-features)
-- [Improving documentation](#improving-documentation)
-
-Other menu items:
-
-- [Values](#values)
-- [Coordinated Disclosure](#coordinated-disclosure)
-- [Thank You](#thank-you)
-
-
-### Contributing code
-
-You can have a huge impact on the project by helping with its code. To contribute code to Caddy, first submit or comment in an issue to discuss your contribution, then open a [pull request](https://github.com/caddyserver/caddy/pulls) (PR). If you're new to our community, that's okay: **we gladly welcome pull requests from anyone, regardless of your native language or coding experience.** You can get familiar with Caddy's code base by using [code search at Sourcegraph](https://sourcegraph.com/github.com/caddyserver/caddy).
-
-We hold contributions to a high standard for quality :bowtie:, so don't be surprised if we ask for revisions—even if it seems small or insignificant. Please don't take it personally. :blue_heart: If your change is on the right track, we can guide you to make it mergable.
-
-Here are some of the expectations we have of contributors:
-
-- **Open an issue to propose your change first.** This way we can avoid confusion, coordinate what everyone is working on, and ensure that any changes are in-line with the project's goals and the best interests of its users. We can also discuss the best possible implementation. If there's already an issue about it, comment on the existing issue to claim it. A lot of valuable time can be saved by discussing a proposal first.
-
-- **Keep pull requests small.** Smaller PRs are more likely to be merged because they are easier to review! We might ask you to break up large PRs into smaller ones. [An example of what we want to avoid.](https://twitter.com/iamdevloper/status/397664295875805184)
-
-- **Keep related commits together in a PR.** We do want pull requests to be small, but you should also keep multiple related commits in the same PR if they rely on each other.
-
-- **Write tests.** Tests are essential! Written properly, they ensure your change works, and that other changes in the future won't break your change. CI checks should pass.
-
-- **Benchmarks should be included for optimizations.** Optimizations sometimes make code harder to read or have changes that are less than obvious. They should be proven with benchmarks or profiling.
-
-- **[Squash](http://gitready.com/advanced/2009/02/10/squashing-commits-with-rebase.html) insignificant commits.** Every commit should be significant. Commits which merely rewrite a comment or fix a typo can be combined into another commit that has more substance. Interactive rebase can do this, or a simpler way is `git reset --soft <diverging-commit>` then `git commit -s`.
-
-- **Own your contributions.** Caddy is a growing project, and it's much better when individual contributors help maintain their change after it is merged.
-
-- **Use comments properly.** We expect good godoc comments for package-level functions, types, and values. Comments are also useful whenever the purpose for a line of code is not obvious.
-
-- **Pull requests may still get closed.** The longer a PR stays open and idle, the more likely it is to be closed. If we haven't reviewed it in a while, it probably means the change is not a priority. Please don't take this personally, we're trying to balance a lot of tasks! If nobody else has commented or reacted to the PR, it likely means your change is useful only to you. The reality is this happens quite a bit. We don't tend to accept PRs that aren't generally helpful. For these reasons or others, the PR may get closed even after a review. We are not obligated to accept all proposed changes, even if the best justification we can give is something vague like, "It doesn't sit right." Sometimes PRs are just the wrong thing or the wrong time. Because it is open source, you can always build your own modified version of Caddy with a change you need, even if we reject it in the official repo.
-
-We often grant [collaborator status](#collaborator-instructions) to contributors who author one or more significant, high-quality PRs that are merged into the code base!
-
-
-#### HOW TO MAKE A PULL REQUEST TO CADDY
-
-Contributing to Go projects on GitHub is fun and easy. After you have proposed your change in an issue, we recommend the following workflow:
-
-1. [Fork this repo](https://github.com/caddyserver/caddy). This makes a copy of the code you can write to.
-
-2. If you don't already have this repo (caddyserver/caddy.git) repo on your computer, clone it down: `git clone https://github.com/caddyserver/caddy.git`
-
-3. Tell git that it can push the caddyserver/caddy.git repo to your fork by adding a remote: `git remote add myfork https://github.com/<your-username>/caddy.git`
-
-4. Make your changes in the caddyserver/caddy.git repo on your computer.
-
-5. Push your changes to your fork: `git push myfork`
-
-6. [Create a pull request](https://github.com/caddyserver/caddy/pull/new/master) to merge your changes into caddyserver/caddy @ master. (Click "compare across forks" and change the head fork.)
-
-This workflow is nice because you don't have to change import paths. You can get fancier by using different branches if you want.
-
-
-### Writing a Caddy module
-
-Caddy can do more with modules! Anyone can write one. Caddy modules are Go libraries that get compiled into Caddy, extending its feature set. They can add directives to the Caddyfile, add new configuration adapters, and even implement new server types (e.g. HTTP, DNS).
-
-[Learn how to write a module here](https://caddyserver.com/docs/extending-caddy). You should also share and discuss your module idea [on the forums](https://caddy.community) to have people test it out. We don't use the Caddy issue tracker for third-party modules.
-
-
-### Getting help using Caddy
-
-If you have a question about using Caddy, [ask on our forum](https://caddy.community)! There will be more people there who can help you than just the Caddy developers who follow our issue tracker. Issues are not the place for usage questions.
-
-Many people on the forums could benefit from your experience and expertise, too. Once you've been helped, consider giving back by answering other people's questions and participating in other discussions.
-
-
-### Reporting bugs
-
-Like every software, Caddy has its flaws. If you find one, [search the issues](https://github.com/caddyserver/caddy/issues) to see if it has already been reported. If not, [open a new issue](https://github.com/caddyserver/caddy/issues/new) and describe the bug, and somebody will look into it! (This repository is only for Caddy and its standard modules.)
-
-**You can help us fix bugs!** Speed up the patch by identifying the bug in the code. This can sometimes be done by adding `fmt.Println()` statements (or similar) in relevant code paths to narrow down where the problem may be. It's a good way to [introduce yourself to the Go language](https://tour.golang.org), too.
-
-We may reply with an issue template. Please follow the template so we have all the needed information. Unredacted&mdash;yes, actual values matter. We need to be able to repeat the bug using your instructions. Please simplify the issue as much as possible. If you don't, we might close your report. The burden is on you to make it easily reproducible and to convince us that it is actually a bug in Caddy. This is easiest to do when you write clear, concise instructions so we can reproduce the behavior (even if it seems obvious). The more detailed and specific you are, the faster we will be able to help you!
-
-We suggest reading [How to Report Bugs Effectively](http://www.chiark.greenend.org.uk/~sgtatham/bugs.html).
-
-Please be kind. :smile: Remember that Caddy comes at no cost to you, and you're getting free support when we fix your issues. If we helped you, please consider helping someone else!
-
-#### Bug reporting expectations
-
-Maintainers---or more generally, developers---need three things to act on bugs:
-
-1. To agree or be convinced that it's a bug (reporter's responsibility).
-	- A bug is unintentional, undesired, or surprising behavior which violates documentation or relevant spec. It might be either a mistake in the documentation or a bug in the code.
-	- This project usually does not work around bugs in other software, systems, and dependencies; instead, we recommend that those bugs are fixed at their source. This sometimes means we close issues or reject PRs that attempt to fix, workaround, or hide bugs in other projects.
-
-2. To be able to understand what is happening (mostly reporter's responsibility).
-	- If the reporter can provide satisfactory instructions such that a developer can reproduce the bug, the developer will likely be able to understand the bug, write a test case, and implement a fix. This is the least amount of work for everyone and path to the fastest resolution.
-	- Otherwise, the burden is on the reporter to test possible solutions. This is less preferable because it loosens the feedback loop, slows down debugging efforts, obscures the true nature of the problem from the developers, and is unlikely to result in new test cases.
-
-3. A solution, or ideas toward a solution (mostly maintainer's responsibility).
-	- Sometimes the best solution is a documentation change.
-	- Usually the developers have the best domain knowledge for inventing a solution, but reporters may have ideas or preferences for how they would like the software to work.
-	- Security, correctness, and project goals/vision all take priority over a user's preferences.
-	- It's simply good business to yield a solution that satisfies the users, and it's even better business to leave them impressed.
-
-Thus, at the very least, the reporter is expected to:
-
-1. Convince the reader that it's a bug in Caddy (if it's not obvious).
-2. Reduce the problem down to the minimum specific steps required to reproduce it.
-
-The maintainer is usually able to do the rest; but of course the reporter may invest additional effort to speed up the process.
-
-
-
-### Suggesting features
-
-First, [search to see if your feature has already been requested](https://github.com/caddyserver/caddy/issues). If it has, you can add a :+1: reaction to vote for it. If your feature idea is new, open an issue to request the feature. Please describe your idea thoroughly so that we know how to implement it! Really vague requests may not be helpful or actionable and, without clarification, will have to be closed.
-
-While we really do value your requests and implement many of them, not all features are a good fit for Caddy. Most of those [make good modules](#writing-a-caddy-module), which can be made by anyone! But if a feature is not in the best interest of the Caddy project or its users in general, we may politely decline to implement it into Caddy core. Additionally, some features are bad ideas altogether (for either obvious or non-obvious reasons) which may be rejected. We'll try to explain why we reject a feature, but sometimes the best we can do is, "It's not a good fit for the project."
-
-
-### Improving documentation
-
-Caddy's documentation is available at [https://caddyserver.com/docs](https://caddyserver.com/docs) and its source is in the [website repo](https://github.com/caddyserver/website). If you would like to make a fix to the docs, please submit an issue there describing the change to make.
-
-Note that third-party module documentation is not hosted by the Caddy website, other than basic usage examples. They are managed by the individual module authors, and you will have to contact them to change their documentation.
-
-Our documentation is scoped to the Caddy project only: it is not for describing how other software or systems work, even if they relate to Caddy or web servers. That kind of content [can be found in our community wiki](https://caddy.community/c/wiki/13), however.
-
-## Collaborator Instructions
-
-Collaborators have push rights to the repository. We grant this permission after one or more successful, high-quality PRs are merged! We thank them for their help. The expectations we have of collaborators are:
-
-- **Help review pull requests.** Be meticulous, but also kind. We love our contributors, but we critique the contribution to make it better. Multiple, thorough reviews make for the best contributions! Here are some questions to consider:
-	- Can the change be made more elegant?
-	- Is this a maintenance burden?
-	- What assumptions does the code make?
-	- Is it well-tested?
-	- Is the change a good fit for the project?
-	- Does it actually fix the problem or is it creating a special case instead?
-	- Does the change incur any new dependencies? (Avoid these!)
-
-- **Answer issues.** If every collaborator helped out with issues, we could count the number of open issues on two hands. This means getting involved in the discussion, investigating the code, and yes, debugging it. It's fun. Really! :smile: Please, please help with open issues. Granted, some issues need to be done before others. And of course some are larger than others: you don't have to do it all yourself. Work with other collaborators as a team!
-
-- **Do not merge pull requests until they have been approved by one or two other collaborators.** If a project owner approves the PR, it can be merged (as long as the conversation has finished too).
-
-- **Prefer squashed commits over a messy merge.** If there are many little commits, please [squash the commits](https://stackoverflow.com/a/11732910/1048862) so we don't clutter the commit history.
-
-- **Don't accept new dependencies lightly.** Dependencies can make the world crash and burn, but they are sometimes necessary. Choose carefully. Extremely small dependencies (a few lines of code) can be inlined. The rest may not be needed. For those that are, Caddy uses [go modules](https://github.com/golang/go/wiki/Modules). All external dependencies must be installed as modules, and _Caddy must not export any types defined by those dependencies_. Check this diligently!
-
-- **Be extra careful in some areas of the code.** There are some critical areas in the Caddy code base that we review extra meticulously: the `caddyhttp` and `caddytls` packages especially.
-
-- **Make sure tests test the actual thing.** Double-check that the tests fail without the change, and pass with it. It's important that they assert what they're purported to assert.
-
-- **Recommended reading**
-	- [CodeReviewComments](https://github.com/golang/go/wiki/CodeReviewComments) for an idea of what we look for in good, clean Go code
-	- [Linus Torvalds describes a good commit message](https://gist.github.com/matthewhudson/1475276)
-	- [Best Practices for Maintainers](https://opensource.guide/best-practices/)
-	- [Shrinking Code Review](https://alexgaynor.net/2015/dec/29/shrinking-code-review/)
-
-
-
-## Values (WIP)
-
-- A person is always more important than code. People don't like being handled "efficiently". But we can still process issues and pull requests efficiently while being kind, patient, and considerate.
-
-- The ends justify the means, if the means are good. A good tree won't produce bad fruit. But if we cut corners or are hasty in our process, the end result will not be good.
-
-
-## Security Policy
-
-If you think you've found a security vulnerability, please refer to our [Security Policy](https://github.com/caddyserver/caddy/security/policy) document.
-
-
-## Thank you
-
-Thanks for your help! Caddy would not be what it is today without your contributions.

.github/FUNDING.yml

@@ -1,12 +0,0 @@
-# These are supported funding model platforms
-
-github: [mholt] # Replace with up to 4 GitHub Sponsors-enabled usernames e.g., [user1, user2]
-patreon: # Replace with a single Patreon username
-open_collective: # Replace with a single Open Collective username
-ko_fi: # Replace with a single Ko-fi username
-tidelift: # Replace with a single Tidelift platform-name/package-name e.g., npm/babel
-community_bridge: # Replace with a single Community Bridge project-name e.g., cloud-foundry
-liberapay: # Replace with a single Liberapay username
-issuehunt: # Replace with a single IssueHunt username
-otechie: # Replace with a single Otechie username
-custom: # Replace with up to 4 custom sponsorship URLs e.g., ['link1', 'link2']

.github/SECURITY.md

@@ -1,59 +0,0 @@
-# Security Policy
-
-The Caddy project would like to make sure that it stays on top of all practically-exploitable vulnerabilities.
-
-
-## Supported Versions
-
-| Version | Supported          |
-| ------- | ------------------ |
-| 2.x     | :white_check_mark: |
-| 1.x     | :x:                |
-| < 1.x   | :x:                |
-
-
-## Acceptable Scope
-
-A security report must demonstrate a security bug in the source code from this repository.
-
-Some security problems are the result of interplay between different components of the Web, rather than a vulnerability in the web server itself. Please only report vulnerabilities in the web server itself, as we cannot coerce the rest of the Web to be fixed (for example, we do not consider IP spoofing, BGP hijacks, or missing/misconfigured HTTP headers a vulnerability in the Caddy web server).
-
-Vulnerabilities caused by misconfigurations are out of scope. Yes, it is entirely possible to craft and use a configuration that is unsafe, just like with every other web server; we recommend against doing that.
-
-We do not accept reports if the steps imply or require a compromised system or third-party software, as we cannot control those. We expect that users secure their own systems and keep all their software patched. For example, if untrusted users are able to upload/write/host arbitrary files in the web root directory, it is NOT a security bug in Caddy if those files get served to clients; however, it _would_ be a valid report if a bug in Caddy's source code unintentionally gave unauthorized users the ability to upload unsafe files or delete files without relying on an unpatched system or piece of software.
-
-Client-side exploits are out of scope. In other words, it is not a bug in Caddy if the web browser does something unsafe, even if the downloaded content was served by Caddy. (Those kinds of exploits can generally be mitigated by proper configuration of HTTP headers.) As a general rule, the content served by Caddy is not considered in scope because content is configurable by the site owner or the associated web application.
-
-Security bugs in code dependencies are out of scope. Instead, if a dependency has patched a relevant security bug, please feel free to open a public issue or pull request to update that dependency in our code.
-
-
-## Reporting a Vulnerability
-
-We get a lot of difficult reports that turn out to be invalid. Clear, obvious reports tend to be the most credible (but are also rare).
-
-First please ensure your report falls within the accepted scope of security bugs (above).
-
-We'll need enough information to verify the bug and make a patch. To speed things up, please include:
-
-- Most minimal possible config (without redactions!)
-- Command(s)
-- Precise HTTP requests (`curl -v` and its output please)
-- Full log output (please enable debug mode)
-- Specific minimal steps to reproduce the issue from scratch
-- A working patch
-
-Please DO NOT use containers, VMs, cloud instances or services, or any other complex infrastructure in your steps. Always prefer `curl` instead of web browsers.
-
-We consider publicly-registered domain names to be public information. This necessary in order to maintain the integrity of certificate transparency, public DNS, and other public trust systems. Do not redact domain names from your reports. The actual content of your domain name affects Caddy's behavior, so we need the exact domain name(s) to reproduce with, or your report will be ignored.
-
-It will speed things up if you suggest a working patch, such as a code diff, and explain why and how it works. Reports that are not actionable, do not contain enough information, are too pushy/demanding, or are not able to convince us that it is a viable and practical attack on the web server itself may be deferred to a later time or possibly ignored, depending on available resources. Priority will be given to credible, responsible reports that are constructive, specific, and actionable. (We get a lot of invalid reports.) Thank you for understanding.
-
-When you are ready, please email Matt Holt (the author) directly: matt at dyanim dot com.
-
-Please don't encrypt the email body. It only makes the process more complicated.
-
-Please also understand that due to our nature as an open source project, we do not have a budget to award security bounties. We can only thank you.
-
-If your report is valid and a patch is released, we will not reveal your identity by default. If you wish to be credited, please give us the name to use and/or your GitHub username. If you don't provide this we can't credit you.
-
-Thanks for responsibly helping Caddy&mdash;and thousands of websites&mdash;be more secure!

.github/workflows/ci.yml

@@ -1,182 +0,0 @@
-# Used as inspiration: https://github.com/mvdan/github-actions-golang
-
-name: Tests
-
-on:
-  push:
-    branches:
-      - master
-      - 2.*
-  pull_request:
-    branches:
-      - master
-      - 2.*
-
-jobs:
-  test:
-    strategy:
-      # Default is true, cancels jobs for other platforms in the matrix if one fails
-      fail-fast: false
-      matrix:
-        os: [ ubuntu-latest, macos-latest, windows-latest ]
-        go: [ '1.18', '1.19' ]
-
-        include:
-        # Set the minimum Go patch version for the given Go minor
-        # Usable via ${{ matrix.GO_SEMVER }}
-        - go: '1.18'
-          GO_SEMVER: '~1.18.4'
-
-        - go: '1.19'
-          GO_SEMVER: '~1.19.0'
-
-        # Set some variables per OS, usable via ${{ matrix.VAR }}
-        # CADDY_BIN_PATH: the path to the compiled Caddy binary, for artifact publishing
-        # SUCCESS: the typical value for $? per OS (Windows/pwsh returns 'True')
-        - os: ubuntu-latest
-          CADDY_BIN_PATH: ./cmd/caddy/caddy
-          SUCCESS: 0
-
-        - os: macos-latest
-          CADDY_BIN_PATH: ./cmd/caddy/caddy
-          SUCCESS: 0
-
-        - os: windows-latest
-          CADDY_BIN_PATH: ./cmd/caddy/caddy.exe
-          SUCCESS: 'True'
-
-    runs-on: ${{ matrix.os }}
-
-    steps:
-    - name: Install Go
-      uses: actions/setup-go@v3
-      with:
-        go-version: ${{ matrix.GO_SEMVER }}
-        check-latest: true
-
-    - name: Checkout code
-      uses: actions/checkout@v3
-
-    # These tools would be useful if we later decide to reinvestigate
-    # publishing test/coverage reports to some tool for easier consumption
-    # - name: Install test and coverage analysis tools
-    #   run: |
-    #     go get github.com/axw/gocov/gocov
-    #     go get github.com/AlekSi/gocov-xml
-    #     go get -u github.com/jstemmer/go-junit-report
-    #     echo "::add-path::$(go env GOPATH)/bin"
-
-    - name: Print Go version and environment
-      id: vars
-      run: |
-        printf "Using go at: $(which go)\n"
-        printf "Go version: $(go version)\n"
-        printf "\n\nGo environment:\n\n"
-        go env
-        printf "\n\nSystem environment:\n\n"
-        env
-        printf "Git version: $(git version)\n\n"
-        # Calculate the short SHA1 hash of the git commit
-        echo "::set-output name=short_sha::$(git rev-parse --short HEAD)"
-
-    - name: Cache the build cache
-      uses: actions/cache@v2
-      with:
-        # In order:
-        # * Module download cache
-        # * Build cache (Linux)
-        # * Build cache (Mac)
-        # * Build cache (Windows)
-        path: |
-          ~/go/pkg/mod
-          ~/.cache/go-build
-          ~/Library/Caches/go-build
-          ~\AppData\Local\go-build
-        key: ${{ runner.os }}-${{ matrix.go }}-go-ci-${{ hashFiles('**/go.sum') }}
-        restore-keys: |
-          ${{ runner.os }}-${{ matrix.go }}-go-ci
-
-    - name: Get dependencies
-      run: |
-        go get -v -t -d ./...
-        # mkdir test-results
-
-    - name: Build Caddy
-      working-directory: ./cmd/caddy
-      env:
-        CGO_ENABLED: 0
-      run: |
-        go build -trimpath -ldflags="-w -s" -v
-
-    - name: Publish Build Artifact
-      uses: actions/upload-artifact@v1
-      with:
-        name: caddy_${{ runner.os }}_go${{ matrix.go }}_${{ steps.vars.outputs.short_sha }}
-        path: ${{ matrix.CADDY_BIN_PATH }}
-
-    # Commented bits below were useful to allow the job to continue
-    # even if the tests fail, so we can publish the report separately
-    # For info about set-output, see https://stackoverflow.com/questions/57850553/github-actions-check-steps-status
-    - name: Run tests
-      # id: step_test
-      # continue-on-error: true
-      run: |
-        # (go test -v -coverprofile=cover-profile.out -race ./... 2>&1) > test-results/test-result.out
-        go test -v -coverprofile="cover-profile.out" -short -race ./...
-        # echo "::set-output name=status::$?"
-
-    # Relevant step if we reinvestigate publishing test/coverage reports
-    # - name: Prepare coverage reports
-    #   run: |
-    #     mkdir coverage
-    #     gocov convert cover-profile.out > coverage/coverage.json
-    #     # Because Windows doesn't work with input redirection like *nix, but output redirection works.
-    #     (cat ./coverage/coverage.json | gocov-xml) > coverage/coverage.xml
-
-    # To return the correct result even though we set 'continue-on-error: true'
-    # - name: Coerce correct build result
-    #   if: matrix.os != 'windows-latest' && steps.step_test.outputs.status != ${{ matrix.SUCCESS }}
-    #   run: |
-    #     echo "step_test ${{ steps.step_test.outputs.status }}\n"
-    #     exit 1
-
-  s390x-test:
-    name: test (s390x on IBM Z)
-    runs-on: ubuntu-latest
-    if: github.event.pull_request.head.repo.full_name == github.repository
-    continue-on-error: true  # August 2020: s390x VM is down due to weather and power issues
-    steps:
-      - name: Checkout code into the Go module directory
-        uses: actions/checkout@v3
-      - name: Run Tests
-        run: |
-          mkdir -p ~/.ssh && echo -e "${SSH_KEY//_/\\n}" > ~/.ssh/id_ecdsa && chmod og-rwx ~/.ssh/id_ecdsa
-
-          # short sha is enough?
-          short_sha=$(git rev-parse --short HEAD)
-
-          # The environment is fresh, so there's no point in keeping accepting and adding the key.
-          rsync -arz -e "ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null" --progress --delete --exclude '.git' . caddy-ci@ci-s390x.caddyserver.com:/var/tmp/"$short_sha"
-          ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -t caddy-ci@ci-s390x.caddyserver.com "cd /var/tmp/$short_sha; go version; go env; printf "\n\n";CGO_ENABLED=0 go test -v ./..."
-          test_result=$?
-
-          # There's no need leaving the files around
-          ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null caddy-ci@ci-s390x.caddyserver.com "rm -rf /var/tmp/'$short_sha'"
-
-          echo "Test exit code: $test_result"
-          exit $test_result
-        env:
-          SSH_KEY: ${{ secrets.S390X_SSH_KEY }}
-
-  goreleaser-check:
-    runs-on: ubuntu-latest
-    steps:
-      - name: checkout
-        uses: actions/checkout@v3
-      
-      - uses: goreleaser/goreleaser-action@v2
-        with:
-          version: latest
-          args: check
-        env:
-          TAG: ${{ steps.vars.outputs.version_tag }}

.github/workflows/cross-build.yml

@@ -1,74 +0,0 @@
-name: Cross-Build
-
-on:
-  push:
-    branches:
-      - master
-      - 2.*
-  pull_request:
-    branches:
-      - master
-      - 2.*
-
-jobs:
-  cross-build-test:
-    strategy:
-      fail-fast: false
-      matrix:
-        goos: ['android', 'linux', 'solaris', 'illumos', 'dragonfly', 'freebsd', 'openbsd', 'plan9', 'windows', 'darwin', 'netbsd']
-        go: [ '1.19' ]
-
-        include:
-        # Set the minimum Go patch version for the given Go minor
-        # Usable via ${{ matrix.GO_SEMVER }}
-        - go: '1.19'
-          GO_SEMVER: '~1.19.0'
-
-    runs-on: ubuntu-latest
-    continue-on-error: true
-    steps:
-      - name: Install Go
-        uses: actions/setup-go@v3
-        with:
-          go-version: ${{ matrix.GO_SEMVER }}
-          check-latest: true
-
-      - name: Print Go version and environment
-        id: vars
-        run: |
-          printf "Using go at: $(which go)\n"
-          printf "Go version: $(go version)\n"
-          printf "\n\nGo environment:\n\n"
-          go env
-          printf "\n\nSystem environment:\n\n"
-          env
-
-      - name: Cache the build cache
-        uses: actions/cache@v2
-        with:
-          # In order:
-          # * Module download cache
-          # * Build cache (Linux)
-          path: |
-            ~/go/pkg/mod
-            ~/.cache/go-build
-          key: cross-build-go${{ matrix.go }}-${{ matrix.goos }}-${{ hashFiles('**/go.sum') }}
-          restore-keys: |
-            cross-build-go${{ matrix.go }}-${{ matrix.goos }}
-
-      - name: Checkout code into the Go module directory
-        uses: actions/checkout@v3
-
-      - name: Run Build
-        env:
-          CGO_ENABLED: 0
-          GOOS: ${{ matrix.goos }}
-        shell: bash
-        continue-on-error: true
-        working-directory: ./cmd/caddy
-        run: |
-          GOOS=$GOOS go build -trimpath -o caddy-"$GOOS"-amd64 2> /dev/null
-          if [ $? -ne 0 ]; then
-            echo "::warning ::$GOOS Build Failed"
-            exit 0
-          fi

.github/workflows/lint.yml

@@ -1,35 +0,0 @@
-name: Lint
-
-on:
-  push:
-    branches:
-      - master
-      - 2.*
-  pull_request:
-    branches:
-      - master
-      - 2.*
-
-jobs:
-  # From https://github.com/golangci/golangci-lint-action
-  golangci:
-    name: lint
-    strategy:
-      matrix:
-        os: [ubuntu-latest, macos-latest, windows-latest]
-    runs-on: ${{ matrix.os }}
-    steps:
-      - uses: actions/checkout@v3
-      - uses: actions/setup-go@v3
-        with:
-          go-version: '~1.18.4'
-          check-latest: true
-
-      - name: golangci-lint
-        uses: golangci/golangci-lint-action@v3
-        with:
-          version: v1.47
-          # Windows times out frequently after about 5m50s if we don't set a longer timeout.
-          args: --timeout 10m
-          # Optional: show only new issues if it's a pull request. The default value is `false`.
-          # only-new-issues: true

.github/workflows/release.yml

@@ -1,184 +0,0 @@
-name: Release
-
-on:
-  push:
-    tags:
-      - 'v*.*.*'
-
-jobs:
-  release:
-    name: Release
-    strategy:
-      matrix:
-        os: [ ubuntu-latest ]
-        go: [ '1.19' ]
-
-        include:
-        # Set the minimum Go patch version for the given Go minor
-        # Usable via ${{ matrix.GO_SEMVER }}
-        - go: '1.19'
-          GO_SEMVER: '~1.19.0'
-
-    runs-on: ${{ matrix.os }}
-    # https://github.com/sigstore/cosign/issues/1258#issuecomment-1002251233
-    # https://docs.github.com/en/actions/deployment/security-hardening-your-deployments/about-security-hardening-with-openid-connect#adding-permissions-settings
-    permissions:
-      id-token: write
-      # https://docs.github.com/en/rest/overview/permissions-required-for-github-apps#permission-on-contents
-      # "Releases" is part of `contents`, so it needs the `write`
-      contents: write
-
-    steps:
-    - name: Install Go
-      uses: actions/setup-go@v3
-      with:
-        go-version: ${{ matrix.GO_SEMVER }}
-        check-latest: true
-
-    - name: Checkout code
-      uses: actions/checkout@v3
-      with:
-        fetch-depth: 0
-
-    # Force fetch upstream tags -- because 65 minutes
-    # tl;dr: actions/checkout@v3 runs this line:
-    #   git -c protocol.version=2 fetch --no-tags --prune --progress --no-recurse-submodules --depth=1 origin +ebc278ec98bb24f2852b61fde2a9bf2e3d83818b:refs/tags/
-    # which makes its own local lightweight tag, losing all the annotations in the process. Our earlier script ran:
-    #   git fetch --prune --unshallow
-    # which doesn't overwrite that tag because that would be destructive.
-    # Credit to @francislavoie for the investigation.
-    # https://github.com/actions/checkout/issues/290#issuecomment-680260080
-    - name: Force fetch upstream tags
-      run: git fetch --tags --force
-
-    # https://github.community/t5/GitHub-Actions/How-to-get-just-the-tag-name/m-p/32167/highlight/true#M1027
-    - name: Print Go version and environment
-      id: vars
-      run: |
-        printf "Using go at: $(which go)\n"
-        printf "Go version: $(go version)\n"
-        printf "\n\nGo environment:\n\n"
-        go env
-        printf "\n\nSystem environment:\n\n"
-        env
-        echo "::set-output name=version_tag::${GITHUB_REF/refs\/tags\//}"
-        echo "::set-output name=short_sha::$(git rev-parse --short HEAD)"
-
-        # Add "pip install" CLI tools to PATH
-        echo ~/.local/bin >> $GITHUB_PATH
-
-        # Parse semver
-        TAG=${GITHUB_REF/refs\/tags\//}
-        SEMVER_RE='[^0-9]*\([0-9]*\)[.]\([0-9]*\)[.]\([0-9]*\)\([0-9A-Za-z\.-]*\)'
-        TAG_MAJOR=`echo ${TAG#v} | sed -e "s#$SEMVER_RE#\1#"`
-        TAG_MINOR=`echo ${TAG#v} | sed -e "s#$SEMVER_RE#\2#"`
-        TAG_PATCH=`echo ${TAG#v} | sed -e "s#$SEMVER_RE#\3#"`
-        TAG_SPECIAL=`echo ${TAG#v} | sed -e "s#$SEMVER_RE#\4#"`
-        echo "::set-output name=tag_major::${TAG_MAJOR}"
-        echo "::set-output name=tag_minor::${TAG_MINOR}"
-        echo "::set-output name=tag_patch::${TAG_PATCH}"
-        echo "::set-output name=tag_special::${TAG_SPECIAL}"
-
-    # Cloudsmith CLI tooling for pushing releases
-    # See https://help.cloudsmith.io/docs/cli
-    - name: Install Cloudsmith CLI
-      run: pip install --upgrade cloudsmith-cli
-
-    - name: Validate commits and tag signatures
-      run: |
-        
-        # Import Matt Holt's key
-        curl 'https://github.com/mholt.gpg' | gpg --import
-
-        echo "Verifying the tag: ${{ steps.vars.outputs.version_tag }}"
-        # tags are only accepted if signed by Matt's key
-        git verify-tag "${{ steps.vars.outputs.version_tag }}" || exit 1
-
-    - name: Cache the build cache
-      uses: actions/cache@v2
-      with:
-        # In order:
-        # * Module download cache
-        # * Build cache (Linux)
-        path: |
-          ~/go/pkg/mod
-          ~/.cache/go-build
-        key: ${{ runner.os }}-go${{ matrix.go }}-release-${{ hashFiles('**/go.sum') }}
-        restore-keys: |
-          ${{ runner.os }}-go${{ matrix.go }}-release
-    - name: Install Cosign
-      uses: sigstore/cosign-installer@main
-    - name: Cosign version
-      run: cosign version
-    - name: Install Syft
-      uses: anchore/sbom-action/download-syft@main
-    - name: Syft version
-      run: syft version
-    # GoReleaser will take care of publishing those artifacts into the release
-    - name: Run GoReleaser
-      uses: goreleaser/goreleaser-action@v2
-      with:
-        version: latest
-        args: release --rm-dist
-      env:
-        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        TAG: ${{ steps.vars.outputs.version_tag }}
-        COSIGN_EXPERIMENTAL: 1
-
-    # Only publish on non-special tags (e.g. non-beta)
-    # We will continue to push to Gemfury for the foreseeable future, although
-    # Cloudsmith is probably better, to not break things for existing users of Gemfury.
-    # See https://gemfury.com/caddy/deb:caddy
-    - name: Publish .deb to Gemfury
-      if: ${{ steps.vars.outputs.tag_special == '' }}
-      env:
-        GEMFURY_PUSH_TOKEN: ${{ secrets.GEMFURY_PUSH_TOKEN }}
-      run: |
-        for filename in dist/*.deb; do
-          # armv6 and armv7 are both "armhf" so we can skip the duplicate
-          if [[ "$filename" == *"armv6"* ]]; then
-            echo "Skipping $filename"
-            continue
-          fi
-
-          curl -F package=@"$filename" https://${GEMFURY_PUSH_TOKEN}:@push.fury.io/caddy/
-        done
-
-    # Publish only special tags (unstable/beta/rc) to the "testing" repo
-    # See https://cloudsmith.io/~caddy/repos/testing/
-    - name: Publish .deb to Cloudsmith (special tags)
-      if: ${{ steps.vars.outputs.tag_special != '' }}
-      env:
-        CLOUDSMITH_API_KEY: ${{ secrets.CLOUDSMITH_API_KEY }}
-      run: |
-        for filename in dist/*.deb; do
-          # armv6 and armv7 are both "armhf" so we can skip the duplicate
-          if [[ "$filename" == *"armv6"* ]]; then
-            echo "Skipping $filename"
-            continue
-          fi
-
-          echo "Pushing $filename to 'testing'"
-          cloudsmith push deb caddy/testing/any-distro/any-version $filename
-        done
-
-    # Publish stable tags to Cloudsmith to both repos, "stable" and "testing"
-    # See https://cloudsmith.io/~caddy/repos/stable/
-    - name: Publish .deb to Cloudsmith (stable tags)
-      if: ${{ steps.vars.outputs.tag_special == '' }}
-      env:
-        CLOUDSMITH_API_KEY: ${{ secrets.CLOUDSMITH_API_KEY }}
-      run: |
-        for filename in dist/*.deb; do
-          # armv6 and armv7 are both "armhf" so we can skip the duplicate
-          if [[ "$filename" == *"armv6"* ]]; then
-            echo "Skipping $filename"
-            continue
-          fi
-
-          echo "Pushing $filename to 'stable'"
-          cloudsmith push deb caddy/stable/any-distro/any-version $filename
-
-          echo "Pushing $filename to 'testing'"
-          cloudsmith push deb caddy/testing/any-distro/any-version $filename
-        done

.github/workflows/release_published.yml

@@ -1,34 +0,0 @@
-name: Release Published
-
-# Event payload: https://developer.github.com/webhooks/event-payloads/#release
-on:
-  release:
-    types: [published]
-
-jobs:
-  release:
-    name: Release Published
-    strategy:
-      matrix:
-        os: [ ubuntu-latest ]
-    runs-on: ${{ matrix.os }}
-
-    steps:
-
-    # See https://github.com/peter-evans/repository-dispatch
-    - name: Trigger event on caddyserver/dist
-      uses: peter-evans/repository-dispatch@v1
-      with:
-        token: ${{ secrets.REPO_DISPATCH_TOKEN }}
-        repository: caddyserver/dist
-        event-type: release-tagged
-        client-payload: '{"tag": "${{ github.event.release.tag_name }}"}'
-
-    - name: Trigger event on caddyserver/caddy-docker
-      uses: peter-evans/repository-dispatch@v1
-      with:
-        token: ${{ secrets.REPO_DISPATCH_TOKEN }}
-        repository: caddyserver/caddy-docker
-        event-type: release-tagged
-        client-payload: '{"tag": "${{ github.event.release.tag_name }}"}'
-

.gitignore

@@ -1,28 +1,16 @@
-_gitignore/
-*.log
-Caddyfile
-Caddyfile.*
-!caddyfile/
-
-# artifacts from pprof tooling
-*.prof
-*.test
-
-# build artifacts and helpers
-cmd/caddy/caddy
-cmd/caddy/caddy.exe
-
-# mac specific
 .DS_Store
+Thumbs.db
+_gitignore/
+Vagrantfile
+.vagrant/
 
-# go modules
-vendor
+dist/builds/
+dist/release/
 
-# goreleaser artifacts
-dist
-caddy-build
-caddy-dist
+error.log
+access.log
 
-# IDE files
-.idea/
-.vscode/
+/*.conf
+Caddyfile
+
+og_static/

.golangci.yml

@@ -1,98 +0,0 @@
-linters-settings:
-  errcheck:
-    ignore: fmt:.*,go.uber.org/zap/zapcore:^Add.*
-    ignoretests: true
-
-linters:
-  disable-all: true
-  enable:
-    - bodyclose
-    - deadcode
-    - errcheck
-    - gofmt
-    - goimports
-    - gosec
-    - gosimple
-    - govet
-    - ineffassign
-    - misspell
-    - prealloc
-    - staticcheck
-    - structcheck
-    - typecheck
-    - unconvert
-    - unused
-    - varcheck
-  # these are implicitly disabled:
-  # - asciicheck
-  # - depguard
-  # - dogsled
-  # - dupl
-  # - exhaustive
-  # - exportloopref
-  # - funlen
-  # - gci
-  # - gochecknoglobals
-  # - gochecknoinits
-  # - gocognit
-  # - goconst
-  # - gocritic
-  # - gocyclo
-  # - godot
-  # - godox
-  # - goerr113
-  # - gofumpt
-  # - goheader
-  # - golint
-  # - gomnd
-  # - gomodguard
-  # - goprintffuncname
-  # - interfacer
-  # - lll
-  # - maligned
-  # - nakedret
-  # - nestif
-  # - nlreturn
-  # - noctx
-  # - nolintlint
-  # - rowserrcheck
-  # - scopelint
-  # - sqlclosecheck
-  # - stylecheck
-  # - testpackage
-  # - unparam
-  # - whitespace
-  # - wsl
-
-run:
-  # default concurrency is a available CPU number.
-  # concurrency: 4 # explicitly omit this value to fully utilize available resources.
-  deadline: 5m
-  issues-exit-code: 1
-  tests: false
-
-# output configuration options
-output:
-  format: 'colored-line-number'
-  print-issued-lines: true
-  print-linter-name: true
-
-issues:
-  exclude-rules:
-    # we aren't calling unknown URL
-    - text: "G107" # G107: Url provided to HTTP request as taint input
-      linters:
-        - gosec
-    # as a web server that's expected to handle any template, this is totally in the hands of the user.
-    - text: "G203" # G203: Use of unescaped data in HTML templates
-      linters:
-        - gosec
-    # we're shelling out to known commands, not relying on user-defined input.
-    - text: "G204" # G204: Audit use of command execution
-      linters:
-        - gosec
-    # the choice of weakrand is deliberate, hence the named import "weakrand"
-    - path: modules/caddyhttp/reverseproxy/selectionpolicies.go
-      text: "G404" # G404: Insecure random number source (rand)
-      linters:
-        - gosec

.goreleaser.yml

@@ -1,148 +0,0 @@
-before:
-  hooks:
-    # The build is done in this particular way to build Caddy in a designated directory named in .gitignore.
-    # This is so we can run goreleaser on tag without Git complaining of being dirty. The main.go in cmd/caddy directory 
-    # cannot be built within that directory due to changes necessary for the build causing Git to be dirty, which
-    # subsequently causes gorleaser to refuse running.
-    - mkdir -p caddy-build
-    - cp cmd/caddy/main.go caddy-build/main.go
-    - /bin/sh -c 'cd ./caddy-build && go mod init caddy'
-    # GoReleaser doesn't seem to offer {{.Tag}} at this stage, so we have to embed it into the env
-    # so we run: TAG=$(git describe --abbrev=0) goreleaser release --rm-dist --skip-publish --skip-validate
-    - go mod edit -require=github.com/caddyserver/caddy/v2@{{.Env.TAG}} ./caddy-build/go.mod
-    # as of Go 1.16, `go` commands no longer automatically change go.{mod,sum}. We now have to explicitly
-    # run `go mod tidy`. The `/bin/sh -c '...'` is because goreleaser can't find cd in PATH without shell invocation.
-    - /bin/sh -c 'cd ./caddy-build && go mod tidy'
-    - git clone --depth 1 https://github.com/caddyserver/dist caddy-dist
-    - mkdir -p caddy-dist/man
-    - go mod download
-    - go run cmd/caddy/main.go manpage --directory ./caddy-dist/man
-    - gzip -r ./caddy-dist/man/
-    - /bin/sh -c 'go run cmd/caddy/main.go completion bash > ./caddy-dist/scripts/bash-completion'
-
-builds:
-- env:
-  - CGO_ENABLED=0
-  - GO111MODULE=on
-  main: main.go
-  dir: ./caddy-build
-  binary: caddy
-  goos:
-  - darwin
-  - linux
-  - windows
-  - freebsd
-  goarch:
-  - amd64
-  - arm
-  - arm64
-  - s390x
-  - ppc64le
-  goarm:
-  - "5"
-  - "6"
-  - "7"
-  ignore:
-    - goos: darwin
-      goarch: arm
-    - goos: darwin
-      goarch: ppc64le
-    - goos: darwin
-      goarch: s390x
-    - goos: windows
-      goarch: ppc64le
-    - goos: windows
-      goarch: s390x
-    - goos: freebsd
-      goarch: ppc64le
-    - goos: freebsd
-      goarch: s390x
-    - goos: freebsd
-      goarch: arm
-      goarm: "5"
-  flags:
-  - -trimpath
-  - -mod=readonly
-  ldflags:
-  - -s -w
-signs:
-  - cmd: cosign
-    signature: "${artifact}.sig"
-    certificate: '{{ trimsuffix .Env.artifact ".tar.gz" }}.pem'
-    args: ["sign-blob", "--output-signature=${signature}", "--output-certificate", "${certificate}", "${artifact}"]
-    artifacts: all
-sboms:
-  - artifacts: binary
-    # defaults to
-    # documents:
-    # - "{{ .Binary }}_{{ .Version }}_{{ .Os }}_{{ .Arch }}.sbom"
-    cmd: syft
-    args: ["$artifact", "--file", "${document}", "--output", "cyclonedx-json"]
-archives:
-  - format_overrides:
-      - goos: windows
-        format: zip
-    replacements:
-      darwin: mac
-checksum:
-  algorithm: sha512
-
-nfpms:
-  - id: default
-    package_name: caddy
-
-    vendor: Dyanim
-    homepage: https://caddyserver.com
-    maintainer: Matthew Holt <mholt@users.noreply.github.com>
-    description: |
-      Caddy - Powerful, enterprise-ready, open source web server with automatic HTTPS written in Go
-    license: Apache 2.0
-
-    formats:
-      - deb
-      # - rpm
-
-    bindir: /usr/bin
-    contents:
-      - src: ./caddy-dist/init/caddy.service
-        dst: /lib/systemd/system/caddy.service
-        
-      - src: ./caddy-dist/init/caddy-api.service
-        dst: /lib/systemd/system/caddy-api.service
-      
-      - src: ./caddy-dist/welcome/index.html
-        dst: /usr/share/caddy/index.html
-      
-      - src: ./caddy-dist/scripts/bash-completion
-        dst: /etc/bash_completion.d/caddy
-    
-      - src: ./caddy-dist/config/Caddyfile
-        dst: /etc/caddy/Caddyfile
-        type: config
-
-      - src: ./caddy-dist/man/*
-        dst: /usr/share/man/man8/
-
-    scripts:
-      postinstall: ./caddy-dist/scripts/postinstall.sh
-      preremove: ./caddy-dist/scripts/preremove.sh
-      postremove: ./caddy-dist/scripts/postremove.sh
-
-
-release:
-  github:
-    owner: caddyserver
-    name: caddy
-  draft: true
-  prerelease: auto
-
-changelog:
-  sort: asc
-  filters:
-    exclude:
-    - '^chore:'
-    - '^ci:'
-    - '^docs?:'
-    - '^readme:'
-    - '^tests?:'
-    - '^\w+\s+' # a hack to remove commit messages without colons thus don't correspond to a package

.travis.yml

@@ -0,0 +1,30 @@
+language: go
+
+go:
+  - 1.6.2
+  - tip
+
+env:
+  - CGO_ENABLED=0
+
+before_install:
+  # Decrypts a script that installs an authenticated cookie
+  # for git to use when cloning from googlesource.com.
+  # Bypasses "bandwidth limit exceeded" errors.
+  # See github.com/golang/go/issues/12933
+ - if [ "$TRAVIS_PULL_REQUEST" = "false" ]; then openssl aes-256-cbc -K $encrypted_3df18f9af81d_key -iv $encrypted_3df18f9af81d_iv -in dist/gitcookie.sh.enc -out dist/gitcookie.sh -d; fi
+
+install:
+  - if [ "$TRAVIS_PULL_REQUEST" = "false" ]; then bash dist/gitcookie.sh; fi
+  - go get -t ./...
+  - go get github.com/golang/lint/golint
+  - go get github.com/gordonklaus/ineffassign
+
+script:
+  - diff <(echo -n) <(gofmt -s -d .)
+  - ineffassign .
+  - go vet ./...
+  - go test ./...
+
+after_script:
+  - golint ./...

AUTHORS

@@ -1,10 +0,0 @@
-# This is the official list of Caddy Authors for copyright purposes.
-# Authors may be either individual people or legal entities.
-#
-# Not all individual contributors are authors. For the full list of
-# contributors, refer to the project's page on GitHub or the repo's
-# commit history.
-
-Matthew Holt <Matthew.Holt@gmail.com>
-Light Code Labs <sales@lightcodelabs.com>
-Ardan Labs <info@ardanlabs.com>

CONTRIBUTING.md

@@ -0,0 +1,105 @@
+## Contributing to Caddy
+
+Welcome! Our community focuses on helping others and making Caddy the best it
+can be. We gladly accept contributions and encourage you to get involved!
+
+
+### Join us in the forum
+
+The [Caddy forum](https://forum.caddyserver.com) is the place for all discussion
+that doesn't belong in issues or pull requests. Feel free to participate with us!
+
+If you want to file a bug report or make an improvement to Caddy, however, you
+should submit an issue or pull request.
+
+
+### Bug reports
+
+Please [search this repository](https://github.com/mholt/caddy/search?q=&type=Issues&utf8=%E2%9C%93)
+with a variety of keywords to ensure your bug is not already reported.
+
+If unique, [open an issue](https://github.com/mholt/caddy/issues) and answer the
+questions so we can understand and reproduce the problematic behavior.
+
+The burden is on you to convince us that it is actually a bug in Caddy. This is
+easiest to do when you write clear, concise instructions so we can reproduce
+the behavior (even if it seems obvious). The more detailed and specific you are,
+the faster we will be able to help you. Check out
+[How to Report Bugs Effectively](http://www.chiark.greenend.org.uk/~sgtatham/bugs.html).
+
+Please be kind. :smile: Remember that Caddy comes at no cost to you, and you're
+getting free help. If we helped you, please consider
+[donating](https://caddyserver.com/donate) - it keeps us motivated!
+
+
+### Minor improvements and new tests
+
+Submit [pull requests](https://github.com/mholt/caddy/pulls) at any time for
+minor changes or new tests. Make sure to write tests to assert your change is
+working properly and is thoroughly covered. We'll ask most pull requests to be
+[squashed](http://gitready.com/advanced/2009/02/10/squashing-commits-with-rebase.html),
+especially with small commits.
+
+Your pull request may be thoroughly reviewed. This is because if we accept the
+PR, we also assume responsibility for it, although we would prefer you to
+help maintain your code after it gets merged. 
+
+
+### Proposals, suggestions, ideas, new features
+
+First, please [search](https://github.com/mholt/caddy/search?q=&type=Issues&utf8=%E2%9C%93)
+with a variety of keywords to ensure your suggestion/proposal is new.
+
+If so, you may open either an issue or a pull request for discussion and
+feedback.
+
+The advantage of issues is that you don't have to spend time implementing your
+idea, but you should still describe it thoroughly as if someone reading it would
+implement the whole thing starting from scratch.
+
+The advantage of pull requests is that we can immediately see the impact the
+change will have on the project, what the code will look like, and how to
+improve it. The disadvantage of pull requests is that they are unlikely to get
+accepted without significant changes first, or it may be rejected entirely.
+Don't worry, that won't happen without an open discussion first.
+
+If you are going to spend significant time writing code for a new pull request,
+best to open an issue to "claim" it and get feedback before you invest a lot of
+time.
+
+Remember: pull requests should always be thoroughly documented both via godoc
+and with at least a rough draft of documentation that might go on the website
+for users to read.
+
+
+### Collaborator status
+
+If your pull request is merged, congratulations! You're technically a
+collaborator. We may also grant you "Collaborator status" which means you can
+push to the repository and merge other pull requests. We hope that you will
+stay involved by reviewing pull requests, submitting more of your own, and
+resolving issues as you are able to. Thanks for making Caddy amazing!
+
+We ask that collaborators will conduct thorough code reviews and be nice to
+new contributors. Before merging a PR, it's best to get the approval of
+at least one or two other collaborators and/or the project owner. We prefer
+squashed commits instead of many little, semantically-unimportant commits. Also,
+CI and other post-commit hooks must pass before being merged except in certain
+unusual circumstances.
+
+Collaborator status may be removed for inactive users from time to time as
+we see fit; this is not an insult, just a basic security precaution in case
+the account becomes inactive or abandoned. Privileges can always be restored
+later.
+
+
+### Vulnerabilities
+
+If you've found a vulnerability that is serious, please email me: Matthew dot 
+Holt at Gmail. If it's not a big deal, a pull request will probably be faster.
+
+
+## Thank you
+
+Thanks for your help! Caddy would not be what it is today without your
+contributions.

ISSUE_TEMPLATE

@@ -0,0 +1,24 @@
+(Are you asking for help with using Caddy? Please use our forum instead: https://forum.caddyserver.com. If you are filing a bug report, please answer the following questions. If your issue is not a bug report, you do not need to use this template. Either way, please consider donating if we've helped you. Thanks!)
+
+#### 1. What version of Caddy are you running (`caddy -version`)?
+
+
+#### 2. What are you trying to do?
+
+
+#### 3. What is your entire Caddyfile?
+```text
+(Put Caddyfile here)
+```
+
+#### 4. How did you run Caddy (give the full command and describe the execution environment)?
+
+
+#### 5. What did you expect to see?
+
+
+#### 6. What did you see instead (give full error messages and/or log)?
+
+
+#### 7. How can someone who is starting from scratch reproduce this behavior as minimally as possible?
+

LICENSE.txt

@@ -1,4 +1,3 @@
-
                                  Apache License
                            Version 2.0, January 2004
                         http://www.apache.org/licenses/
@@ -179,7 +178,7 @@
    APPENDIX: How to apply the Apache License to your work.
 
       To apply the Apache License to your work, attach the following
-      boilerplate notice, with the fields enclosed by brackets "[]"
+      boilerplate notice, with the fields enclosed by brackets "{}"
       replaced with your own identifying information. (Don't include
       the brackets!)  The text should be enclosed in the appropriate
       comment syntax for the file format. We also recommend that a
@@ -187,7 +186,7 @@
       same "printed page" as the copyright notice for easier
       identification within third-party archives.
 
-   Copyright [yyyy] [name of copyright owner]
+   Copyright {yyyy} {name of copyright owner}
 
    Licensed under the Apache License, Version 2.0 (the "License");
    you may not use this file except in compliance with the License.

README.md

@@ -1,188 +1,158 @@
-<p align="center">
-	<a href="https://caddyserver.com"><img src="https://user-images.githubusercontent.com/1128849/36338535-05fb646a-136f-11e8-987b-e6901e717d5a.png" alt="Caddy" width="450"></a>
-	<br>
-	<h3 align="center">a <a href="https://zerossl.com"><img src="https://caddyserver.com/resources/images/zerossl-logo.svg" height="28" valign="middle"></a> project</h3>
-</p>
-<hr>
-<h3 align="center">Every site on HTTPS</h3>
-<p align="center">Caddy is an extensible server platform that uses TLS by default.</p>
-<p align="center">
-	<a href="https://github.com/caddyserver/caddy/actions?query=workflow%3ACross-Platform"><img src="https://github.com/caddyserver/caddy/workflows/Cross-Platform/badge.svg"></a>
-	<a href="https://pkg.go.dev/github.com/caddyserver/caddy/v2"><img src="https://img.shields.io/badge/godoc-reference-%23007d9c.svg"></a>
-	<br>
-	<a href="https://twitter.com/caddyserver" title="@caddyserver on Twitter"><img src="https://img.shields.io/badge/twitter-@caddyserver-55acee.svg" alt="@caddyserver on Twitter"></a>
-	<a href="https://caddy.community" title="Caddy Forum"><img src="https://img.shields.io/badge/community-forum-ff69b4.svg" alt="Caddy Forum"></a>
-	<br>
-	<a href="https://sourcegraph.com/github.com/caddyserver/caddy?badge" title="Caddy on Sourcegraph"><img src="https://sourcegraph.com/github.com/caddyserver/caddy/-/badge.svg" alt="Caddy on Sourcegraph"></a>
-	<a href="https://cloudsmith.io/~caddy/repos/"><img src="https://img.shields.io/badge/OSS%20hosting%20by-cloudsmith-blue?logo=cloudsmith" alt="Cloudsmith"></a>
-</p>
-<p align="center">
-	<a href="https://github.com/caddyserver/caddy/releases">Releases</a> ·
-	<a href="https://caddyserver.com/docs/">Documentation</a> ·
-	<a href="https://caddy.community">Get Help</a>
-</p>
+<a href="https://caddyserver.com"><img src="https://caddyserver.com/resources/images/caddy-lower.png" alt="Caddy" width="350"></a>
+
+[![community](https://img.shields.io/badge/community-forum-ff69b4.svg?style=flat-square)](https://forum.caddyserver.com) [![twitter](https://img.shields.io/badge/twitter-@caddyserver-55acee.svg?style=flat-square)](https://twitter.com/caddyserver) [![Documentation](https://img.shields.io/badge/godoc-reference-blue.svg?style=flat-square)](https://godoc.org/github.com/mholt/caddy) [![Linux Build Status](https://img.shields.io/travis/mholt/caddy.svg?style=flat-square&label=linux+build)](https://travis-ci.org/mholt/caddy) [![Windows Build Status](https://img.shields.io/appveyor/ci/mholt/caddy.svg?style=flat-square&label=windows+build)](https://ci.appveyor.com/project/mholt/caddy)
+
+Caddy is a general-purpose web server for Windows, Mac, Linux, BSD, and
+[Android](https://github.com/mholt/caddy/wiki/Running-Caddy-on-Android). It is
+a capable but easier alternative to other popular web servers.
+
+[Releases](https://github.com/mholt/caddy/releases) · 
+[User Guide](https://caddyserver.com/docs) · 
+[Community](https://forum.caddyserver.com)
 
 
-
-### Menu
+## Menu
 
 - [Features](#features)
-- [Install](#install)
-- [Build from source](#build-from-source)
-	- [For development](#for-development)
-	- [With version information and/or plugins](#with-version-information-andor-plugins)
-- [Quick start](#quick-start)
-- [Overview](#overview)
-- [Full documentation](#full-documentation)
-- [Getting help](#getting-help)
-- [About](#about)
-
-<p align="center">
-	<b>Powered by</b>
-	<br>
-	<a href="https://github.com/caddyserver/certmagic"><img src="https://user-images.githubusercontent.com/1128849/49704830-49d37200-fbd5-11e8-8385-767e0cd033c3.png" alt="CertMagic" width="250"></a>
-</p>
+- [Quick Start](#quick-start)
+- [Running from Source](#running-from-source)
+- [Running in Production](#running-in-production)
+- [Contributing](#contributing)
+- [About the Project](#about-the-project)
 
 
-## [Features](https://caddyserver.com/v2)
 
-- **Easy configuration** with the [Caddyfile](https://caddyserver.com/docs/caddyfile)
-- **Powerful configuration** with its [native JSON config](https://caddyserver.com/docs/json/)
-- **Dynamic configuration** with the [JSON API](https://caddyserver.com/docs/api)
-- [**Config adapters**](https://caddyserver.com/docs/config-adapters) if you don't like JSON
-- **Automatic HTTPS** by default
-	- [ZeroSSL](https://zerossl.com) and [Let's Encrypt](https://letsencrypt.org) for public names
-	- Fully-managed local CA for internal names & IPs
-	- Can coordinate with other Caddy instances in a cluster
-	- Multi-issuer fallback
-- **Stays up when other servers go down** due to TLS/OCSP/certificate-related issues
-- **Production-ready** after serving trillions of requests and managing millions of TLS certificates
-- **Scales to hundreds of thousands of sites** as proven in production
-- **HTTP/1.1, HTTP/2, and HTTP/3** supported all by default
-- **Highly extensible** [modular architecture](https://caddyserver.com/docs/architecture) lets Caddy do anything without bloat
+## Features
+
+- **Easy configuration** with Caddyfile
+- **Automatic HTTPS** via [Let's Encrypt](https://letsencrypt.org); Caddy
+obtains and manages all cryptographic assets for you
+- **HTTP/2** enabled by default (powered by Go standard library)
+- **Virtual hosting** for hundreds of sites per server instance, including TLS
+SNI
+- Experimental **QUIC support** for those that like speed
+- TLS session ticket **key rotation** for more secure connections
+- **Brilliant extensibility** so Caddy can be customized for your needs
 - **Runs anywhere** with **no external dependencies** (not even libc)
-- Written in Go, a language with higher **memory safety guarantees** than other servers
-- Actually **fun to use**
-- So much more to [discover](https://caddyserver.com/v2)
 
-## Install
 
-The simplest, cross-platform way to get started is to download Caddy from [GitHub Releases](https://github.com/caddyserver/caddy/releases) and place the executable file in your PATH.
 
-See [our online documentation](https://caddyserver.com/docs/install) for other install instructions.
+## Quick Start
 
-## Build from source
+Caddy binaries have no dependencies and are available for every platform.
+Install Caddy any one of these ways:
 
-Requirements:
+- **[Download page](https://caddyserver.com/download)** allows you to
+customize your build in the browser
+- **[Latest release](https://github.com/mholt/caddy/releases/latest)** for
+pre-built binaries
+- **curl [getcaddy.com](https://getcaddy.com)** for auto install:
+`curl https://getcaddy.com | bash`
 
-- [Go 1.18 or newer](https://golang.org/dl/)
+Once `caddy` is in your PATH, you can `cd` to your website's folder and run
+`caddy` to serve it. By default, Caddy serves the current directory at 
+[localhost:2015](http://localhost:2015).
 
-### For development
- 
-_**Note:** These steps [will not embed proper version information](https://github.com/golang/go/issues/29228). For that, please follow the instructions in the next section._
+To customize how your site is served, create a file named Caddyfile by your
+site and paste this into it:
 
-```bash
-$ git clone "https://github.com/caddyserver/caddy.git"
-$ cd caddy/cmd/caddy/
-$ go build
+```plain
+localhost
+
+gzip
+browse
+websocket /echo cat
+ext    .html
+log    /var/log/access.log
+proxy  /api 127.0.0.1:7005
+header /api Access-Control-Allow-Origin *
 ```
 
-When you run Caddy, it may try to bind to low ports unless otherwise specified in your config. If your OS requires elevated privileges for this, you will need to give your new binary permission to do so. On Linux, this can be done easily with: `sudo setcap cap_net_bind_service=+ep ./caddy`
+When you run `caddy` in that directory, it will automatically find and use
+that Caddyfile to configure itself.
 
-If you prefer to use `go run` which only creates temporary binaries, you can still do this with the included `setcap.sh` like so:
+This simple file enables compression, allows directory browsing (for folders 
+without an index file), hosts a WebSocket echo server at /echo, serves clean
+URLs, logs requests to access.log, proxies all API requests to a backend on
+port 7005, and adds the coveted  `Access-Control-Allow-Origin: *` header for
+all responses from the API.
 
-```bash
-$ go run -exec ./setcap.sh main.go
-```
+Wow! Caddy can do a lot with just a few lines.
 
-If you don't want to type your password for `setcap`, use `sudo visudo` to edit your sudoers file and allow your user account to run that command without a password, for example:
+To host multiple sites and do more with the Caddyfile, please see the
+[Caddyfile documentation](https://caddyserver.com/docs/caddyfile).
 
-```
-username ALL=(ALL:ALL) NOPASSWD: /usr/sbin/setcap
-```
+Note that production sites (not localhost) are served over
+[HTTPS by default](https://caddyserver.com/docs/automatic-https).
 
-replacing `username` with your actual username. Please be careful and only do this if you know what you are doing! We are only qualified to document how to use Caddy, not Go tooling or your computer, and we are providing these instructions for convenience only; please learn how to use your own computer at your own risk and make any needful adjustments.
+Caddy has a command line interface. Run `caddy -h` to view basic help or see
+the [CLI documentation](https://caddyserver.com/docs/cli) for details.
 
-### With version information and/or plugins
-
-Using [our builder tool, `xcaddy`](https://github.com/caddyserver/xcaddy)...
-
-```
-$ xcaddy build
-```
-
-...the following steps are automated:
-
-1. Create a new folder: `mkdir caddy`
-2. Change into it: `cd caddy`
-3. Copy [Caddy's main.go](https://github.com/caddyserver/caddy/blob/master/cmd/caddy/main.go) into the empty folder. Add imports for any custom plugins you want to add.
-4. Initialize a Go module: `go mod init caddy`
-5. (Optional) Pin Caddy version: `go get github.com/caddyserver/caddy/v2@version` replacing `version` with a git tag, commit, or branch name.
-6. (Optional) Add plugins by adding their import: `_ "import/path/here"`
-7. Compile: `go build`
+**Running as root:** We advise against this. You can still listen on ports
+< 1024 using setcap like so: `sudo setcap cap_net_bind_service=+ep ./caddy`
 
 
 
+## Running from Source
 
-## Quick start
+Note: You will need **[Go 1.6](https://golang.org/dl/)** or newer.
 
-The [Caddy website](https://caddyserver.com/docs/) has documentation that includes tutorials, quick-start guides, reference, and more.
+1. `go get github.com/mholt/caddy`
+2. `cd` into your website's directory
+3. Run `caddy` (assumes `$GOPATH/bin` is in your `$PATH`)
 
-**We recommend that all users -- regardless of experience level -- do our [Getting Started](https://caddyserver.com/docs/getting-started) guide to become familiar with using Caddy.**
-
-If you've only got a minute, [the website has several quick-start tutorials](https://caddyserver.com/docs/quick-starts) to choose from! However, after finishing a quick-start tutorial, please read more documentation to understand how the software works. 🙂
+Caddy's `main()` is in the caddy subfolder. To recompile Caddy, use
+`build.bash` found in that folder.
 
 
 
+## Running in Production
 
-## Overview
+The Caddy project does not officially maintain any system-specific
+integrations, but your download file includes
+[unofficial resources](https://github.com/mholt/caddy/tree/master/dist/init)
+contributed by the community that you may find helpful for running Caddy in
+production.
 
-Caddy is most often used as an HTTPS server, but it is suitable for any long-running Go program. First and foremost, it is a platform to run Go applications. Caddy "apps" are just Go programs that are implemented as Caddy modules. Two apps -- `tls` and `http` -- ship standard with Caddy.
-
-Caddy apps instantly benefit from [automated documentation](https://caddyserver.com/docs/json/), graceful on-line [config changes via API](https://caddyserver.com/docs/api), and unification with other Caddy apps.
-
-Although [JSON](https://caddyserver.com/docs/json/) is Caddy's native config language, Caddy can accept input from [config adapters](https://caddyserver.com/docs/config-adapters) which can essentially convert any config format of your choice into JSON: Caddyfile, JSON 5, YAML, TOML, NGINX config, and more.
-
-The primary way to configure Caddy is through [its API](https://caddyserver.com/docs/api), but if you prefer config files, the [command-line interface](https://caddyserver.com/docs/command-line) supports those too.
-
-Caddy exposes an unprecedented level of control compared to any web server in existence. In Caddy, you are usually setting the actual values of the initialized types in memory that power everything from your HTTP handlers and TLS handshakes to your storage medium. Caddy is also ridiculously extensible, with a powerful plugin system that makes vast improvements over other web servers.
-
-To wield the power of this design, you need to know how the config document is structured. Please see [our documentation site](https://caddyserver.com/docs/) for details about [Caddy's config structure](https://caddyserver.com/docs/json/).
-
-Nearly all of Caddy's configuration is contained in a single config document, rather than being scattered across CLI flags and env variables and a configuration file as with other web servers. This makes managing your server config more straightforward and reduces hidden variables/factors.
-
-
-## Full documentation
-
-Our website has complete documentation:
-
-**https://caddyserver.com/docs/**
-
-The docs are also open source. You can contribute to them here: https://github.com/caddyserver/website
+How you choose to run Caddy is up to you. Many users are satisfied with
+`nohup caddy &`. Others use `screen`. Users who need Caddy to come back up
+after reboots either do so in the script that caused the reboot, add a command
+to an init script, or configure a service with their OS.
 
 
 
-## Getting help
+## Contributing
 
-- We advise companies using Caddy to secure a support contract through [Ardan Labs](https://www.ardanlabs.com/my/contact-us?dd=caddy) before help is needed.
+**[Join our community](https://forum.caddyserver.com) where you can chat with
+other Caddy users and developers!**
 
-- A [sponsorship](https://github.com/sponsors/mholt) goes a long way! We can offer private help to sponsors. If Caddy is benefitting your company, please consider a sponsorship. This not only helps fund full-time work to ensure the longevity of the project, it provides your company the resources, support, and discounts you need; along with being a great look for your company to your customers and potential customers!
+Please see our [contributing guidelines](https://github.com/mholt/caddy/blob/master/CONTRIBUTING.md)
+and check out the [developer wiki](https://github.com/mholt/caddy/wiki).
 
-- Individuals can exchange help for free on our community forum at https://caddy.community. Remember that people give help out of their spare time and good will. The best way to get help is to give it first!
+We use GitHub issues and pull requests only for discussing bug reports and
+the development of specific changes. We welcome all other topics on the
+[forum](https://forum.caddyserver.com)!
 
-Please use our [issue tracker](https://github.com/caddyserver/caddy/issues) only for bug reports and feature requests, i.e. actionable development items (support questions will usually be referred to the forums).
+Thanks for making Caddy -- and the Web -- better!
+
+Special thanks to
+[![DigitalOcean](https://i.imgur.com/sfGr0eY.png)](https://www.digitalocean.com)
+for hosting the Caddy project.
 
 
 
-## About
+## About the Project
 
-Matthew Holt began developing Caddy in 2014 while studying computer science at Brigham Young University. (The name "Caddy" was chosen because this software helps with the tedious, mundane tasks of serving the Web, and is also a single place for multiple things to be organized together.) It soon became the first web server to use HTTPS automatically and by default, and now has hundreds of contributors and has served trillions of HTTPS requests.
+Caddy was born out of the need for a "batteries-included" web server that runs
+anywhere and doesn't have to take its configuration with it. Caddy took
+inspiration from [spark](https://github.com/rif/spark),
+[nginx](https://github.com/nginx/nginx), lighttpd,
+[Websocketd](https://github.com/joewalnes/websocketd)
+and [Vagrant](https://www.vagrantup.com/),
+which provides a pleasant mixture of features from each of them.
 
-**The name "Caddy" is trademarked.** The name of the software is "Caddy", not "Caddy Server" or "CaddyServer". Please call it "Caddy" or, if you wish to clarify, "the Caddy web server". Caddy is a registered trademark of Stack Holdings GmbH.
+**The name "Caddy":** The name of the software is "Caddy", not "Caddy Server"
+or "CaddyServer". Please call it "Caddy" or, if you wish to clarify, "the
+Caddy web server". See [brand guidelines](https://caddyserver.com/brand).
 
-- _Project on Twitter: [@caddyserver](https://twitter.com/caddyserver)_
-- _Author on Twitter: [@mholt6](https://twitter.com/mholt6)_
-
-Caddy is a project of [ZeroSSL](https://zerossl.com), a Stack Holdings company.
-
-Debian package repository hosting is graciously provided by [Cloudsmith](https://cloudsmith.com). Cloudsmith is the only fully hosted, cloud-native, universal package management solution, that enables your organization to create, store and share packages in any format, to any place, with total confidence.
+*Author on Twitter: [@mholt6](https://twitter.com/mholt6)*

admin_test.go

@@ -1,199 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddy
-
-import (
-	"encoding/json"
-	"fmt"
-	"net/http"
-	"reflect"
-	"sync"
-	"testing"
-)
-
-var testCfg = []byte(`{
-			"apps": {
-				"http": {
-					"servers": {
-						"myserver": {
-							"listen": ["tcp/localhost:8080-8084"],
-							"read_timeout": "30s"
-						},
-						"yourserver": {
-							"listen": ["127.0.0.1:5000"],
-							"read_header_timeout": "15s"
-						}
-					}
-				}
-			}
-		}
-		`)
-
-func TestUnsyncedConfigAccess(t *testing.T) {
-	// each test is performed in sequence, so
-	// each change builds on the previous ones;
-	// the config is not reset between tests
-	for i, tc := range []struct {
-		method    string
-		path      string // rawConfigKey will be prepended
-		payload   string
-		expect    string // JSON representation of what the whole config is expected to be after the request
-		shouldErr bool
-	}{
-		{
-			method:  "POST",
-			path:    "",
-			payload: `{"foo": "bar", "list": ["a", "b", "c"]}`, // starting value
-			expect:  `{"foo": "bar", "list": ["a", "b", "c"]}`,
-		},
-		{
-			method:  "POST",
-			path:    "/foo",
-			payload: `"jet"`,
-			expect:  `{"foo": "jet", "list": ["a", "b", "c"]}`,
-		},
-		{
-			method:  "POST",
-			path:    "/bar",
-			payload: `{"aa": "bb", "qq": "zz"}`,
-			expect:  `{"foo": "jet", "bar": {"aa": "bb", "qq": "zz"}, "list": ["a", "b", "c"]}`,
-		},
-		{
-			method: "DELETE",
-			path:   "/bar/qq",
-			expect: `{"foo": "jet", "bar": {"aa": "bb"}, "list": ["a", "b", "c"]}`,
-		},
-		{
-			method:  "POST",
-			path:    "/list",
-			payload: `"e"`,
-			expect:  `{"foo": "jet", "bar": {"aa": "bb"}, "list": ["a", "b", "c", "e"]}`,
-		},
-		{
-			method:  "PUT",
-			path:    "/list/3",
-			payload: `"d"`,
-			expect:  `{"foo": "jet", "bar": {"aa": "bb"}, "list": ["a", "b", "c", "d", "e"]}`,
-		},
-		{
-			method: "DELETE",
-			path:   "/list/3",
-			expect: `{"foo": "jet", "bar": {"aa": "bb"}, "list": ["a", "b", "c", "e"]}`,
-		},
-		{
-			method:  "PATCH",
-			path:    "/list/3",
-			payload: `"d"`,
-			expect:  `{"foo": "jet", "bar": {"aa": "bb"}, "list": ["a", "b", "c", "d"]}`,
-		},
-		{
-			method:  "POST",
-			path:    "/list/...",
-			payload: `["e", "f", "g"]`,
-			expect:  `{"foo": "jet", "bar": {"aa": "bb"}, "list": ["a", "b", "c", "d", "e", "f", "g"]}`,
-		},
-	} {
-		err := unsyncedConfigAccess(tc.method, rawConfigKey+tc.path, []byte(tc.payload), nil)
-
-		if tc.shouldErr && err == nil {
-			t.Fatalf("Test %d: Expected error return value, but got: %v", i, err)
-		}
-		if !tc.shouldErr && err != nil {
-			t.Fatalf("Test %d: Should not have had error return value, but got: %v", i, err)
-		}
-
-		// decode the expected config so we can do a convenient DeepEqual
-		var expectedDecoded any
-		err = json.Unmarshal([]byte(tc.expect), &expectedDecoded)
-		if err != nil {
-			t.Fatalf("Test %d: Unmarshaling expected config: %v", i, err)
-		}
-
-		// make sure the resulting config is as we expect it
-		if !reflect.DeepEqual(rawCfg[rawConfigKey], expectedDecoded) {
-			t.Fatalf("Test %d:\nExpected:\n\t%#v\nActual:\n\t%#v",
-				i, expectedDecoded, rawCfg[rawConfigKey])
-		}
-	}
-}
-
-// TestLoadConcurrent exercises Load under concurrent conditions
-// and is most useful under test with `-race` enabled.
-func TestLoadConcurrent(t *testing.T) {
-	var wg sync.WaitGroup
-
-	for i := 0; i < 100; i++ {
-		wg.Add(1)
-		go func() {
-			_ = Load(testCfg, true)
-			wg.Done()
-		}()
-	}
-	wg.Wait()
-}
-
-type fooModule struct {
-	IntField int
-	StrField string
-}
-
-func (fooModule) CaddyModule() ModuleInfo {
-	return ModuleInfo{
-		ID:  "foo",
-		New: func() Module { return new(fooModule) },
-	}
-}
-func (fooModule) Start() error { return nil }
-func (fooModule) Stop() error  { return nil }
-
-func TestETags(t *testing.T) {
-	RegisterModule(fooModule{})
-
-	if err := Load([]byte(`{"apps": {"foo": {"strField": "abc", "intField": 0}}}`), true); err != nil {
-		t.Fatalf("loading: %s", err)
-	}
-
-	const key = "/" + rawConfigKey + "/apps/foo"
-
-	// try update the config with the wrong etag
-	err := changeConfig(http.MethodPost, key, []byte(`{"strField": "abc", "intField": 1}}`), fmt.Sprintf(`"/%s not_an_etag"`, rawConfigKey), false)
-	if apiErr, ok := err.(APIError); !ok || apiErr.HTTPStatus != http.StatusPreconditionFailed {
-		t.Fatalf("expected precondition failed; got %v", err)
-	}
-
-	// get the etag
-	hash := etagHasher()
-	if err := readConfig(key, hash); err != nil {
-		t.Fatalf("reading: %s", err)
-	}
-
-	// do the same update with the correct key
-	err = changeConfig(http.MethodPost, key, []byte(`{"strField": "abc", "intField": 1}`), makeEtag(key, hash), false)
-	if err != nil {
-		t.Fatalf("expected update to work; got %v", err)
-	}
-
-	// now try another update. The hash should no longer match and we should get precondition failed
-	err = changeConfig(http.MethodPost, key, []byte(`{"strField": "abc", "intField": 2}`), makeEtag(key, hash), false)
-	if apiErr, ok := err.(APIError); !ok || apiErr.HTTPStatus != http.StatusPreconditionFailed {
-		t.Fatalf("expected precondition failed; got %v", err)
-	}
-}
-
-func BenchmarkLoad(b *testing.B) {
-	for i := 0; i < b.N; i++ {
-		Load(testCfg, true)
-	}
-}

appveyor.yml

@@ -0,0 +1,32 @@
+version: "{build}"
+
+os: Windows Server 2012 R2
+
+clone_folder: c:\gopath\src\github.com\mholt\caddy
+
+environment:
+  GOPATH: c:\gopath
+  CGO_ENABLED: 0
+
+install:
+  - rmdir c:\go /s /q
+  - appveyor DownloadFile https://storage.googleapis.com/golang/go1.6.2.windows-amd64.zip
+  - 7z x go1.6.2.windows-amd64.zip -y -oC:\ > NUL
+  - go version
+  - go env
+  - go get -t ./...
+  - go get github.com/golang/lint/golint
+  - go get github.com/gordonklaus/ineffassign
+  - set PATH=%GOPATH%\bin;%PATH%
+
+build: off
+
+test_script:
+  - go vet ./...
+  - go test ./...
+  - ineffassign .
+
+after_test:
+  - golint ./...
+
+deploy: off

assets.go

@@ -0,0 +1,34 @@
+package caddy
+
+import (
+	"os"
+	"path/filepath"
+	"runtime"
+)
+
+// AssetsPath returns the path to the folder
+// where the application may store data. If
+// CADDYPATH env variable is set, that value
+// is used. Otherwise, the path is the result
+// of evaluating "$HOME/.caddy".
+func AssetsPath() string {
+	if caddyPath := os.Getenv("CADDYPATH"); caddyPath != "" {
+		return caddyPath
+	}
+	return filepath.Join(userHomeDir(), ".caddy")
+}
+
+// userHomeDir returns the user's home directory according to
+// environment variables.
+//
+// Credit: http://stackoverflow.com/a/7922977/1048862
+func userHomeDir() string {
+	if runtime.GOOS == "windows" {
+		home := os.Getenv("HOMEDRIVE") + os.Getenv("HOMEPATH")
+		if home == "" {
+			home = os.Getenv("USERPROFILE")
+		}
+		return home
+	}
+	return os.Getenv("HOME")
+}

assets_test.go

@@ -0,0 +1,19 @@
+package caddy
+
+import (
+	"os"
+	"strings"
+	"testing"
+)
+
+func TestAssetsPath(t *testing.T) {
+	if actual := AssetsPath(); !strings.HasSuffix(actual, ".caddy") {
+		t.Errorf("Expected path to be a .caddy folder, got: %v", actual)
+	}
+
+	os.Setenv("CADDYPATH", "testpath")
+	if actual, expected := AssetsPath(), "testpath"; actual != expected {
+		t.Errorf("Expected path to be %v, got: %v", expected, actual)
+	}
+	os.Setenv("CADDYPATH", "")
+}

caddy/build.bash

@@ -0,0 +1,56 @@
+#!/usr/bin/env bash
+#
+# Caddy build script. Automates proper versioning.
+#
+# Usage:
+#
+#     $ ./build.bash [output_filename] [git_repo]
+#
+# Outputs compiled program in current directory.
+# Default git repo is current directory.
+# Builds always take place from current directory.
+
+set -euo pipefail
+
+: ${output_filename:="${1:-}"}
+: ${output_filename:="caddy"}
+
+: ${git_repo:="${2:-}"}
+: ${git_repo:="."}
+
+pkg=github.com/mholt/caddy/caddy/caddymain
+ldflags=()
+
+# Timestamp of build
+name="${pkg}.buildDate"
+value=$(date -u +"%a %b %d %H:%M:%S %Z %Y")
+ldflags+=("-X" "\"${name}=${value}\"")
+
+# Current tag, if HEAD is on a tag
+name="${pkg}.gitTag"
+set +e
+value="$(git -C "${git_repo}" describe --exact-match HEAD 2>/dev/null)"
+set -e
+ldflags+=("-X" "\"${name}=${value}\"")
+
+# Nearest tag on branch
+name="${pkg}.gitNearestTag"
+value="$(git -C "${git_repo}" describe --abbrev=0 --tags HEAD)"
+ldflags+=("-X" "\"${name}=${value}\"")
+
+# Commit SHA
+name="${pkg}.gitCommit"
+value="$(git -C "${git_repo}" rev-parse --short HEAD)"
+ldflags+=("-X" "\"${name}=${value}\"")
+
+# Summary of uncommitted changes
+name="${pkg}.gitShortStat"
+value="$(git -C "${git_repo}" diff-index --shortstat HEAD)"
+ldflags+=("-X" "\"${name}=${value}\"")
+
+# List of modified files
+name="${pkg}.gitFilesModified"
+value="$(git -C "${git_repo}" diff-index --name-only HEAD)"
+ldflags+=("-X" "\"${name}=${value}\"")
+
+go build -ldflags "${ldflags[*]}" -o "${output_filename}"

caddy/caddymain/run.go

@@ -0,0 +1,279 @@
+package caddymain
+
+import (
+	"errors"
+	"flag"
+	"fmt"
+	"io/ioutil"
+	"log"
+	"os"
+	"path/filepath"
+	"runtime"
+	"strconv"
+	"strings"
+
+	"gopkg.in/natefinch/lumberjack.v2"
+
+	"github.com/xenolf/lego/acme"
+
+	"github.com/mholt/caddy"
+	// plug in the HTTP server type
+	_ "github.com/mholt/caddy/caddyhttp"
+
+	"github.com/mholt/caddy/caddytls"
+	// This is where other plugins get plugged in (imported)
+)
+
+func init() {
+	caddy.TrapSignals()
+	setVersion()
+
+	flag.BoolVar(&caddytls.Agreed, "agree", false, "Agree to the CA's Subscriber Agreement")
+	flag.StringVar(&caddytls.DefaultCAUrl, "ca", "https://acme-v01.api.letsencrypt.org/directory", "URL to certificate authority's ACME server directory")
+	flag.StringVar(&conf, "conf", "", "Caddyfile to load (default \""+caddy.DefaultConfigFile+"\")")
+	flag.StringVar(&cpu, "cpu", "100%", "CPU cap")
+	flag.BoolVar(&plugins, "plugins", false, "List installed plugins")
+	flag.StringVar(&caddytls.DefaultEmail, "email", "", "Default ACME CA account email address")
+	flag.StringVar(&logfile, "log", "", "Process log file")
+	flag.StringVar(&caddy.PidFile, "pidfile", "", "Path to write pid file")
+	flag.BoolVar(&caddy.Quiet, "quiet", false, "Quiet mode (no initialization output)")
+	flag.StringVar(&revoke, "revoke", "", "Hostname for which to revoke the certificate")
+	flag.StringVar(&serverType, "type", "http", "Type of server to run")
+	flag.BoolVar(&version, "version", false, "Show version")
+
+	caddy.RegisterCaddyfileLoader("flag", caddy.LoaderFunc(confLoader))
+	caddy.SetDefaultCaddyfileLoader("default", caddy.LoaderFunc(defaultLoader))
+}
+
+// Run is Caddy's main() function.
+func Run() {
+	flag.Parse()
+	moveStorage() // TODO: This is temporary for the 0.9 release, or until most users upgrade to 0.9+
+
+	caddy.AppName = appName
+	caddy.AppVersion = appVersion
+	acme.UserAgent = appName + "/" + appVersion
+
+	// Set up process log before anything bad happens
+	switch logfile {
+	case "stdout":
+		log.SetOutput(os.Stdout)
+	case "stderr":
+		log.SetOutput(os.Stderr)
+	case "":
+		log.SetOutput(ioutil.Discard)
+	default:
+		log.SetOutput(&lumberjack.Logger{
+			Filename:   logfile,
+			MaxSize:    100,
+			MaxAge:     14,
+			MaxBackups: 10,
+		})
+	}
+
+	// Check for one-time actions
+	if revoke != "" {
+		err := caddytls.Revoke(revoke)
+		if err != nil {
+			log.Fatal(err)
+		}
+		fmt.Printf("Revoked certificate for %s\n", revoke)
+		os.Exit(0)
+	}
+	if version {
+		fmt.Printf("%s %s\n", appName, appVersion)
+		if devBuild && gitShortStat != "" {
+			fmt.Printf("%s\n%s\n", gitShortStat, gitFilesModified)
+		}
+		os.Exit(0)
+	}
+	if plugins {
+		fmt.Println(caddy.DescribePlugins())
+		os.Exit(0)
+	}
+
+	// Set CPU cap
+	err := setCPU(cpu)
+	if err != nil {
+		mustLogFatal(err)
+	}
+
+	// Get Caddyfile input
+	caddyfile, err := caddy.LoadCaddyfile(serverType)
+	if err != nil {
+		mustLogFatal(err)
+	}
+
+	// Start your engines
+	instance, err := caddy.Start(caddyfile)
+	if err != nil {
+		mustLogFatal(err)
+	}
+
+	// Twiddle your thumbs
+	instance.Wait()
+}
+
+// mustLogFatal wraps log.Fatal() in a way that ensures the
+// output is always printed to stderr so the user can see it
+// if the user is still there, even if the process log was not
+// enabled. If this process is an upgrade, however, and the user
+// might not be there anymore, this just logs to the process
+// log and exits.
+func mustLogFatal(args ...interface{}) {
+	if !caddy.IsUpgrade() {
+		log.SetOutput(os.Stderr)
+	}
+	log.Fatal(args...)
+}
+
+// confLoader loads the Caddyfile using the -conf flag.
+func confLoader(serverType string) (caddy.Input, error) {
+	if conf == "" {
+		return nil, nil
+	}
+
+	if conf == "stdin" {
+		return caddy.CaddyfileFromPipe(os.Stdin)
+	}
+
+	contents, err := ioutil.ReadFile(conf)
+	if err != nil {
+		return nil, err
+	}
+	return caddy.CaddyfileInput{
+		Contents:       contents,
+		Filepath:       conf,
+		ServerTypeName: serverType,
+	}, nil
+}
+
+// defaultLoader loads the Caddyfile from the current working directory.
+func defaultLoader(serverType string) (caddy.Input, error) {
+	contents, err := ioutil.ReadFile(caddy.DefaultConfigFile)
+	if err != nil {
+		if os.IsNotExist(err) {
+			return nil, nil
+		}
+		return nil, err
+	}
+	return caddy.CaddyfileInput{
+		Contents:       contents,
+		Filepath:       caddy.DefaultConfigFile,
+		ServerTypeName: serverType,
+	}, nil
+}
+
+// moveStorage moves the old certificate storage location by
+// renaming the "letsencrypt" folder to the hostname of the
+// CA URL. This is TEMPORARY until most users have upgraded to 0.9+.
+func moveStorage() {
+	oldPath := filepath.Join(caddy.AssetsPath(), "letsencrypt")
+	_, err := os.Stat(oldPath)
+	if os.IsNotExist(err) {
+		return
+	}
+	newPath, err := caddytls.StorageFor(caddytls.DefaultCAUrl)
+	if err != nil {
+		log.Fatalf("[ERROR] Unable to get new path for certificate storage: %v", err)
+	}
+	err = os.MkdirAll(string(newPath), 0700)
+	if err != nil {
+		log.Fatalf("[ERROR] Unable to make new certificate storage path: %v", err)
+	}
+	err = os.Rename(oldPath, string(newPath))
+	if err != nil {
+		log.Fatalf("[ERROR] Unable to migrate certificate storage: %v", err)
+	}
+	// convert mixed case folder and file names to lowercase
+	filepath.Walk(string(newPath), func(path string, info os.FileInfo, err error) error {
+		// must be careful to only lowercase the base of the path, not the whole thing!!
+		base := filepath.Base(path)
+		if lowerBase := strings.ToLower(base); base != lowerBase {
+			lowerPath := filepath.Join(filepath.Dir(path), lowerBase)
+			err = os.Rename(path, lowerPath)
+			if err != nil {
+				log.Fatalf("[ERROR] Unable to lower-case: %v", err)
+			}
+		}
+		return nil
+	})
+}
+
+// setVersion figures out the version information
+// based on variables set by -ldflags.
+func setVersion() {
+	// A development build is one that's not at a tag or has uncommitted changes
+	devBuild = gitTag == "" || gitShortStat != ""
+
+	// Only set the appVersion if -ldflags was used
+	if gitNearestTag != "" || gitTag != "" {
+		if devBuild && gitNearestTag != "" {
+			appVersion = fmt.Sprintf("%s (+%s %s)",
+				strings.TrimPrefix(gitNearestTag, "v"), gitCommit, buildDate)
+		} else if gitTag != "" {
+			appVersion = strings.TrimPrefix(gitTag, "v")
+		}
+	}
+}
+
+// setCPU parses string cpu and sets GOMAXPROCS
+// according to its value. It accepts either
+// a number (e.g. 3) or a percent (e.g. 50%).
+func setCPU(cpu string) error {
+	var numCPU int
+
+	availCPU := runtime.NumCPU()
+
+	if strings.HasSuffix(cpu, "%") {
+		// Percent
+		var percent float32
+		pctStr := cpu[:len(cpu)-1]
+		pctInt, err := strconv.Atoi(pctStr)
+		if err != nil || pctInt < 1 || pctInt > 100 {
+			return errors.New("invalid CPU value: percentage must be between 1-100")
+		}
+		percent = float32(pctInt) / 100
+		numCPU = int(float32(availCPU) * percent)
+	} else {
+		// Number
+		num, err := strconv.Atoi(cpu)
+		if err != nil || num < 1 {
+			return errors.New("invalid CPU value: provide a number or percent greater than 0")
+		}
+		numCPU = num
+	}
+
+	if numCPU > availCPU {
+		numCPU = availCPU
+	}
+
+	runtime.GOMAXPROCS(numCPU)
+	return nil
+}
+
+const appName = "Caddy"
+
+// Flags that control program flow or startup
+var (
+	serverType string
+	conf       string
+	cpu        string
+	logfile    string
+	revoke     string
+	version    bool
+	plugins    bool
+)
+
+// Build information obtained with the help of -ldflags
+var (
+	appVersion = "(untracked dev build)" // inferred at startup
+	devBuild   = true                    // inferred at startup
+
+	buildDate        string // date -u
+	gitTag           string // git describe --exact-match HEAD 2> /dev/null
+	gitNearestTag    string // git describe --abbrev=0 --tags HEAD
+	gitCommit        string // git rev-parse HEAD
+	gitShortStat     string // git diff-index --shortstat
+	gitFilesModified string // git diff-index --name-only HEAD
+)

caddy/caddymain/run_test.go

@@ -0,0 +1,44 @@
+package caddymain
+
+import (
+	"runtime"
+	"testing"
+)
+
+func TestSetCPU(t *testing.T) {
+	currentCPU := runtime.GOMAXPROCS(-1)
+	maxCPU := runtime.NumCPU()
+	halfCPU := int(0.5 * float32(maxCPU))
+	if halfCPU < 1 {
+		halfCPU = 1
+	}
+	for i, test := range []struct {
+		input     string
+		output    int
+		shouldErr bool
+	}{
+		{"1", 1, false},
+		{"-1", currentCPU, true},
+		{"0", currentCPU, true},
+		{"100%", maxCPU, false},
+		{"50%", halfCPU, false},
+		{"110%", currentCPU, true},
+		{"-10%", currentCPU, true},
+		{"invalid input", currentCPU, true},
+		{"invalid input%", currentCPU, true},
+		{"9999", maxCPU, false}, // over available CPU
+	} {
+		err := setCPU(test.input)
+		if test.shouldErr && err == nil {
+			t.Errorf("Test %d: Expected error, but there wasn't any", i)
+		}
+		if !test.shouldErr && err != nil {
+			t.Errorf("Test %d: Expected no error, but there was one: %v", i, err)
+		}
+		if actual, expected := runtime.GOMAXPROCS(-1), test.output; actual != expected {
+			t.Errorf("Test %d: GOMAXPROCS was %d but expected %d", i, actual, expected)
+		}
+		// teardown
+		runtime.GOMAXPROCS(currentCPU)
+	}
+}

caddy/main.go

@@ -0,0 +1,14 @@
+// By moving the application's package main logic into
+// a package other than main, it becomes much easier to
+// wrap caddy for custom builds that are go-gettable.
+// https://forum.caddyserver.com/t/my-wish-for-0-9-go-gettable-custom-builds/59?u=matt
+
+package main
+
+import "github.com/mholt/caddy/caddy/caddymain"
+
+var run = caddymain.Run // replaced for tests
+
+func main() {
+	run()
+}

caddy/main_test.go

@@ -0,0 +1,17 @@
+package main
+
+import "testing"
+
+// This works because it does not have the same signature as the
+// conventional "TestMain" function described in the testing package
+// godoc.
+func TestMain(t *testing.T) {
+	var ran bool
+	run = func() {
+		ran = true
+	}
+	main()
+	if !ran {
+		t.Error("Expected Run() to be called, but it wasn't")
+	}
+}

caddy_test.go

@@ -1,74 +1,58 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
 package caddy
 
-import (
-	"testing"
-	"time"
-)
+import "testing"
 
-func TestParseDuration(t *testing.T) {
-	const day = 24 * time.Hour
-	for i, tc := range []struct {
-		input  string
-		expect time.Duration
-	}{
-		{
-			input:  "3h",
-			expect: 3 * time.Hour,
-		},
-		{
-			input:  "1d",
-			expect: day,
-		},
-		{
-			input:  "1d30m",
-			expect: day + 30*time.Minute,
-		},
-		{
-			input:  "1m2d",
-			expect: time.Minute + day*2,
-		},
-		{
-			input:  "1m2d30s",
-			expect: time.Minute + day*2 + 30*time.Second,
-		},
-		{
-			input:  "1d2d",
-			expect: 3 * day,
-		},
-		{
-			input:  "1.5d",
-			expect: time.Duration(1.5 * float64(day)),
-		},
-		{
-			input:  "4m1.25d",
-			expect: 4*time.Minute + time.Duration(1.25*float64(day)),
-		},
-		{
-			input:  "-1.25d12h",
-			expect: time.Duration(-1.25*float64(day)) - 12*time.Hour,
-		},
-	} {
-		actual, err := ParseDuration(tc.input)
+/*
+// TODO
+func TestCaddyStartStop(t *testing.T) {
+	caddyfile := "localhost:1984"
+
+	for i := 0; i < 2; i++ {
+		_, err := Start(CaddyfileInput{Contents: []byte(caddyfile)})
 		if err != nil {
-			t.Errorf("Test %d ('%s'): Got error: %v", i, tc.input, err)
-			continue
+			t.Fatalf("Error starting, iteration %d: %v", i, err)
 		}
-		if actual != tc.expect {
-			t.Errorf("Test %d ('%s'): Expected=%s Actual=%s", i, tc.input, tc.expect, actual)
+
+		client := http.Client{
+			Timeout: time.Duration(2 * time.Second),
+		}
+		resp, err := client.Get("http://localhost:1984")
+		if err != nil {
+			t.Fatalf("Expected GET request to succeed (iteration %d), but it failed: %v", i, err)
+		}
+		resp.Body.Close()
+
+		err = Stop()
+		if err != nil {
+			t.Fatalf("Error stopping, iteration %d: %v", i, err)
+		}
+	}
+}
+*/
+
+func TestIsLoopback(t *testing.T) {
+	for i, test := range []struct {
+		input  string
+		expect bool
+	}{
+		{"example.com", false},
+		{"localhost", true},
+		{"localhost:1234", true},
+		{"localhost:", true},
+		{"127.0.0.1", true},
+		{"127.0.0.1:443", true},
+		{"127.0.1.5", true},
+		{"10.0.0.5", false},
+		{"12.7.0.1", false},
+		{"[::1]", true},
+		{"[::1]:1234", true},
+		{"::1", true},
+		{"::", false},
+		{"[::]", false},
+		{"local", false},
+	} {
+		if got, want := IsLoopback(test.input), test.expect; got != want {
+			t.Errorf("Test %d (%s): expected %v but was %v", i, test.input, want, got)
 		}
 	}
 }

caddyconfig/caddyfile/adapter.go

@@ -1,149 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyfile
-
-import (
-	"bytes"
-	"encoding/json"
-	"fmt"
-
-	"github.com/caddyserver/caddy/v2"
-	"github.com/caddyserver/caddy/v2/caddyconfig"
-)
-
-// Adapter adapts Caddyfile to Caddy JSON.
-type Adapter struct {
-	ServerType ServerType
-}
-
-// Adapt converts the Caddyfile config in body to Caddy JSON.
-func (a Adapter) Adapt(body []byte, options map[string]any) ([]byte, []caddyconfig.Warning, error) {
-	if a.ServerType == nil {
-		return nil, nil, fmt.Errorf("no server type")
-	}
-	if options == nil {
-		options = make(map[string]any)
-	}
-
-	filename, _ := options["filename"].(string)
-	if filename == "" {
-		filename = "Caddyfile"
-	}
-
-	serverBlocks, err := Parse(filename, body)
-	if err != nil {
-		return nil, nil, err
-	}
-
-	cfg, warnings, err := a.ServerType.Setup(serverBlocks, options)
-	if err != nil {
-		return nil, warnings, err
-	}
-
-	// lint check: see if input was properly formatted; sometimes messy files files parse
-	// successfully but result in logical errors (the Caddyfile is a bad format, I'm sorry)
-	if warning, different := formattingDifference(filename, body); different {
-		warnings = append(warnings, warning)
-	}
-
-	result, err := json.Marshal(cfg)
-
-	return result, warnings, err
-}
-
-// formattingDifference returns a warning and true if the formatted version
-// is any different from the input; empty warning and false otherwise.
-// TODO: also perform this check on imported files
-func formattingDifference(filename string, body []byte) (caddyconfig.Warning, bool) {
-	// replace windows-style newlines to normalize comparison
-	normalizedBody := bytes.Replace(body, []byte("\r\n"), []byte("\n"), -1)
-
-	formatted := Format(normalizedBody)
-	if bytes.Equal(formatted, normalizedBody) {
-		return caddyconfig.Warning{}, false
-	}
-
-	// find where the difference is
-	line := 1
-	for i, ch := range normalizedBody {
-		if i >= len(formatted) || ch != formatted[i] {
-			break
-		}
-		if ch == '\n' {
-			line++
-		}
-	}
-	return caddyconfig.Warning{
-		File:    filename,
-		Line:    line,
-		Message: "Caddyfile input is not formatted; run the 'caddy fmt' command to fix inconsistencies",
-	}, true
-}
-
-// Unmarshaler is a type that can unmarshal
-// Caddyfile tokens to set itself up for a
-// JSON encoding. The goal of an unmarshaler
-// is not to set itself up for actual use,
-// but to set itself up for being marshaled
-// into JSON. Caddyfile-unmarshaled values
-// will not be used directly; they will be
-// encoded as JSON and then used from that.
-// Implementations must be able to support
-// multiple segments (instances of their
-// directive or batch of tokens); typically
-// this means wrapping all token logic in
-// a loop: `for d.Next() { ... }`.
-type Unmarshaler interface {
-	UnmarshalCaddyfile(d *Dispenser) error
-}
-
-// ServerType is a type that can evaluate a Caddyfile and set up a caddy config.
-type ServerType interface {
-	// Setup takes the server blocks which
-	// contain tokens, as well as options
-	// (e.g. CLI flags) and creates a Caddy
-	// config, along with any warnings or
-	// an error.
-	Setup([]ServerBlock, map[string]any) (*caddy.Config, []caddyconfig.Warning, error)
-}
-
-// UnmarshalModule instantiates a module with the given ID and invokes
-// UnmarshalCaddyfile on the new value using the immediate next segment
-// of d as input. In other words, d's next token should be the first
-// token of the module's Caddyfile input.
-//
-// This function is used when the next segment of Caddyfile tokens
-// belongs to another Caddy module. The returned value is often
-// type-asserted to the module's associated type for practical use
-// when setting up a config.
-func UnmarshalModule(d *Dispenser, moduleID string) (Unmarshaler, error) {
-	mod, err := caddy.GetModule(moduleID)
-	if err != nil {
-		return nil, d.Errf("getting module named '%s': %v", moduleID, err)
-	}
-	inst := mod.New()
-	unm, ok := inst.(Unmarshaler)
-	if !ok {
-		return nil, d.Errf("module %s is not a Caddyfile unmarshaler; is %T", mod.ID, inst)
-	}
-	err = unm.UnmarshalCaddyfile(d.NewFromNextSegment())
-	if err != nil {
-		return nil, err
-	}
-	return unm, nil
-}
-
-// Interface guard
-var _ caddyconfig.Adapter = (*Adapter)(nil)

caddyconfig/caddyfile/dispenser.go

@@ -1,517 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyfile
-
-import (
-	"errors"
-	"fmt"
-	"io"
-	"log"
-	"strconv"
-	"strings"
-)
-
-// Dispenser is a type that dispenses tokens, similarly to a lexer,
-// except that it can do so with some notion of structure. An empty
-// Dispenser is invalid; call NewDispenser to make a proper instance.
-type Dispenser struct {
-	tokens  []Token
-	cursor  int
-	nesting int
-}
-
-// NewDispenser returns a Dispenser filled with the given tokens.
-func NewDispenser(tokens []Token) *Dispenser {
-	return &Dispenser{
-		tokens: tokens,
-		cursor: -1,
-	}
-}
-
-// NewTestDispenser parses input into tokens and creates a new
-// Dispenser for test purposes only; any errors are fatal.
-func NewTestDispenser(input string) *Dispenser {
-	tokens, err := allTokens("Testfile", []byte(input))
-	if err != nil && err != io.EOF {
-		log.Fatalf("getting all tokens from input: %v", err)
-	}
-	return NewDispenser(tokens)
-}
-
-// Next loads the next token. Returns true if a token
-// was loaded; false otherwise. If false, all tokens
-// have been consumed.
-func (d *Dispenser) Next() bool {
-	if d.cursor < len(d.tokens)-1 {
-		d.cursor++
-		return true
-	}
-	return false
-}
-
-// Prev moves to the previous token. It does the inverse
-// of Next(), except this function may decrement the cursor
-// to -1 so that the next call to Next() points to the
-// first token; this allows dispensing to "start over". This
-// method returns true if the cursor ends up pointing to a
-// valid token.
-func (d *Dispenser) Prev() bool {
-	if d.cursor > -1 {
-		d.cursor--
-		return d.cursor > -1
-	}
-	return false
-}
-
-// NextArg loads the next token if it is on the same
-// line and if it is not a block opening (open curly
-// brace). Returns true if an argument token was
-// loaded; false otherwise. If false, all tokens on
-// the line have been consumed except for potentially
-// a block opening. It handles imported tokens
-// correctly.
-func (d *Dispenser) NextArg() bool {
-	if !d.nextOnSameLine() {
-		return false
-	}
-	if d.Val() == "{" {
-		// roll back; a block opening is not an argument
-		d.cursor--
-		return false
-	}
-	return true
-}
-
-// nextOnSameLine advances the cursor if the next
-// token is on the same line of the same file.
-func (d *Dispenser) nextOnSameLine() bool {
-	if d.cursor < 0 {
-		d.cursor++
-		return true
-	}
-	if d.cursor >= len(d.tokens) {
-		return false
-	}
-	if d.cursor < len(d.tokens)-1 &&
-		d.tokens[d.cursor].File == d.tokens[d.cursor+1].File &&
-		d.tokens[d.cursor].Line+d.numLineBreaks(d.cursor) == d.tokens[d.cursor+1].Line {
-		d.cursor++
-		return true
-	}
-	return false
-}
-
-// NextLine loads the next token only if it is not on the same
-// line as the current token, and returns true if a token was
-// loaded; false otherwise. If false, there is not another token
-// or it is on the same line. It handles imported tokens correctly.
-func (d *Dispenser) NextLine() bool {
-	if d.cursor < 0 {
-		d.cursor++
-		return true
-	}
-	if d.cursor >= len(d.tokens) {
-		return false
-	}
-	if d.cursor < len(d.tokens)-1 &&
-		(d.tokens[d.cursor].File != d.tokens[d.cursor+1].File ||
-			d.tokens[d.cursor].Line+d.numLineBreaks(d.cursor) < d.tokens[d.cursor+1].Line) {
-		d.cursor++
-		return true
-	}
-	return false
-}
-
-// NextBlock can be used as the condition of a for loop
-// to load the next token as long as it opens a block or
-// is already in a block nested more than initialNestingLevel.
-// In other words, a loop over NextBlock() will iterate
-// all tokens in the block assuming the next token is an
-// open curly brace, until the matching closing brace.
-// The open and closing brace tokens for the outer-most
-// block will be consumed internally and omitted from
-// the iteration.
-//
-// Proper use of this method looks like this:
-//
-//	for nesting := d.Nesting(); d.NextBlock(nesting); {
-//	}
-//
-// However, in simple cases where it is known that the
-// Dispenser is new and has not already traversed state
-// by a loop over NextBlock(), this will do:
-//
-//	for d.NextBlock(0) {
-//	}
-//
-// As with other token parsing logic, a loop over
-// NextBlock() should be contained within a loop over
-// Next(), as it is usually prudent to skip the initial
-// token.
-func (d *Dispenser) NextBlock(initialNestingLevel int) bool {
-	if d.nesting > initialNestingLevel {
-		if !d.Next() {
-			return false // should be EOF error
-		}
-		if d.Val() == "}" && !d.nextOnSameLine() {
-			d.nesting--
-		} else if d.Val() == "{" && !d.nextOnSameLine() {
-			d.nesting++
-		}
-		return d.nesting > initialNestingLevel
-	}
-	if !d.nextOnSameLine() { // block must open on same line
-		return false
-	}
-	if d.Val() != "{" {
-		d.cursor-- // roll back if not opening brace
-		return false
-	}
-	d.Next() // consume open curly brace
-	if d.Val() == "}" {
-		return false // open and then closed right away
-	}
-	d.nesting++
-	return true
-}
-
-// Nesting returns the current nesting level. Necessary
-// if using NextBlock()
-func (d *Dispenser) Nesting() int {
-	return d.nesting
-}
-
-// Val gets the text of the current token. If there is no token
-// loaded, it returns empty string.
-func (d *Dispenser) Val() string {
-	if d.cursor < 0 || d.cursor >= len(d.tokens) {
-		return ""
-	}
-	return d.tokens[d.cursor].Text
-}
-
-// ValRaw gets the raw text of the current token (including quotes).
-// If there is no token loaded, it returns empty string.
-func (d *Dispenser) ValRaw() string {
-	if d.cursor < 0 || d.cursor >= len(d.tokens) {
-		return ""
-	}
-	quote := d.tokens[d.cursor].wasQuoted
-	if quote > 0 {
-		return string(quote) + d.tokens[d.cursor].Text + string(quote) // string literal
-	}
-	return d.tokens[d.cursor].Text
-}
-
-// ScalarVal gets value of the current token, converted to the closest
-// scalar type. If there is no token loaded, it returns nil.
-func (d *Dispenser) ScalarVal() any {
-	if d.cursor < 0 || d.cursor >= len(d.tokens) {
-		return nil
-	}
-	quote := d.tokens[d.cursor].wasQuoted
-	text := d.tokens[d.cursor].Text
-
-	if quote > 0 {
-		return text // string literal
-	}
-	if num, err := strconv.Atoi(text); err == nil {
-		return num
-	}
-	if num, err := strconv.ParseFloat(text, 64); err == nil {
-		return num
-	}
-	if bool, err := strconv.ParseBool(text); err == nil {
-		return bool
-	}
-	return text
-}
-
-// Line gets the line number of the current token.
-// If there is no token loaded, it returns 0.
-func (d *Dispenser) Line() int {
-	if d.cursor < 0 || d.cursor >= len(d.tokens) {
-		return 0
-	}
-	return d.tokens[d.cursor].Line
-}
-
-// File gets the filename where the current token originated.
-func (d *Dispenser) File() string {
-	if d.cursor < 0 || d.cursor >= len(d.tokens) {
-		return ""
-	}
-	return d.tokens[d.cursor].File
-}
-
-// Args is a convenience function that loads the next arguments
-// (tokens on the same line) into an arbitrary number of strings
-// pointed to in targets. If there are not enough argument tokens
-// available to fill targets, false is returned and the remaining
-// targets are left unchanged. If all the targets are filled,
-// then true is returned.
-func (d *Dispenser) Args(targets ...*string) bool {
-	for i := 0; i < len(targets); i++ {
-		if !d.NextArg() {
-			return false
-		}
-		*targets[i] = d.Val()
-	}
-	return true
-}
-
-// AllArgs is like Args, but if there are more argument tokens
-// available than there are targets, false is returned. The
-// number of available argument tokens must match the number of
-// targets exactly to return true.
-func (d *Dispenser) AllArgs(targets ...*string) bool {
-	if !d.Args(targets...) {
-		return false
-	}
-	if d.NextArg() {
-		d.Prev()
-		return false
-	}
-	return true
-}
-
-// CountRemainingArgs counts the amount of remaining arguments
-// (tokens on the same line) without consuming the tokens.
-func (d *Dispenser) CountRemainingArgs() int {
-	count := 0
-	for d.NextArg() {
-		count++
-	}
-	for i := 0; i < count; i++ {
-		d.Prev()
-	}
-	return count
-}
-
-// RemainingArgs loads any more arguments (tokens on the same line)
-// into a slice and returns them. Open curly brace tokens also indicate
-// the end of arguments, and the curly brace is not included in
-// the return value nor is it loaded.
-func (d *Dispenser) RemainingArgs() []string {
-	var args []string
-	for d.NextArg() {
-		args = append(args, d.Val())
-	}
-	return args
-}
-
-// RemainingArgsRaw loads any more arguments (tokens on the same line,
-// retaining quotes) into a slice and returns them. Open curly brace
-// tokens also indicate the end of arguments, and the curly brace is
-// not included in the return value nor is it loaded.
-func (d *Dispenser) RemainingArgsRaw() []string {
-	var args []string
-	for d.NextArg() {
-		args = append(args, d.ValRaw())
-	}
-	return args
-}
-
-// NewFromNextSegment returns a new dispenser with a copy of
-// the tokens from the current token until the end of the
-// "directive" whether that be to the end of the line or
-// the end of a block that starts at the end of the line;
-// in other words, until the end of the segment.
-func (d *Dispenser) NewFromNextSegment() *Dispenser {
-	return NewDispenser(d.NextSegment())
-}
-
-// NextSegment returns a copy of the tokens from the current
-// token until the end of the line or block that starts at
-// the end of the line.
-func (d *Dispenser) NextSegment() Segment {
-	tkns := Segment{d.Token()}
-	for d.NextArg() {
-		tkns = append(tkns, d.Token())
-	}
-	var openedBlock bool
-	for nesting := d.Nesting(); d.NextBlock(nesting); {
-		if !openedBlock {
-			// because NextBlock() consumes the initial open
-			// curly brace, we rewind here to append it, since
-			// our case is special in that we want the new
-			// dispenser to have all the tokens including
-			// surrounding curly braces
-			d.Prev()
-			tkns = append(tkns, d.Token())
-			d.Next()
-			openedBlock = true
-		}
-		tkns = append(tkns, d.Token())
-	}
-	if openedBlock {
-		// include closing brace
-		tkns = append(tkns, d.Token())
-
-		// do not consume the closing curly brace; the
-		// next iteration of the enclosing loop will
-		// call Next() and consume it
-	}
-	return tkns
-}
-
-// Token returns the current token.
-func (d *Dispenser) Token() Token {
-	if d.cursor < 0 || d.cursor >= len(d.tokens) {
-		return Token{}
-	}
-	return d.tokens[d.cursor]
-}
-
-// Reset sets d's cursor to the beginning, as
-// if this was a new and unused dispenser.
-func (d *Dispenser) Reset() {
-	d.cursor = -1
-	d.nesting = 0
-}
-
-// ArgErr returns an argument error, meaning that another
-// argument was expected but not found. In other words,
-// a line break or open curly brace was encountered instead of
-// an argument.
-func (d *Dispenser) ArgErr() error {
-	if d.Val() == "{" {
-		return d.Err("Unexpected token '{', expecting argument")
-	}
-	return d.Errf("Wrong argument count or unexpected line ending after '%s'", d.Val())
-}
-
-// SyntaxErr creates a generic syntax error which explains what was
-// found and what was expected.
-func (d *Dispenser) SyntaxErr(expected string) error {
-	msg := fmt.Sprintf("%s:%d - Syntax error: Unexpected token '%s', expecting '%s'", d.File(), d.Line(), d.Val(), expected)
-	return errors.New(msg)
-}
-
-// EOFErr returns an error indicating that the dispenser reached
-// the end of the input when searching for the next token.
-func (d *Dispenser) EOFErr() error {
-	return d.Errf("Unexpected EOF")
-}
-
-// Err generates a custom parse-time error with a message of msg.
-func (d *Dispenser) Err(msg string) error {
-	return d.Errf(msg)
-}
-
-// Errf is like Err, but for formatted error messages
-func (d *Dispenser) Errf(format string, args ...any) error {
-	return d.WrapErr(fmt.Errorf(format, args...))
-}
-
-// WrapErr takes an existing error and adds the Caddyfile file and line number.
-func (d *Dispenser) WrapErr(err error) error {
-	return fmt.Errorf("%s:%d - Error during parsing: %w", d.File(), d.Line(), err)
-}
-
-// Delete deletes the current token and returns the updated slice
-// of tokens. The cursor is not advanced to the next token.
-// Because deletion modifies the underlying slice, this method
-// should only be called if you have access to the original slice
-// of tokens and/or are using the slice of tokens outside this
-// Dispenser instance. If you do not re-assign the slice with the
-// return value of this method, inconsistencies in the token
-// array will become apparent (or worse, hide from you like they
-// did me for 3 and a half freaking hours late one night).
-func (d *Dispenser) Delete() []Token {
-	if d.cursor >= 0 && d.cursor <= len(d.tokens)-1 {
-		d.tokens = append(d.tokens[:d.cursor], d.tokens[d.cursor+1:]...)
-		d.cursor--
-	}
-	return d.tokens
-}
-
-// numLineBreaks counts how many line breaks are in the token
-// value given by the token index tknIdx. It returns 0 if the
-// token does not exist or there are no line breaks.
-func (d *Dispenser) numLineBreaks(tknIdx int) int {
-	if tknIdx < 0 || tknIdx >= len(d.tokens) {
-		return 0
-	}
-	return strings.Count(d.tokens[tknIdx].Text, "\n")
-}
-
-// isNewLine determines whether the current token is on a different
-// line (higher line number) than the previous token. It handles imported
-// tokens correctly. If there isn't a previous token, it returns true.
-func (d *Dispenser) isNewLine() bool {
-	if d.cursor < 1 {
-		return true
-	}
-	if d.cursor > len(d.tokens)-1 {
-		return false
-	}
-
-	prev := d.tokens[d.cursor-1]
-	curr := d.tokens[d.cursor]
-
-	// If the previous token is from a different file,
-	// we can assume it's from a different line
-	if prev.File != curr.File {
-		return true
-	}
-
-	// The previous token may contain line breaks if
-	// it was quoted and spanned multiple lines. e.g:
-	//
-	// dir "foo
-	//   bar
-	//   baz"
-	prevLineBreaks := d.numLineBreaks(d.cursor - 1)
-
-	// If the previous token (incl line breaks) ends
-	// on a line earlier than the current token,
-	// then the current token is on a new line
-	return prev.Line+prevLineBreaks < curr.Line
-}
-
-// isNextOnNewLine determines whether the current token is on a different
-// line (higher line number) than the next token. It handles imported
-// tokens correctly. If there isn't a next token, it returns true.
-func (d *Dispenser) isNextOnNewLine() bool {
-	if d.cursor < 0 {
-		return false
-	}
-	if d.cursor >= len(d.tokens)-1 {
-		return true
-	}
-
-	curr := d.tokens[d.cursor]
-	next := d.tokens[d.cursor+1]
-
-	// If the next token is from a different file,
-	// we can assume it's from a different line
-	if curr.File != next.File {
-		return true
-	}
-
-	// The current token may contain line breaks if
-	// it was quoted and spanned multiple lines. e.g:
-	//
-	// dir "foo
-	//   bar
-	//   baz"
-	currLineBreaks := d.numLineBreaks(d.cursor)
-
-	// If the current token (incl line breaks) ends
-	// on a line earlier than the next token,
-	// then the next token is on a new line
-	return curr.Line+currLineBreaks < next.Line
-}

caddyconfig/caddyfile/formatter.go

@@ -1,219 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyfile
-
-import (
-	"bytes"
-	"io"
-	"unicode"
-)
-
-// Format formats the input Caddyfile to a standard, nice-looking
-// appearance. It works by reading each rune of the input and taking
-// control over all the bracing and whitespace that is written; otherwise,
-// words, comments, placeholders, and escaped characters are all treated
-// literally and written as they appear in the input.
-func Format(input []byte) []byte {
-	input = bytes.TrimSpace(input)
-
-	out := new(bytes.Buffer)
-	rdr := bytes.NewReader(input)
-
-	var (
-		last rune // the last character that was written to the result
-
-		space           = true // whether current/previous character was whitespace (beginning of input counts as space)
-		beginningOfLine = true // whether we are at beginning of line
-
-		openBrace        bool // whether current word/token is or started with open curly brace
-		openBraceWritten bool // if openBrace, whether that brace was written or not
-		openBraceSpace   bool // whether there was a non-newline space before open brace
-
-		newLines int // count of newlines consumed
-
-		comment bool // whether we're in a comment
-		quoted  bool // whether we're in a quoted segment
-		escaped bool // whether current char is escaped
-
-		nesting int // indentation level
-	)
-
-	write := func(ch rune) {
-		out.WriteRune(ch)
-		last = ch
-	}
-
-	indent := func() {
-		for tabs := nesting; tabs > 0; tabs-- {
-			write('\t')
-		}
-	}
-
-	nextLine := func() {
-		write('\n')
-		beginningOfLine = true
-	}
-
-	for {
-		ch, _, err := rdr.ReadRune()
-		if err != nil {
-			if err == io.EOF {
-				break
-			}
-			panic(err)
-		}
-
-		if comment {
-			if ch == '\n' {
-				comment = false
-				space = true
-				nextLine()
-				continue
-			} else {
-				write(ch)
-				continue
-			}
-		}
-
-		if !escaped && ch == '\\' {
-			if space {
-				write(' ')
-				space = false
-			}
-			write(ch)
-			escaped = true
-			continue
-		}
-
-		if escaped {
-			write(ch)
-			escaped = false
-			continue
-		}
-
-		if quoted {
-			if ch == '"' {
-				quoted = false
-			}
-			write(ch)
-			continue
-		}
-
-		if space && ch == '"' {
-			quoted = true
-		}
-
-		if unicode.IsSpace(ch) {
-			space = true
-			if ch == '\n' {
-				newLines++
-			}
-			continue
-		}
-		spacePrior := space
-		space = false
-
-		//////////////////////////////////////////////////////////
-		// I find it helpful to think of the formatting loop in two
-		// main sections; by the time we reach this point, we
-		// know we are in a "regular" part of the file: we know
-		// the character is not a space, not in a literal segment
-		// like a comment or quoted, it's not escaped, etc.
-		//////////////////////////////////////////////////////////
-
-		if ch == '#' {
-			comment = true
-		}
-
-		if openBrace && spacePrior && !openBraceWritten {
-			if nesting == 0 && last == '}' {
-				nextLine()
-				nextLine()
-			}
-
-			openBrace = false
-			if beginningOfLine {
-				indent()
-			} else if !openBraceSpace {
-				write(' ')
-			}
-			write('{')
-			openBraceWritten = true
-			nextLine()
-			newLines = 0
-			// prevent infinite nesting from ridiculous inputs (issue #4169)
-			if nesting < 10 {
-				nesting++
-			}
-		}
-
-		switch {
-		case ch == '{':
-			openBrace = true
-			openBraceWritten = false
-			openBraceSpace = spacePrior && !beginningOfLine
-			if openBraceSpace {
-				write(' ')
-			}
-			continue
-
-		case ch == '}' && (spacePrior || !openBrace):
-			if last != '\n' {
-				nextLine()
-			}
-			if nesting > 0 {
-				nesting--
-			}
-			indent()
-			write('}')
-			newLines = 0
-			continue
-		}
-
-		if newLines > 2 {
-			newLines = 2
-		}
-		for i := 0; i < newLines; i++ {
-			nextLine()
-		}
-		newLines = 0
-		if beginningOfLine {
-			indent()
-		}
-		if nesting == 0 && last == '}' && beginningOfLine {
-			nextLine()
-			nextLine()
-		}
-
-		if !beginningOfLine && spacePrior {
-			write(' ')
-		}
-
-		if openBrace && !openBraceWritten {
-			write('{')
-			openBraceWritten = true
-		}
-		write(ch)
-
-		beginningOfLine = false
-	}
-
-	// the Caddyfile does not need any leading or trailing spaces, but...
-	trimmedResult := bytes.TrimSpace(out.Bytes())
-
-	// ...Caddyfiles should, however, end with a newline because
-	// newlines are significant to the syntax of the file
-	return append(trimmedResult, '\n')
-}

caddyconfig/caddyfile/formatter_fuzz.go

@@ -1,27 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-//go:build gofuzz
-
-package caddyfile
-
-import "bytes"
-
-func FuzzFormat(input []byte) int {
-	formatted := Format(input)
-	if bytes.Equal(formatted, Format(formatted)) {
-		return 1
-	}
-	return 0
-}

caddyconfig/caddyfile/formatter_test.go

@@ -1,381 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyfile
-
-import (
-	"strings"
-	"testing"
-)
-
-func TestFormatter(t *testing.T) {
-	for i, tc := range []struct {
-		description string
-		input       string
-		expect      string
-	}{
-		{
-			description: "very simple",
-			input: `abc   def
-	g hi jkl
-mn`,
-			expect: `abc def
-g hi jkl
-mn`,
-		},
-		{
-			description: "basic indentation, line breaks, and nesting",
-			input: `  a
-b
-
-	c {
-		d
-}
-
-e { f
-}
-
-
-
-g {
-h {
-i
-}
-}
-
-j { k {
-l
-}
-}
-
-m {
-	n { o
-	}
-	p { q r
-s }
-}
-
-	{
-{ t
-		u
-
-	v
-
-w
-}
-}`,
-			expect: `a
-b
-
-c {
-	d
-}
-
-e {
-	f
-}
-
-g {
-	h {
-		i
-	}
-}
-
-j {
-	k {
-		l
-	}
-}
-
-m {
-	n {
-		o
-	}
-	p {
-		q r
-		s
-	}
-}
-
-{
-	{
-		t
-		u
-
-		v
-
-		w
-	}
-}`,
-		},
-		{
-			description: "block spacing",
-			input: `a{
-	b
-}
-
-c{ d
-}`,
-			expect: `a {
-	b
-}
-
-c {
-	d
-}`,
-		},
-		{
-			description: "advanced spacing",
-			input: `abc {
-	def
-}ghi{
-	jkl mno
-pqr}`,
-			expect: `abc {
-	def
-}
-
-ghi {
-	jkl mno
-	pqr
-}`,
-		},
-		{
-			description: "env var placeholders",
-			input: `{$A}
-
-b {
-{$C}
-}
-
-d { {$E}
-}
-
-{ {$F}
-}
-`,
-			expect: `{$A}
-
-b {
-	{$C}
-}
-
-d {
-	{$E}
-}
-
-{
-	{$F}
-}`,
-		},
-		{
-			description: "env var placeholders with port",
-			input:       `:{$PORT}`,
-			expect:      `:{$PORT}`,
-		},
-		{
-			description: "comments",
-			input: `#a "\n"
-
- #b {
-	c
-}
-
-d {
-e#f
-# g
-}
-
-h { # i
-}`,
-			expect: `#a "\n"
-
-#b {
-c
-}
-
-d {
-	e#f
-	# g
-}
-
-h {
-	# i
-}`,
-		},
-		{
-			description: "quotes and escaping",
-			input: `"a \"b\" "#c
-	d
-
-e {
-"f"
-}
-
-g { "h"
-}
-
-i {
-	"foo
-bar"
-}
-
-j {
-"\"k\" l m"
-}`,
-			expect: `"a \"b\" "#c
-d
-
-e {
-	"f"
-}
-
-g {
-	"h"
-}
-
-i {
-	"foo
-bar"
-}
-
-j {
-	"\"k\" l m"
-}`,
-		},
-		{
-			description: "bad nesting (too many open)",
-			input: `a
-{
-	{
-}`,
-			expect: `a {
-	{
-	}
-`,
-		},
-		{
-			description: "bad nesting (too many close)",
-			input: `a
-{
-	{
-}}}`,
-			expect: `a {
-	{
-	}
-}
-}
-`,
-		},
-		{
-			description: "json",
-			input: `foo
-bar      "{\"key\":34}"
-`,
-			expect: `foo
-bar "{\"key\":34}"`,
-		},
-		{
-			description: "escaping after spaces",
-			input:       `foo \"literal\"`,
-			expect:      `foo \"literal\"`,
-		},
-		{
-			description: "simple placeholders as standalone tokens",
-			input:       `foo {bar}`,
-			expect:      `foo {bar}`,
-		},
-		{
-			description: "simple placeholders within tokens",
-			input:       `foo{bar} foo{bar}baz`,
-			expect:      `foo{bar} foo{bar}baz`,
-		},
-		{
-			description: "placeholders and malformed braces",
-			input:       `foo{bar} foo{ bar}baz`,
-			expect: `foo{bar} foo {
-	bar
-}
-
-baz`,
-		},
-		{
-			description: "hash within string is not a comment",
-			input:       `redir / /some/#/path`,
-			expect:      `redir / /some/#/path`,
-		},
-		{
-			description: "brace does not fold into comment above",
-			input: `# comment
-{
-	foo
-}`,
-			expect: `# comment
-{
-	foo
-}`,
-		},
-		{
-			description: "matthewpi/vscode-caddyfile-support#13",
-			input: `{
-	email {$ACMEEMAIL}
-	#debug
-}
-
-block {
-}
-`,
-			expect: `{
-	email {$ACMEEMAIL}
-	#debug
-}
-
-block {
-}
-`,
-		},
-		{
-			description: "matthewpi/vscode-caddyfile-support#13 - bad formatting",
-			input: `{
-	email {$ACMEEMAIL}
-	#debug
-	}
-
-	block {
-	}
-`,
-			expect: `{
-	email {$ACMEEMAIL}
-	#debug
-}
-
-block {
-}
-`,
-		},
-	} {
-		// the formatter should output a trailing newline,
-		// even if the tests aren't written to expect that
-		if !strings.HasSuffix(tc.expect, "\n") {
-			tc.expect += "\n"
-		}
-
-		actual := Format([]byte(tc.input))
-
-		if string(actual) != tc.expect {
-			t.Errorf("\n[TEST %d: %s]\n====== EXPECTED ======\n%s\n====== ACTUAL ======\n%s^^^^^^^^^^^^^^^^^^^^^",
-				i, tc.description, string(tc.expect), string(actual))
-		}
-	}
-}

caddyconfig/caddyfile/importgraph.go

@@ -1,127 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyfile
-
-import (
-	"fmt"
-)
-
-type adjacency map[string][]string
-
-type importGraph struct {
-	nodes map[string]bool
-	edges adjacency
-}
-
-func (i *importGraph) addNode(name string) {
-	if i.nodes == nil {
-		i.nodes = make(map[string]bool)
-	}
-	if _, exists := i.nodes[name]; exists {
-		return
-	}
-	i.nodes[name] = true
-}
-func (i *importGraph) addNodes(names []string) {
-	for _, name := range names {
-		i.addNode(name)
-	}
-}
-
-func (i *importGraph) removeNode(name string) {
-	delete(i.nodes, name)
-}
-func (i *importGraph) removeNodes(names []string) {
-	for _, name := range names {
-		i.removeNode(name)
-	}
-}
-
-func (i *importGraph) addEdge(from, to string) error {
-	if !i.exists(from) || !i.exists(to) {
-		return fmt.Errorf("one of the nodes does not exist")
-	}
-
-	if i.willCycle(to, from) {
-		return fmt.Errorf("a cycle of imports exists between %s and %s", from, to)
-	}
-
-	if i.areConnected(from, to) {
-		// if connected, there's nothing to do
-		return nil
-	}
-
-	if i.nodes == nil {
-		i.nodes = make(map[string]bool)
-	}
-	if i.edges == nil {
-		i.edges = make(adjacency)
-	}
-
-	i.edges[from] = append(i.edges[from], to)
-	return nil
-}
-func (i *importGraph) addEdges(from string, tos []string) error {
-	for _, to := range tos {
-		err := i.addEdge(from, to)
-		if err != nil {
-			return err
-		}
-	}
-	return nil
-}
-
-func (i *importGraph) areConnected(from, to string) bool {
-	al, ok := i.edges[from]
-	if !ok {
-		return false
-	}
-	for _, v := range al {
-		if v == to {
-			return true
-		}
-	}
-	return false
-}
-
-func (i *importGraph) willCycle(from, to string) bool {
-	collector := make(map[string]bool)
-
-	var visit func(string)
-	visit = func(start string) {
-		if !collector[start] {
-			collector[start] = true
-			for _, v := range i.edges[start] {
-				visit(v)
-			}
-		}
-	}
-
-	for _, v := range i.edges[from] {
-		visit(v)
-	}
-	for k := range collector {
-		if to == k {
-			return true
-		}
-	}
-
-	return false
-}
-
-func (i *importGraph) exists(key string) bool {
-	_, exists := i.nodes[key]
-	return exists
-}

caddyconfig/caddyfile/lexer.go

@@ -1,197 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyfile
-
-import (
-	"bufio"
-	"bytes"
-	"io"
-	"unicode"
-)
-
-type (
-	// lexer is a utility which can get values, token by
-	// token, from a Reader. A token is a word, and tokens
-	// are separated by whitespace. A word can be enclosed
-	// in quotes if it contains whitespace.
-	lexer struct {
-		reader       *bufio.Reader
-		token        Token
-		line         int
-		skippedLines int
-	}
-
-	// Token represents a single parsable unit.
-	Token struct {
-		File        string
-		Line        int
-		Text        string
-		wasQuoted   rune // enclosing quote character, if any
-		inSnippet   bool
-		snippetName string
-	}
-)
-
-// load prepares the lexer to scan an input for tokens.
-// It discards any leading byte order mark.
-func (l *lexer) load(input io.Reader) error {
-	l.reader = bufio.NewReader(input)
-	l.line = 1
-
-	// discard byte order mark, if present
-	firstCh, _, err := l.reader.ReadRune()
-	if err != nil {
-		return err
-	}
-	if firstCh != 0xFEFF {
-		err := l.reader.UnreadRune()
-		if err != nil {
-			return err
-		}
-	}
-
-	return nil
-}
-
-// next loads the next token into the lexer.
-// A token is delimited by whitespace, unless
-// the token starts with a quotes character (")
-// in which case the token goes until the closing
-// quotes (the enclosing quotes are not included).
-// Inside quoted strings, quotes may be escaped
-// with a preceding \ character. No other chars
-// may be escaped. The rest of the line is skipped
-// if a "#" character is read in. Returns true if
-// a token was loaded; false otherwise.
-func (l *lexer) next() bool {
-	var val []rune
-	var comment, quoted, btQuoted, escaped bool
-
-	makeToken := func(quoted rune) bool {
-		l.token.Text = string(val)
-		l.token.wasQuoted = quoted
-		return true
-	}
-
-	for {
-		ch, _, err := l.reader.ReadRune()
-		if err != nil {
-			if len(val) > 0 {
-				return makeToken(0)
-			}
-			if err == io.EOF {
-				return false
-			}
-			panic(err)
-		}
-
-		if !escaped && !btQuoted && ch == '\\' {
-			escaped = true
-			continue
-		}
-
-		if quoted || btQuoted {
-			if quoted && escaped {
-				// all is literal in quoted area,
-				// so only escape quotes
-				if ch != '"' {
-					val = append(val, '\\')
-				}
-				escaped = false
-			} else {
-				if quoted && ch == '"' {
-					return makeToken('"')
-				}
-				if btQuoted && ch == '`' {
-					return makeToken('`')
-				}
-			}
-			if ch == '\n' {
-				l.line += 1 + l.skippedLines
-				l.skippedLines = 0
-			}
-			val = append(val, ch)
-			continue
-		}
-
-		if unicode.IsSpace(ch) {
-			if ch == '\r' {
-				continue
-			}
-			if ch == '\n' {
-				if escaped {
-					l.skippedLines++
-					escaped = false
-				} else {
-					l.line += 1 + l.skippedLines
-					l.skippedLines = 0
-				}
-				comment = false
-			}
-			if len(val) > 0 {
-				return makeToken(0)
-			}
-			continue
-		}
-
-		if ch == '#' && len(val) == 0 {
-			comment = true
-		}
-		if comment {
-			continue
-		}
-
-		if len(val) == 0 {
-			l.token = Token{Line: l.line}
-			if ch == '"' {
-				quoted = true
-				continue
-			}
-			if ch == '`' {
-				btQuoted = true
-				continue
-			}
-		}
-
-		if escaped {
-			val = append(val, '\\')
-			escaped = false
-		}
-
-		val = append(val, ch)
-	}
-}
-
-// Tokenize takes bytes as input and lexes it into
-// a list of tokens that can be parsed as a Caddyfile.
-// Also takes a filename to fill the token's File as
-// the source of the tokens, which is important to
-// determine relative paths for `import` directives.
-func Tokenize(input []byte, filename string) ([]Token, error) {
-	l := lexer{}
-	if err := l.load(bytes.NewReader(input)); err != nil {
-		return nil, err
-	}
-	var tokens []Token
-	for l.next() {
-		l.token.File = filename
-		tokens = append(tokens, l.token)
-	}
-	return tokens, nil
-}
-
-func (t Token) Quoted() bool {
-	return t.wasQuoted > 0
-}

caddyconfig/caddyfile/lexer_fuzz.go

@@ -1,28 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-//go:build gofuzz
-
-package caddyfile
-
-func FuzzTokenize(input []byte) int {
-	tokens, err := Tokenize(input, "Caddyfile")
-	if err != nil {
-		return 0
-	}
-	if len(tokens) == 0 {
-		return -1
-	}
-	return 1
-}

caddyconfig/caddyfile/lexer_test.go

@@ -1,280 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyfile
-
-import (
-	"testing"
-)
-
-type lexerTestCase struct {
-	input    []byte
-	expected []Token
-}
-
-func TestLexer(t *testing.T) {
-	testCases := []lexerTestCase{
-		{
-			input: []byte(`host:123`),
-			expected: []Token{
-				{Line: 1, Text: "host:123"},
-			},
-		},
-		{
-			input: []byte(`host:123
-
-					directive`),
-			expected: []Token{
-				{Line: 1, Text: "host:123"},
-				{Line: 3, Text: "directive"},
-			},
-		},
-		{
-			input: []byte(`host:123 {
-						directive
-					}`),
-			expected: []Token{
-				{Line: 1, Text: "host:123"},
-				{Line: 1, Text: "{"},
-				{Line: 2, Text: "directive"},
-				{Line: 3, Text: "}"},
-			},
-		},
-		{
-			input: []byte(`host:123 { directive }`),
-			expected: []Token{
-				{Line: 1, Text: "host:123"},
-				{Line: 1, Text: "{"},
-				{Line: 1, Text: "directive"},
-				{Line: 1, Text: "}"},
-			},
-		},
-		{
-			input: []byte(`host:123 {
-						#comment
-						directive
-						# comment
-						foobar # another comment
-					}`),
-			expected: []Token{
-				{Line: 1, Text: "host:123"},
-				{Line: 1, Text: "{"},
-				{Line: 3, Text: "directive"},
-				{Line: 5, Text: "foobar"},
-				{Line: 6, Text: "}"},
-			},
-		},
-		{
-			input: []byte(`host:123 {
-						# hash inside string is not a comment
-						redir / /some/#/path
-					}`),
-			expected: []Token{
-				{Line: 1, Text: "host:123"},
-				{Line: 1, Text: "{"},
-				{Line: 3, Text: "redir"},
-				{Line: 3, Text: "/"},
-				{Line: 3, Text: "/some/#/path"},
-				{Line: 4, Text: "}"},
-			},
-		},
-		{
-			input: []byte("# comment at beginning of file\n# comment at beginning of line\nhost:123"),
-			expected: []Token{
-				{Line: 3, Text: "host:123"},
-			},
-		},
-		{
-			input: []byte(`a "quoted value" b
-					foobar`),
-			expected: []Token{
-				{Line: 1, Text: "a"},
-				{Line: 1, Text: "quoted value"},
-				{Line: 1, Text: "b"},
-				{Line: 2, Text: "foobar"},
-			},
-		},
-		{
-			input: []byte(`A "quoted \"value\" inside" B`),
-			expected: []Token{
-				{Line: 1, Text: "A"},
-				{Line: 1, Text: `quoted "value" inside`},
-				{Line: 1, Text: "B"},
-			},
-		},
-		{
-			input: []byte("An escaped \"newline\\\ninside\" quotes"),
-			expected: []Token{
-				{Line: 1, Text: "An"},
-				{Line: 1, Text: "escaped"},
-				{Line: 1, Text: "newline\\\ninside"},
-				{Line: 2, Text: "quotes"},
-			},
-		},
-		{
-			input: []byte("An escaped newline\\\noutside quotes"),
-			expected: []Token{
-				{Line: 1, Text: "An"},
-				{Line: 1, Text: "escaped"},
-				{Line: 1, Text: "newline"},
-				{Line: 1, Text: "outside"},
-				{Line: 1, Text: "quotes"},
-			},
-		},
-		{
-			input: []byte("line1\\\nescaped\nline2\nline3"),
-			expected: []Token{
-				{Line: 1, Text: "line1"},
-				{Line: 1, Text: "escaped"},
-				{Line: 3, Text: "line2"},
-				{Line: 4, Text: "line3"},
-			},
-		},
-		{
-			input: []byte("line1\\\nescaped1\\\nescaped2\nline4\nline5"),
-			expected: []Token{
-				{Line: 1, Text: "line1"},
-				{Line: 1, Text: "escaped1"},
-				{Line: 1, Text: "escaped2"},
-				{Line: 4, Text: "line4"},
-				{Line: 5, Text: "line5"},
-			},
-		},
-		{
-			input: []byte(`"unescapable\ in quotes"`),
-			expected: []Token{
-				{Line: 1, Text: `unescapable\ in quotes`},
-			},
-		},
-		{
-			input: []byte(`"don't\escape"`),
-			expected: []Token{
-				{Line: 1, Text: `don't\escape`},
-			},
-		},
-		{
-			input: []byte(`"don't\\escape"`),
-			expected: []Token{
-				{Line: 1, Text: `don't\\escape`},
-			},
-		},
-		{
-			input: []byte(`un\escapable`),
-			expected: []Token{
-				{Line: 1, Text: `un\escapable`},
-			},
-		},
-		{
-			input: []byte(`A "quoted value with line
-					break inside" {
-						foobar
-					}`),
-			expected: []Token{
-				{Line: 1, Text: "A"},
-				{Line: 1, Text: "quoted value with line\n\t\t\t\t\tbreak inside"},
-				{Line: 2, Text: "{"},
-				{Line: 3, Text: "foobar"},
-				{Line: 4, Text: "}"},
-			},
-		},
-		{
-			input: []byte(`"C:\php\php-cgi.exe"`),
-			expected: []Token{
-				{Line: 1, Text: `C:\php\php-cgi.exe`},
-			},
-		},
-		{
-			input: []byte(`empty "" string`),
-			expected: []Token{
-				{Line: 1, Text: `empty`},
-				{Line: 1, Text: ``},
-				{Line: 1, Text: `string`},
-			},
-		},
-		{
-			input: []byte("skip those\r\nCR characters"),
-			expected: []Token{
-				{Line: 1, Text: "skip"},
-				{Line: 1, Text: "those"},
-				{Line: 2, Text: "CR"},
-				{Line: 2, Text: "characters"},
-			},
-		},
-		{
-			input: []byte("\xEF\xBB\xBF:8080"), // test with leading byte order mark
-			expected: []Token{
-				{Line: 1, Text: ":8080"},
-			},
-		},
-		{
-			input: []byte("simple `backtick quoted` string"),
-			expected: []Token{
-				{Line: 1, Text: `simple`},
-				{Line: 1, Text: `backtick quoted`},
-				{Line: 1, Text: `string`},
-			},
-		},
-		{
-			input: []byte("multiline `backtick\nquoted\n` string"),
-			expected: []Token{
-				{Line: 1, Text: `multiline`},
-				{Line: 1, Text: "backtick\nquoted\n"},
-				{Line: 3, Text: `string`},
-			},
-		},
-		{
-			input: []byte("nested `\"quotes inside\" backticks` string"),
-			expected: []Token{
-				{Line: 1, Text: `nested`},
-				{Line: 1, Text: `"quotes inside" backticks`},
-				{Line: 1, Text: `string`},
-			},
-		},
-		{
-			input: []byte("reverse-nested \"`backticks` inside\" quotes"),
-			expected: []Token{
-				{Line: 1, Text: `reverse-nested`},
-				{Line: 1, Text: "`backticks` inside"},
-				{Line: 1, Text: `quotes`},
-			},
-		},
-	}
-
-	for i, testCase := range testCases {
-		actual, err := Tokenize(testCase.input, "")
-		if err != nil {
-			t.Errorf("%v", err)
-		}
-		lexerCompare(t, i, testCase.expected, actual)
-	}
-}
-
-func lexerCompare(t *testing.T, n int, expected, actual []Token) {
-	if len(expected) != len(actual) {
-		t.Errorf("Test case %d: expected %d token(s) but got %d", n, len(expected), len(actual))
-	}
-
-	for i := 0; i < len(actual) && i < len(expected); i++ {
-		if actual[i].Line != expected[i].Line {
-			t.Errorf("Test case %d token %d ('%s'): expected line %d but was line %d",
-				n, i, expected[i].Text, expected[i].Line, actual[i].Line)
-			break
-		}
-		if actual[i].Text != expected[i].Text {
-			t.Errorf("Test case %d token %d: expected text '%s' but was '%s'",
-				n, i, expected[i].Text, actual[i].Text)
-			break
-		}
-	}
-}

caddyconfig/caddyfile/parse.go

@@ -1,630 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyfile
-
-import (
-	"bytes"
-	"fmt"
-	"io"
-	"os"
-	"path/filepath"
-	"strconv"
-	"strings"
-
-	"github.com/caddyserver/caddy/v2"
-	"go.uber.org/zap"
-)
-
-// Parse parses the input just enough to group tokens, in
-// order, by server block. No further parsing is performed.
-// Server blocks are returned in the order in which they appear.
-// Directives that do not appear in validDirectives will cause
-// an error. If you do not want to check for valid directives,
-// pass in nil instead.
-//
-// Environment variables in {$ENVIRONMENT_VARIABLE} notation
-// will be replaced before parsing begins.
-func Parse(filename string, input []byte) ([]ServerBlock, error) {
-	// unfortunately, we must copy the input because parsing must
-	// remain a read-only operation, but we have to expand environment
-	// variables before we parse, which changes the underlying array (#4422)
-	inputCopy := make([]byte, len(input))
-	copy(inputCopy, input)
-
-	tokens, err := allTokens(filename, inputCopy)
-	if err != nil {
-		return nil, err
-	}
-	p := parser{
-		Dispenser: NewDispenser(tokens),
-		importGraph: importGraph{
-			nodes: make(map[string]bool),
-			edges: make(adjacency),
-		},
-	}
-	return p.parseAll()
-}
-
-// allTokens lexes the entire input, but does not parse it.
-// It returns all the tokens from the input, unstructured
-// and in order. It may mutate input as it expands env vars.
-func allTokens(filename string, input []byte) ([]Token, error) {
-	inputCopy, err := replaceEnvVars(input)
-	if err != nil {
-		return nil, err
-	}
-	tokens, err := Tokenize(inputCopy, filename)
-	if err != nil {
-		return nil, err
-	}
-	return tokens, nil
-}
-
-// replaceEnvVars replaces all occurrences of environment variables.
-// It mutates the underlying array and returns the updated slice.
-func replaceEnvVars(input []byte) ([]byte, error) {
-	var offset int
-	for {
-		begin := bytes.Index(input[offset:], spanOpen)
-		if begin < 0 {
-			break
-		}
-		begin += offset // make beginning relative to input, not offset
-		end := bytes.Index(input[begin+len(spanOpen):], spanClose)
-		if end < 0 {
-			break
-		}
-		end += begin + len(spanOpen) // make end relative to input, not begin
-
-		// get the name; if there is no name, skip it
-		envString := input[begin+len(spanOpen) : end]
-		if len(envString) == 0 {
-			offset = end + len(spanClose)
-			continue
-		}
-
-		// split the string into a key and an optional default
-		envParts := strings.SplitN(string(envString), envVarDefaultDelimiter, 2)
-
-		// do a lookup for the env var, replace with the default if not found
-		envVarValue, found := os.LookupEnv(envParts[0])
-		if !found && len(envParts) == 2 {
-			envVarValue = envParts[1]
-		}
-
-		// get the value of the environment variable
-		// note that this causes one-level deep chaining
-		envVarBytes := []byte(envVarValue)
-
-		// splice in the value
-		input = append(input[:begin],
-			append(envVarBytes, input[end+len(spanClose):]...)...)
-
-		// continue at the end of the replacement
-		offset = begin + len(envVarBytes)
-	}
-	return input, nil
-}
-
-type parser struct {
-	*Dispenser
-	block           ServerBlock // current server block being parsed
-	eof             bool        // if we encounter a valid EOF in a hard place
-	definedSnippets map[string][]Token
-	nesting         int
-	importGraph     importGraph
-}
-
-func (p *parser) parseAll() ([]ServerBlock, error) {
-	var blocks []ServerBlock
-
-	for p.Next() {
-		err := p.parseOne()
-		if err != nil {
-			return blocks, err
-		}
-		if len(p.block.Keys) > 0 || len(p.block.Segments) > 0 {
-			blocks = append(blocks, p.block)
-		}
-		if p.nesting > 0 {
-			return blocks, p.EOFErr()
-		}
-	}
-
-	return blocks, nil
-}
-
-func (p *parser) parseOne() error {
-	p.block = ServerBlock{}
-	return p.begin()
-}
-
-func (p *parser) begin() error {
-	if len(p.tokens) == 0 {
-		return nil
-	}
-
-	err := p.addresses()
-
-	if err != nil {
-		return err
-	}
-
-	if p.eof {
-		// this happens if the Caddyfile consists of only
-		// a line of addresses and nothing else
-		return nil
-	}
-
-	if ok, name := p.isSnippet(); ok {
-		if p.definedSnippets == nil {
-			p.definedSnippets = map[string][]Token{}
-		}
-		if _, found := p.definedSnippets[name]; found {
-			return p.Errf("redeclaration of previously declared snippet %s", name)
-		}
-		// consume all tokens til matched close brace
-		tokens, err := p.snippetTokens()
-		if err != nil {
-			return err
-		}
-		// Just as we need to track which file the token comes from, we need to
-		// keep track of which snippets do the tokens come from. This is helpful
-		// in tracking import cycles across files/snippets by namespacing them. Without
-		// this we end up with false-positives in cycle-detection.
-		for k, v := range tokens {
-			v.inSnippet = true
-			v.snippetName = name
-			tokens[k] = v
-		}
-		p.definedSnippets[name] = tokens
-		// empty block keys so we don't save this block as a real server.
-		p.block.Keys = nil
-		return nil
-	}
-
-	return p.blockContents()
-}
-
-func (p *parser) addresses() error {
-	var expectingAnother bool
-
-	for {
-		tkn := p.Val()
-
-		// special case: import directive replaces tokens during parse-time
-		if tkn == "import" && p.isNewLine() {
-			err := p.doImport()
-			if err != nil {
-				return err
-			}
-			continue
-		}
-
-		// Open brace definitely indicates end of addresses
-		if tkn == "{" {
-			if expectingAnother {
-				return p.Errf("Expected another address but had '%s' - check for extra comma", tkn)
-			}
-			// Mark this server block as being defined with braces.
-			// This is used to provide a better error message when
-			// the user may have tried to define two server blocks
-			// without having used braces, which are required in
-			// that case.
-			p.block.HasBraces = true
-			break
-		}
-
-		// Users commonly forget to place a space between the address and the '{'
-		if strings.HasSuffix(tkn, "{") {
-			return p.Errf("Site addresses cannot end with a curly brace: '%s' - put a space between the token and the brace", tkn)
-		}
-
-		if tkn != "" { // empty token possible if user typed ""
-			// Trailing comma indicates another address will follow, which
-			// may possibly be on the next line
-			if tkn[len(tkn)-1] == ',' {
-				tkn = tkn[:len(tkn)-1]
-				expectingAnother = true
-			} else {
-				expectingAnother = false // but we may still see another one on this line
-			}
-
-			// If there's a comma here, it's probably because they didn't use a space
-			// between their two domains, e.g. "foo.com,bar.com", which would not be
-			// parsed as two separate site addresses.
-			if strings.Contains(tkn, ",") {
-				return p.Errf("Site addresses cannot contain a comma ',': '%s' - put a space after the comma to separate site addresses", tkn)
-			}
-
-			p.block.Keys = append(p.block.Keys, tkn)
-		}
-
-		// Advance token and possibly break out of loop or return error
-		hasNext := p.Next()
-		if expectingAnother && !hasNext {
-			return p.EOFErr()
-		}
-		if !hasNext {
-			p.eof = true
-			break // EOF
-		}
-		if !expectingAnother && p.isNewLine() {
-			break
-		}
-	}
-
-	return nil
-}
-
-func (p *parser) blockContents() error {
-	errOpenCurlyBrace := p.openCurlyBrace()
-	if errOpenCurlyBrace != nil {
-		// single-server configs don't need curly braces
-		p.cursor--
-	}
-
-	err := p.directives()
-	if err != nil {
-		return err
-	}
-
-	// only look for close curly brace if there was an opening
-	if errOpenCurlyBrace == nil {
-		err = p.closeCurlyBrace()
-		if err != nil {
-			return err
-		}
-	}
-
-	return nil
-}
-
-// directives parses through all the lines for directives
-// and it expects the next token to be the first
-// directive. It goes until EOF or closing curly brace
-// which ends the server block.
-func (p *parser) directives() error {
-	for p.Next() {
-		// end of server block
-		if p.Val() == "}" {
-			// p.nesting has already been decremented
-			break
-		}
-
-		// special case: import directive replaces tokens during parse-time
-		if p.Val() == "import" {
-			err := p.doImport()
-			if err != nil {
-				return err
-			}
-			p.cursor-- // cursor is advanced when we continue, so roll back one more
-			continue
-		}
-
-		// normal case: parse a directive as a new segment
-		// (a "segment" is a line which starts with a directive
-		// and which ends at the end of the line or at the end of
-		// the block that is opened at the end of the line)
-		if err := p.directive(); err != nil {
-			return err
-		}
-	}
-
-	return nil
-}
-
-// doImport swaps out the import directive and its argument
-// (a total of 2 tokens) with the tokens in the specified file
-// or globbing pattern. When the function returns, the cursor
-// is on the token before where the import directive was. In
-// other words, call Next() to access the first token that was
-// imported.
-func (p *parser) doImport() error {
-	// syntax checks
-	if !p.NextArg() {
-		return p.ArgErr()
-	}
-	importPattern := p.Val()
-	if importPattern == "" {
-		return p.Err("Import requires a non-empty filepath")
-	}
-
-	// grab remaining args as placeholder replacements
-	args := p.RemainingArgs()
-
-	// add args to the replacer
-	repl := caddy.NewEmptyReplacer()
-	for index, arg := range args {
-		repl.Set("args."+strconv.Itoa(index), arg)
-	}
-
-	// splice out the import directive and its arguments
-	// (2 tokens, plus the length of args)
-	tokensBefore := p.tokens[:p.cursor-1-len(args)]
-	tokensAfter := p.tokens[p.cursor+1:]
-	var importedTokens []Token
-	var nodes []string
-
-	// first check snippets. That is a simple, non-recursive replacement
-	if p.definedSnippets != nil && p.definedSnippets[importPattern] != nil {
-		importedTokens = p.definedSnippets[importPattern]
-		if len(importedTokens) > 0 {
-			// just grab the first one
-			nodes = append(nodes, fmt.Sprintf("%s:%s", importedTokens[0].File, importedTokens[0].snippetName))
-		}
-	} else {
-		// make path relative to the file of the _token_ being processed rather
-		// than current working directory (issue #867) and then use glob to get
-		// list of matching filenames
-		absFile, err := filepath.Abs(p.Dispenser.File())
-		if err != nil {
-			return p.Errf("Failed to get absolute path of file: %s: %v", p.Dispenser.File(), err)
-		}
-
-		var matches []string
-		var globPattern string
-		if !filepath.IsAbs(importPattern) {
-			globPattern = filepath.Join(filepath.Dir(absFile), importPattern)
-		} else {
-			globPattern = importPattern
-		}
-		if strings.Count(globPattern, "*") > 1 || strings.Count(globPattern, "?") > 1 ||
-			(strings.Contains(globPattern, "[") && strings.Contains(globPattern, "]")) {
-			// See issue #2096 - a pattern with many glob expansions can hang for too long
-			return p.Errf("Glob pattern may only contain one wildcard (*), but has others: %s", globPattern)
-		}
-		matches, err = filepath.Glob(globPattern)
-
-		if err != nil {
-			return p.Errf("Failed to use import pattern %s: %v", importPattern, err)
-		}
-		if len(matches) == 0 {
-			if strings.ContainsAny(globPattern, "*?[]") {
-				caddy.Log().Warn("No files matching import glob pattern", zap.String("pattern", importPattern))
-			} else {
-				return p.Errf("File to import not found: %s", importPattern)
-			}
-		}
-
-		// collect all the imported tokens
-		for _, importFile := range matches {
-			newTokens, err := p.doSingleImport(importFile)
-			if err != nil {
-				return err
-			}
-			importedTokens = append(importedTokens, newTokens...)
-		}
-		nodes = matches
-	}
-
-	nodeName := p.File()
-	if p.Token().inSnippet {
-		nodeName += fmt.Sprintf(":%s", p.Token().snippetName)
-	}
-	p.importGraph.addNode(nodeName)
-	p.importGraph.addNodes(nodes)
-	if err := p.importGraph.addEdges(nodeName, nodes); err != nil {
-		p.importGraph.removeNodes(nodes)
-		return err
-	}
-
-	// copy the tokens so we don't overwrite p.definedSnippets
-	tokensCopy := make([]Token, len(importedTokens))
-	copy(tokensCopy, importedTokens)
-
-	// run the argument replacer on the tokens
-	for index, token := range tokensCopy {
-		token.Text = repl.ReplaceKnown(token.Text, "")
-		tokensCopy[index] = token
-	}
-
-	// splice the imported tokens in the place of the import statement
-	// and rewind cursor so Next() will land on first imported token
-	p.tokens = append(tokensBefore, append(tokensCopy, tokensAfter...)...)
-	p.cursor -= len(args) + 1
-
-	return nil
-}
-
-// doSingleImport lexes the individual file at importFile and returns
-// its tokens or an error, if any.
-func (p *parser) doSingleImport(importFile string) ([]Token, error) {
-	file, err := os.Open(importFile)
-	if err != nil {
-		return nil, p.Errf("Could not import %s: %v", importFile, err)
-	}
-	defer file.Close()
-
-	if info, err := file.Stat(); err != nil {
-		return nil, p.Errf("Could not import %s: %v", importFile, err)
-	} else if info.IsDir() {
-		return nil, p.Errf("Could not import %s: is a directory", importFile)
-	}
-
-	input, err := io.ReadAll(file)
-	if err != nil {
-		return nil, p.Errf("Could not read imported file %s: %v", importFile, err)
-	}
-
-	importedTokens, err := allTokens(importFile, input)
-	if err != nil {
-		return nil, p.Errf("Could not read tokens while importing %s: %v", importFile, err)
-	}
-
-	// Tack the file path onto these tokens so errors show the imported file's name
-	// (we use full, absolute path to avoid bugs: issue #1892)
-	filename, err := filepath.Abs(importFile)
-	if err != nil {
-		return nil, p.Errf("Failed to get absolute path of file: %s: %v", importFile, err)
-	}
-	for i := 0; i < len(importedTokens); i++ {
-		importedTokens[i].File = filename
-	}
-
-	return importedTokens, nil
-}
-
-// directive collects tokens until the directive's scope
-// closes (either end of line or end of curly brace block).
-// It expects the currently-loaded token to be a directive
-// (or } that ends a server block). The collected tokens
-// are loaded into the current server block for later use
-// by directive setup functions.
-func (p *parser) directive() error {
-
-	// a segment is a list of tokens associated with this directive
-	var segment Segment
-
-	// the directive itself is appended as a relevant token
-	segment = append(segment, p.Token())
-
-	for p.Next() {
-		if p.Val() == "{" {
-			p.nesting++
-			if !p.isNextOnNewLine() && p.Token().wasQuoted == 0 {
-				return p.Err("Unexpected next token after '{' on same line")
-			}
-		} else if p.Val() == "{}" {
-			if p.isNextOnNewLine() && p.Token().wasQuoted == 0 {
-				return p.Err("Unexpected '{}' at end of line")
-			}
-		} else if p.isNewLine() && p.nesting == 0 {
-			p.cursor-- // read too far
-			break
-		} else if p.Val() == "}" && p.nesting > 0 {
-			p.nesting--
-		} else if p.Val() == "}" && p.nesting == 0 {
-			return p.Err("Unexpected '}' because no matching opening brace")
-		} else if p.Val() == "import" && p.isNewLine() {
-			if err := p.doImport(); err != nil {
-				return err
-			}
-			p.cursor-- // cursor is advanced when we continue, so roll back one more
-			continue
-		}
-
-		segment = append(segment, p.Token())
-	}
-
-	p.block.Segments = append(p.block.Segments, segment)
-
-	if p.nesting > 0 {
-		return p.EOFErr()
-	}
-
-	return nil
-}
-
-// openCurlyBrace expects the current token to be an
-// opening curly brace. This acts like an assertion
-// because it returns an error if the token is not
-// a opening curly brace. It does NOT advance the token.
-func (p *parser) openCurlyBrace() error {
-	if p.Val() != "{" {
-		return p.SyntaxErr("{")
-	}
-	return nil
-}
-
-// closeCurlyBrace expects the current token to be
-// a closing curly brace. This acts like an assertion
-// because it returns an error if the token is not
-// a closing curly brace. It does NOT advance the token.
-func (p *parser) closeCurlyBrace() error {
-	if p.Val() != "}" {
-		return p.SyntaxErr("}")
-	}
-	return nil
-}
-
-func (p *parser) isSnippet() (bool, string) {
-	keys := p.block.Keys
-	// A snippet block is a single key with parens. Nothing else qualifies.
-	if len(keys) == 1 && strings.HasPrefix(keys[0], "(") && strings.HasSuffix(keys[0], ")") {
-		return true, strings.TrimSuffix(keys[0][1:], ")")
-	}
-	return false, ""
-}
-
-// read and store everything in a block for later replay.
-func (p *parser) snippetTokens() ([]Token, error) {
-	// snippet must have curlies.
-	err := p.openCurlyBrace()
-	if err != nil {
-		return nil, err
-	}
-	nesting := 1 // count our own nesting in snippets
-	tokens := []Token{}
-	for p.Next() {
-		if p.Val() == "}" {
-			nesting--
-			if nesting == 0 {
-				break
-			}
-		}
-		if p.Val() == "{" {
-			nesting++
-		}
-		tokens = append(tokens, p.tokens[p.cursor])
-	}
-	// make sure we're matched up
-	if nesting != 0 {
-		return nil, p.SyntaxErr("}")
-	}
-	return tokens, nil
-}
-
-// ServerBlock associates any number of keys from the
-// head of the server block with tokens, which are
-// grouped by segments.
-type ServerBlock struct {
-	HasBraces bool
-	Keys      []string
-	Segments  []Segment
-}
-
-// DispenseDirective returns a dispenser that contains
-// all the tokens in the server block.
-func (sb ServerBlock) DispenseDirective(dir string) *Dispenser {
-	var tokens []Token
-	for _, seg := range sb.Segments {
-		if len(seg) > 0 && seg[0].Text == dir {
-			tokens = append(tokens, seg...)
-		}
-	}
-	return NewDispenser(tokens)
-}
-
-// Segment is a list of tokens which begins with a directive
-// and ends at the end of the directive (either at the end of
-// the line, or at the end of a block it opens).
-type Segment []Token
-
-// Directive returns the directive name for the segment.
-// The directive name is the text of the first token.
-func (s Segment) Directive() string {
-	if len(s) > 0 {
-		return s[0].Text
-	}
-	return ""
-}
-
-// spanOpen and spanClose are used to bound spans that
-// contain the name of an environment variable.
-var (
-	spanOpen, spanClose    = []byte{'{', '$'}, []byte{'}'}
-	envVarDefaultDelimiter = ":"
-)

caddyconfig/caddyfile/parse_test.go

@@ -1,722 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyfile
-
-import (
-	"bytes"
-	"os"
-	"path/filepath"
-	"testing"
-)
-
-func TestAllTokens(t *testing.T) {
-	input := []byte("a b c\nd e")
-	expected := []string{"a", "b", "c", "d", "e"}
-	tokens, err := allTokens("TestAllTokens", input)
-
-	if err != nil {
-		t.Fatalf("Expected no error, got %v", err)
-	}
-	if len(tokens) != len(expected) {
-		t.Fatalf("Expected %d tokens, got %d", len(expected), len(tokens))
-	}
-
-	for i, val := range expected {
-		if tokens[i].Text != val {
-			t.Errorf("Token %d should be '%s' but was '%s'", i, val, tokens[i].Text)
-		}
-	}
-}
-
-func TestParseOneAndImport(t *testing.T) {
-	testParseOne := func(input string) (ServerBlock, error) {
-		p := testParser(input)
-		p.Next() // parseOne doesn't call Next() to start, so we must
-		err := p.parseOne()
-		return p.block, err
-	}
-
-	for i, test := range []struct {
-		input     string
-		shouldErr bool
-		keys      []string
-		numTokens []int // number of tokens to expect in each segment
-	}{
-		{`localhost`, false, []string{
-			"localhost",
-		}, []int{}},
-
-		{`localhost
-		  dir1`, false, []string{
-			"localhost",
-		}, []int{1}},
-
-		{`localhost:1234
-		  dir1 foo bar`, false, []string{
-			"localhost:1234",
-		}, []int{3},
-		},
-
-		{`localhost {
-		    dir1
-		  }`, false, []string{
-			"localhost",
-		}, []int{1}},
-
-		{`localhost:1234 {
-		    dir1 foo bar
-		    dir2
-		  }`, false, []string{
-			"localhost:1234",
-		}, []int{3, 1}},
-
-		{`http://localhost https://localhost
-		  dir1 foo bar`, false, []string{
-			"http://localhost",
-			"https://localhost",
-		}, []int{3}},
-
-		{`http://localhost https://localhost {
-		    dir1 foo bar
-		  }`, false, []string{
-			"http://localhost",
-			"https://localhost",
-		}, []int{3}},
-
-		{`http://localhost, https://localhost {
-		    dir1 foo bar
-		  }`, false, []string{
-			"http://localhost",
-			"https://localhost",
-		}, []int{3}},
-
-		{`http://localhost, {
-		  }`, true, []string{
-			"http://localhost",
-		}, []int{}},
-
-		{`host1:80, http://host2.com
-		  dir1 foo bar
-		  dir2 baz`, false, []string{
-			"host1:80",
-			"http://host2.com",
-		}, []int{3, 2}},
-
-		{`http://host1.com,
-		  http://host2.com,
-		  https://host3.com`, false, []string{
-			"http://host1.com",
-			"http://host2.com",
-			"https://host3.com",
-		}, []int{}},
-
-		{`http://host1.com:1234, https://host2.com
-		  dir1 foo {
-		    bar baz
-		  }
-		  dir2`, false, []string{
-			"http://host1.com:1234",
-			"https://host2.com",
-		}, []int{6, 1}},
-
-		{`127.0.0.1
-		  dir1 {
-		    bar baz
-		  }
-		  dir2 {
-		    foo bar
-		  }`, false, []string{
-			"127.0.0.1",
-		}, []int{5, 5}},
-
-		{`localhost
-		  dir1 {
-		    foo`, true, []string{
-			"localhost",
-		}, []int{3}},
-
-		{`localhost
-		  dir1 {
-		  }`, false, []string{
-			"localhost",
-		}, []int{3}},
-
-		{`localhost
-		  dir1 {
-		  } }`, true, []string{
-			"localhost",
-		}, []int{}},
-
-		{`localhost{
-		    dir1
-		  }`, true, []string{}, []int{}},
-
-		{`localhost
-		  dir1 {
-		    nested {
-		      foo
-		    }
-		  }
-		  dir2 foo bar`, false, []string{
-			"localhost",
-		}, []int{7, 3}},
-
-		{``, false, []string{}, []int{}},
-
-		{`localhost
-		  dir1 arg1
-		  import testdata/import_test1.txt`, false, []string{
-			"localhost",
-		}, []int{2, 3, 1}},
-
-		{`import testdata/import_test2.txt`, false, []string{
-			"host1",
-		}, []int{1, 2}},
-
-		{`import testdata/not_found.txt`, true, []string{}, []int{}},
-
-		{`""`, false, []string{}, []int{}},
-
-		{``, false, []string{}, []int{}},
-
-		// Unexpected next token after '{' on same line
-		{`localhost
-		  dir1 { a b }`, true, []string{"localhost"}, []int{}},
-		// Workaround with quotes
-		{`localhost
-		  dir1 "{" a b "}"`, false, []string{"localhost"}, []int{5}},
-
-		// Unexpected '{}' at end of line
-		{`localhost
-		  dir1 {}`, true, []string{"localhost"}, []int{}},
-		// Workaround with quotes
-		{`localhost
-		  dir1 "{}"`, false, []string{"localhost"}, []int{2}},
-
-		// import with args
-		{`import testdata/import_args0.txt a`, false, []string{"a"}, []int{}},
-		{`import testdata/import_args1.txt a b`, false, []string{"a", "b"}, []int{}},
-		{`import testdata/import_args*.txt a b`, false, []string{"a"}, []int{2}},
-
-		// test cases found by fuzzing!
-		{`import }{$"`, true, []string{}, []int{}},
-		{`import /*/*.txt`, true, []string{}, []int{}},
-		{`import /???/?*?o`, true, []string{}, []int{}},
-		{`import /??`, true, []string{}, []int{}},
-		{`import /[a-z]`, true, []string{}, []int{}},
-		{`import {$}`, true, []string{}, []int{}},
-		{`import {%}`, true, []string{}, []int{}},
-		{`import {$$}`, true, []string{}, []int{}},
-		{`import {%%}`, true, []string{}, []int{}},
-	} {
-		result, err := testParseOne(test.input)
-
-		if test.shouldErr && err == nil {
-			t.Errorf("Test %d: Expected an error, but didn't get one", i)
-		}
-		if !test.shouldErr && err != nil {
-			t.Errorf("Test %d: Expected no error, but got: %v", i, err)
-		}
-
-		// t.Logf("%+v\n", result)
-		if len(result.Keys) != len(test.keys) {
-			t.Errorf("Test %d: Expected %d keys, got %d",
-				i, len(test.keys), len(result.Keys))
-			continue
-		}
-		for j, addr := range result.Keys {
-			if addr != test.keys[j] {
-				t.Errorf("Test %d, key %d: Expected '%s', but was '%s'",
-					i, j, test.keys[j], addr)
-			}
-		}
-
-		if len(result.Segments) != len(test.numTokens) {
-			t.Errorf("Test %d: Expected %d segments, had %d",
-				i, len(test.numTokens), len(result.Segments))
-			continue
-		}
-
-		for j, seg := range result.Segments {
-			if len(seg) != test.numTokens[j] {
-				t.Errorf("Test %d, segment %d: Expected %d tokens, counted %d",
-					i, j, test.numTokens[j], len(seg))
-				continue
-			}
-		}
-	}
-}
-
-func TestRecursiveImport(t *testing.T) {
-	testParseOne := func(input string) (ServerBlock, error) {
-		p := testParser(input)
-		p.Next() // parseOne doesn't call Next() to start, so we must
-		err := p.parseOne()
-		return p.block, err
-	}
-
-	isExpected := func(got ServerBlock) bool {
-		if len(got.Keys) != 1 || got.Keys[0] != "localhost" {
-			t.Errorf("got keys unexpected: expect localhost, got %v", got.Keys)
-			return false
-		}
-		if len(got.Segments) != 2 {
-			t.Errorf("got wrong number of segments: expect 2, got %d", len(got.Segments))
-			return false
-		}
-		if len(got.Segments[0]) != 1 || len(got.Segments[1]) != 2 {
-			t.Errorf("got unexpected tokens: %v", got.Segments)
-			return false
-		}
-		return true
-	}
-
-	recursiveFile1, err := filepath.Abs("testdata/recursive_import_test1")
-	if err != nil {
-		t.Fatal(err)
-	}
-	recursiveFile2, err := filepath.Abs("testdata/recursive_import_test2")
-	if err != nil {
-		t.Fatal(err)
-	}
-
-	// test relative recursive import
-	err = os.WriteFile(recursiveFile1, []byte(
-		`localhost
-		dir1
-		import recursive_import_test2`), 0644)
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.Remove(recursiveFile1)
-
-	err = os.WriteFile(recursiveFile2, []byte("dir2 1"), 0644)
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.Remove(recursiveFile2)
-
-	// import absolute path
-	result, err := testParseOne("import " + recursiveFile1)
-	if err != nil {
-		t.Fatal(err)
-	}
-	if !isExpected(result) {
-		t.Error("absolute+relative import failed")
-	}
-
-	// import relative path
-	result, err = testParseOne("import testdata/recursive_import_test1")
-	if err != nil {
-		t.Fatal(err)
-	}
-	if !isExpected(result) {
-		t.Error("relative+relative import failed")
-	}
-
-	// test absolute recursive import
-	err = os.WriteFile(recursiveFile1, []byte(
-		`localhost
-		dir1
-		import `+recursiveFile2), 0644)
-	if err != nil {
-		t.Fatal(err)
-	}
-
-	// import absolute path
-	result, err = testParseOne("import " + recursiveFile1)
-	if err != nil {
-		t.Fatal(err)
-	}
-	if !isExpected(result) {
-		t.Error("absolute+absolute import failed")
-	}
-
-	// import relative path
-	result, err = testParseOne("import testdata/recursive_import_test1")
-	if err != nil {
-		t.Fatal(err)
-	}
-	if !isExpected(result) {
-		t.Error("relative+absolute import failed")
-	}
-}
-
-func TestDirectiveImport(t *testing.T) {
-	testParseOne := func(input string) (ServerBlock, error) {
-		p := testParser(input)
-		p.Next() // parseOne doesn't call Next() to start, so we must
-		err := p.parseOne()
-		return p.block, err
-	}
-
-	isExpected := func(got ServerBlock) bool {
-		if len(got.Keys) != 1 || got.Keys[0] != "localhost" {
-			t.Errorf("got keys unexpected: expect localhost, got %v", got.Keys)
-			return false
-		}
-		if len(got.Segments) != 2 {
-			t.Errorf("got wrong number of segments: expect 2, got %d", len(got.Segments))
-			return false
-		}
-		if len(got.Segments[0]) != 1 || len(got.Segments[1]) != 8 {
-			t.Errorf("got unexpected tokens: %v", got.Segments)
-			return false
-		}
-		return true
-	}
-
-	directiveFile, err := filepath.Abs("testdata/directive_import_test")
-	if err != nil {
-		t.Fatal(err)
-	}
-
-	err = os.WriteFile(directiveFile, []byte(`prop1 1
-	prop2 2`), 0644)
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer os.Remove(directiveFile)
-
-	// import from existing file
-	result, err := testParseOne(`localhost
-	dir1
-	proxy {
-		import testdata/directive_import_test
-		transparent
-	}`)
-	if err != nil {
-		t.Fatal(err)
-	}
-	if !isExpected(result) {
-		t.Error("directive import failed")
-	}
-
-	// import from nonexistent file
-	_, err = testParseOne(`localhost
-	dir1
-	proxy {
-		import testdata/nonexistent_file
-		transparent
-	}`)
-	if err == nil {
-		t.Fatal("expected error when importing a nonexistent file")
-	}
-}
-
-func TestParseAll(t *testing.T) {
-	for i, test := range []struct {
-		input     string
-		shouldErr bool
-		keys      [][]string // keys per server block, in order
-	}{
-		{`localhost`, false, [][]string{
-			{"localhost"},
-		}},
-
-		{`localhost:1234`, false, [][]string{
-			{"localhost:1234"},
-		}},
-
-		{`localhost:1234 {
-		  }
-		  localhost:2015 {
-		  }`, false, [][]string{
-			{"localhost:1234"},
-			{"localhost:2015"},
-		}},
-
-		{`localhost:1234, http://host2`, false, [][]string{
-			{"localhost:1234", "http://host2"},
-		}},
-
-		{`localhost:1234, http://host2,`, true, [][]string{}},
-
-		{`http://host1.com, http://host2.com {
-		  }
-		  https://host3.com, https://host4.com {
-		  }`, false, [][]string{
-			{"http://host1.com", "http://host2.com"},
-			{"https://host3.com", "https://host4.com"},
-		}},
-
-		{`import testdata/import_glob*.txt`, false, [][]string{
-			{"glob0.host0"},
-			{"glob0.host1"},
-			{"glob1.host0"},
-			{"glob2.host0"},
-		}},
-
-		{`import notfound/*`, false, [][]string{}},        // glob needn't error with no matches
-		{`import notfound/file.conf`, true, [][]string{}}, // but a specific file should
-
-		// recursive self-import
-		{`import testdata/import_recursive0.txt`, true, [][]string{}},
-		{`import testdata/import_recursive3.txt
-		import testdata/import_recursive1.txt`, true, [][]string{}},
-
-		// cyclic imports
-		{`(A) {
-			import A
-		}
-		:80
-		import A
-		`, true, [][]string{}},
-		{`(A) {
-			import B
-		}
-		(B) {
-			import A
-		}
-		:80
-		import A
-		`, true, [][]string{}},
-	} {
-		p := testParser(test.input)
-		blocks, err := p.parseAll()
-
-		if test.shouldErr && err == nil {
-			t.Errorf("Test %d: Expected an error, but didn't get one", i)
-		}
-		if !test.shouldErr && err != nil {
-			t.Errorf("Test %d: Expected no error, but got: %v", i, err)
-		}
-
-		if len(blocks) != len(test.keys) {
-			t.Errorf("Test %d: Expected %d server blocks, got %d",
-				i, len(test.keys), len(blocks))
-			continue
-		}
-		for j, block := range blocks {
-			if len(block.Keys) != len(test.keys[j]) {
-				t.Errorf("Test %d: Expected %d keys in block %d, got %d",
-					i, len(test.keys[j]), j, len(block.Keys))
-				continue
-			}
-			for k, addr := range block.Keys {
-				if addr != test.keys[j][k] {
-					t.Errorf("Test %d, block %d, key %d: Expected '%s', but got '%s'",
-						i, j, k, test.keys[j][k], addr)
-				}
-			}
-		}
-	}
-}
-
-func TestEnvironmentReplacement(t *testing.T) {
-	os.Setenv("FOOBAR", "foobar")
-	os.Setenv("CHAINED", "$FOOBAR")
-
-	for i, test := range []struct {
-		input  string
-		expect string
-	}{
-		{
-			input:  "",
-			expect: "",
-		},
-		{
-			input:  "foo",
-			expect: "foo",
-		},
-		{
-			input:  "{$NOT_SET}",
-			expect: "",
-		},
-		{
-			input:  "foo{$NOT_SET}bar",
-			expect: "foobar",
-		},
-		{
-			input:  "{$FOOBAR}",
-			expect: "foobar",
-		},
-		{
-			input:  "foo {$FOOBAR} bar",
-			expect: "foo foobar bar",
-		},
-		{
-			input:  "foo{$FOOBAR}bar",
-			expect: "foofoobarbar",
-		},
-		{
-			input:  "foo\n{$FOOBAR}\nbar",
-			expect: "foo\nfoobar\nbar",
-		},
-		{
-			input:  "{$FOOBAR} {$FOOBAR}",
-			expect: "foobar foobar",
-		},
-		{
-			input:  "{$FOOBAR}{$FOOBAR}",
-			expect: "foobarfoobar",
-		},
-		{
-			input:  "{$CHAINED}",
-			expect: "$FOOBAR", // should not chain env expands
-		},
-		{
-			input:  "{$FOO:default}",
-			expect: "default",
-		},
-		{
-			input:  "foo{$BAR:bar}baz",
-			expect: "foobarbaz",
-		},
-		{
-			input:  "foo{$BAR:$FOOBAR}baz",
-			expect: "foo$FOOBARbaz", // should not chain env expands
-		},
-		{
-			input:  "{$FOOBAR",
-			expect: "{$FOOBAR",
-		},
-		{
-			input:  "{$LONGER_NAME $FOOBAR}",
-			expect: "",
-		},
-		{
-			input:  "{$}",
-			expect: "{$}",
-		},
-		{
-			input:  "{$$}",
-			expect: "",
-		},
-		{
-			input:  "{$",
-			expect: "{$",
-		},
-		{
-			input:  "}{$",
-			expect: "}{$",
-		},
-	} {
-		actual, err := replaceEnvVars([]byte(test.input))
-		if err != nil {
-			t.Fatal(err)
-		}
-		if !bytes.Equal(actual, []byte(test.expect)) {
-			t.Errorf("Test %d: Expected: '%s' but got '%s'", i, test.expect, actual)
-		}
-	}
-}
-
-func TestSnippets(t *testing.T) {
-	p := testParser(`
-		(common) {
-			gzip foo
-			errors stderr
-		}
-		http://example.com {
-			import common
-		}
-	`)
-	blocks, err := p.parseAll()
-	if err != nil {
-		t.Fatal(err)
-	}
-	if len(blocks) != 1 {
-		t.Fatalf("Expect exactly one server block. Got %d.", len(blocks))
-	}
-	if actual, expected := blocks[0].Keys[0], "http://example.com"; expected != actual {
-		t.Errorf("Expected server name to be '%s' but was '%s'", expected, actual)
-	}
-	if len(blocks[0].Segments) != 2 {
-		t.Fatalf("Server block should have tokens from import, got: %+v", blocks[0])
-	}
-	if actual, expected := blocks[0].Segments[0][0].Text, "gzip"; expected != actual {
-		t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual)
-	}
-	if actual, expected := blocks[0].Segments[1][1].Text, "stderr"; expected != actual {
-		t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual)
-	}
-}
-
-func writeStringToTempFileOrDie(t *testing.T, str string) (pathToFile string) {
-	file, err := os.CreateTemp("", t.Name())
-	if err != nil {
-		panic(err) // get a stack trace so we know where this was called from.
-	}
-	if _, err := file.WriteString(str); err != nil {
-		panic(err)
-	}
-	if err := file.Close(); err != nil {
-		panic(err)
-	}
-	return file.Name()
-}
-
-func TestImportedFilesIgnoreNonDirectiveImportTokens(t *testing.T) {
-	fileName := writeStringToTempFileOrDie(t, `
-		http://example.com {
-			# This isn't an import directive, it's just an arg with value 'import'
-			basicauth / import password
-		}
-	`)
-	// Parse the root file that imports the other one.
-	p := testParser(`import ` + fileName)
-	blocks, err := p.parseAll()
-	if err != nil {
-		t.Fatal(err)
-	}
-	auth := blocks[0].Segments[0]
-	line := auth[0].Text + " " + auth[1].Text + " " + auth[2].Text + " " + auth[3].Text
-	if line != "basicauth / import password" {
-		// Previously, it would be changed to:
-		//   basicauth / import /path/to/test/dir/password
-		// referencing a file that (probably) doesn't exist and changing the
-		// password!
-		t.Errorf("Expected basicauth tokens to be 'basicauth / import password' but got %#q", line)
-	}
-}
-
-func TestSnippetAcrossMultipleFiles(t *testing.T) {
-	// Make the derived Caddyfile that expects (common) to be defined.
-	fileName := writeStringToTempFileOrDie(t, `
-		http://example.com {
-			import common
-		}
-	`)
-
-	// Parse the root file that defines (common) and then imports the other one.
-	p := testParser(`
-		(common) {
-			gzip foo
-		}
-		import ` + fileName + `
-	`)
-
-	blocks, err := p.parseAll()
-	if err != nil {
-		t.Fatal(err)
-	}
-	if len(blocks) != 1 {
-		t.Fatalf("Expect exactly one server block. Got %d.", len(blocks))
-	}
-	if actual, expected := blocks[0].Keys[0], "http://example.com"; expected != actual {
-		t.Errorf("Expected server name to be '%s' but was '%s'", expected, actual)
-	}
-	if len(blocks[0].Segments) != 1 {
-		t.Fatalf("Server block should have tokens from import")
-	}
-	if actual, expected := blocks[0].Segments[0][0].Text, "gzip"; expected != actual {
-		t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual)
-	}
-}
-
-func testParser(input string) parser {
-	return parser{Dispenser: NewTestDispenser(input)}
-}

caddyconfig/caddyfile/testdata/import_args0.txt

@@ -1 +0,0 @@
-{args.0}

caddyconfig/caddyfile/testdata/import_args1.txt

@@ -1 +0,0 @@
-{args.0} {args.1}

caddyconfig/caddyfile/testdata/import_recursive0.txt

@@ -1 +0,0 @@
-import import_recursive0.txt

caddyconfig/caddyfile/testdata/import_recursive1.txt

@@ -1 +0,0 @@
-import import_recursive2.txt

caddyconfig/caddyfile/testdata/import_recursive2.txt

@@ -1 +0,0 @@
-import import_recursive3.txt

caddyconfig/caddyfile/testdata/import_recursive3.txt

@@ -1 +0,0 @@
-import import_recursive1.txt

caddyconfig/configadapters.go

@@ -1,138 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyconfig
-
-import (
-	"encoding/json"
-	"fmt"
-
-	"github.com/caddyserver/caddy/v2"
-)
-
-// Adapter is a type which can adapt a configuration to Caddy JSON.
-// It returns the results and any warnings, or an error.
-type Adapter interface {
-	Adapt(body []byte, options map[string]any) ([]byte, []Warning, error)
-}
-
-// Warning represents a warning or notice related to conversion.
-type Warning struct {
-	File      string `json:"file,omitempty"`
-	Line      int    `json:"line,omitempty"`
-	Directive string `json:"directive,omitempty"`
-	Message   string `json:"message,omitempty"`
-}
-
-func (w Warning) String() string {
-	var directive string
-	if w.Directive != "" {
-		directive = fmt.Sprintf(" (%s)", w.Directive)
-	}
-	return fmt.Sprintf("%s:%d%s: %s", w.File, w.Line, directive, w.Message)
-}
-
-// JSON encodes val as JSON, returning it as a json.RawMessage. Any
-// marshaling errors (which are highly unlikely with correct code)
-// are converted to warnings. This is convenient when filling config
-// structs that require a json.RawMessage, without having to worry
-// about errors.
-func JSON(val any, warnings *[]Warning) json.RawMessage {
-	b, err := json.Marshal(val)
-	if err != nil {
-		if warnings != nil {
-			*warnings = append(*warnings, Warning{Message: err.Error()})
-		}
-		return nil
-	}
-	return b
-}
-
-// JSONModuleObject is like JSON(), except it marshals val into a JSON object
-// with an added key named fieldName with the value fieldVal. This is useful
-// for encoding module values where the module name has to be described within
-// the object by a certain key; for example, `"handler": "file_server"` for a
-// file server HTTP handler (fieldName="handler" and fieldVal="file_server").
-// The val parameter must encode into a map[string]any (i.e. it must be
-// a struct or map). Any errors are converted into warnings.
-func JSONModuleObject(val any, fieldName, fieldVal string, warnings *[]Warning) json.RawMessage {
-	// encode to a JSON object first
-	enc, err := json.Marshal(val)
-	if err != nil {
-		if warnings != nil {
-			*warnings = append(*warnings, Warning{Message: err.Error()})
-		}
-		return nil
-	}
-
-	// then decode the object
-	var tmp map[string]any
-	err = json.Unmarshal(enc, &tmp)
-	if err != nil {
-		if warnings != nil {
-			*warnings = append(*warnings, Warning{Message: err.Error()})
-		}
-		return nil
-	}
-
-	// so we can easily add the module's field with its appointed value
-	tmp[fieldName] = fieldVal
-
-	// then re-marshal as JSON
-	result, err := json.Marshal(tmp)
-	if err != nil {
-		if warnings != nil {
-			*warnings = append(*warnings, Warning{Message: err.Error()})
-		}
-		return nil
-	}
-
-	return result
-}
-
-// RegisterAdapter registers a config adapter with the given name.
-// This should usually be done at init-time. It panics if the
-// adapter cannot be registered successfully.
-func RegisterAdapter(name string, adapter Adapter) {
-	if _, ok := configAdapters[name]; ok {
-		panic(fmt.Errorf("%s: already registered", name))
-	}
-	configAdapters[name] = adapter
-	caddy.RegisterModule(adapterModule{name, adapter})
-}
-
-// GetAdapter returns the adapter with the given name,
-// or nil if one with that name is not registered.
-func GetAdapter(name string) Adapter {
-	return configAdapters[name]
-}
-
-// adapterModule is a wrapper type that can turn any config
-// adapter into a Caddy module, which has the benefit of being
-// counted with other modules, even though they do not
-// technically extend the Caddy configuration structure.
-// See caddyserver/caddy#3132.
-type adapterModule struct {
-	name string
-	Adapter
-}
-
-func (am adapterModule) CaddyModule() caddy.ModuleInfo {
-	return caddy.ModuleInfo{
-		ID:  caddy.ModuleID("caddy.adapters." + am.name),
-		New: func() caddy.Module { return am },
-	}
-}
-
-var configAdapters = make(map[string]Adapter)

caddyconfig/httpcaddyfile/addresses.go

@@ -1,411 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package httpcaddyfile
-
-import (
-	"fmt"
-	"net"
-	"net/netip"
-	"reflect"
-	"sort"
-	"strconv"
-	"strings"
-	"unicode"
-
-	"github.com/caddyserver/caddy/v2"
-	"github.com/caddyserver/caddy/v2/caddyconfig/caddyfile"
-	"github.com/caddyserver/caddy/v2/modules/caddyhttp"
-	"github.com/caddyserver/certmagic"
-)
-
-// mapAddressToServerBlocks returns a map of listener address to list of server
-// blocks that will be served on that address. To do this, each server block is
-// expanded so that each one is considered individually, although keys of a
-// server block that share the same address stay grouped together so the config
-// isn't repeated unnecessarily. For example, this Caddyfile:
-//
-//	example.com {
-//		bind 127.0.0.1
-//	}
-//	www.example.com, example.net/path, localhost:9999 {
-//		bind 127.0.0.1 1.2.3.4
-//	}
-//
-// has two server blocks to start with. But expressed in this Caddyfile are
-// actually 4 listener addresses: 127.0.0.1:443, 1.2.3.4:443, 127.0.0.1:9999,
-// and 127.0.0.1:9999. This is because the bind directive is applied to each
-// key of its server block (specifying the host part), and each key may have
-// a different port. And we definitely need to be sure that a site which is
-// bound to be served on a specific interface is not served on others just
-// because that is more convenient: it would be a potential security risk
-// if the difference between interfaces means private vs. public.
-//
-// So what this function does for the example above is iterate each server
-// block, and for each server block, iterate its keys. For the first, it
-// finds one key (example.com) and determines its listener address
-// (127.0.0.1:443 - because of 'bind' and automatic HTTPS). It then adds
-// the listener address to the map value returned by this function, with
-// the first server block as one of its associations.
-//
-// It then iterates each key on the second server block and associates them
-// with one or more listener addresses. Indeed, each key in this block has
-// two listener addresses because of the 'bind' directive. Once we know
-// which addresses serve which keys, we can create a new server block for
-// each address containing the contents of the server block and only those
-// specific keys of the server block which use that address.
-//
-// It is possible and even likely that some keys in the returned map have
-// the exact same list of server blocks (i.e. they are identical). This
-// happens when multiple hosts are declared with a 'bind' directive and
-// the resulting listener addresses are not shared by any other server
-// block (or the other server blocks are exactly identical in their token
-// contents). This happens with our example above because 1.2.3.4:443
-// and 1.2.3.4:9999 are used exclusively with the second server block. This
-// repetition may be undesirable, so call consolidateAddrMappings() to map
-// multiple addresses to the same lists of server blocks (a many:many mapping).
-// (Doing this is essentially a map-reduce technique.)
-func (st *ServerType) mapAddressToServerBlocks(originalServerBlocks []serverBlock,
-	options map[string]any) (map[string][]serverBlock, error) {
-	sbmap := make(map[string][]serverBlock)
-
-	for i, sblock := range originalServerBlocks {
-		// within a server block, we need to map all the listener addresses
-		// implied by the server block to the keys of the server block which
-		// will be served by them; this has the effect of treating each
-		// key of a server block as its own, but without having to repeat its
-		// contents in cases where multiple keys really can be served together
-		addrToKeys := make(map[string][]string)
-		for j, key := range sblock.block.Keys {
-			// a key can have multiple listener addresses if there are multiple
-			// arguments to the 'bind' directive (although they will all have
-			// the same port, since the port is defined by the key or is implicit
-			// through automatic HTTPS)
-			addrs, err := st.listenerAddrsForServerBlockKey(sblock, key, options)
-			if err != nil {
-				return nil, fmt.Errorf("server block %d, key %d (%s): determining listener address: %v", i, j, key, err)
-			}
-
-			// associate this key with each listener address it is served on
-			for _, addr := range addrs {
-				addrToKeys[addr] = append(addrToKeys[addr], key)
-			}
-		}
-
-		// make a slice of the map keys so we can iterate in sorted order
-		addrs := make([]string, 0, len(addrToKeys))
-		for k := range addrToKeys {
-			addrs = append(addrs, k)
-		}
-		sort.Strings(addrs)
-
-		// now that we know which addresses serve which keys of this
-		// server block, we iterate that mapping and create a list of
-		// new server blocks for each address where the keys of the
-		// server block are only the ones which use the address; but
-		// the contents (tokens) are of course the same
-		for _, addr := range addrs {
-			keys := addrToKeys[addr]
-			// parse keys so that we only have to do it once
-			parsedKeys := make([]Address, 0, len(keys))
-			for _, key := range keys {
-				addr, err := ParseAddress(key)
-				if err != nil {
-					return nil, fmt.Errorf("parsing key '%s': %v", key, err)
-				}
-				parsedKeys = append(parsedKeys, addr.Normalize())
-			}
-			sbmap[addr] = append(sbmap[addr], serverBlock{
-				block: caddyfile.ServerBlock{
-					Keys:     keys,
-					Segments: sblock.block.Segments,
-				},
-				pile: sblock.pile,
-				keys: parsedKeys,
-			})
-		}
-	}
-
-	return sbmap, nil
-}
-
-// consolidateAddrMappings eliminates repetition of identical server blocks in a mapping of
-// single listener addresses to lists of server blocks. Since multiple addresses may serve
-// identical sites (server block contents), this function turns a 1:many mapping into a
-// many:many mapping. Server block contents (tokens) must be exactly identical so that
-// reflect.DeepEqual returns true in order for the addresses to be combined. Identical
-// entries are deleted from the addrToServerBlocks map. Essentially, each pairing (each
-// association from multiple addresses to multiple server blocks; i.e. each element of
-// the returned slice) becomes a server definition in the output JSON.
-func (st *ServerType) consolidateAddrMappings(addrToServerBlocks map[string][]serverBlock) []sbAddrAssociation {
-	sbaddrs := make([]sbAddrAssociation, 0, len(addrToServerBlocks))
-	for addr, sblocks := range addrToServerBlocks {
-		// we start with knowing that at least this address
-		// maps to these server blocks
-		a := sbAddrAssociation{
-			addresses:    []string{addr},
-			serverBlocks: sblocks,
-		}
-
-		// now find other addresses that map to identical
-		// server blocks and add them to our list of
-		// addresses, while removing them from the map
-		for otherAddr, otherSblocks := range addrToServerBlocks {
-			if addr == otherAddr {
-				continue
-			}
-			if reflect.DeepEqual(sblocks, otherSblocks) {
-				a.addresses = append(a.addresses, otherAddr)
-				delete(addrToServerBlocks, otherAddr)
-			}
-		}
-		sort.Strings(a.addresses)
-
-		sbaddrs = append(sbaddrs, a)
-	}
-
-	// sort them by their first address (we know there will always be at least one)
-	// to avoid problems with non-deterministic ordering (makes tests flaky)
-	sort.Slice(sbaddrs, func(i, j int) bool {
-		return sbaddrs[i].addresses[0] < sbaddrs[j].addresses[0]
-	})
-
-	return sbaddrs
-}
-
-// listenerAddrsForServerBlockKey essentially converts the Caddyfile
-// site addresses to Caddy listener addresses for each server block.
-func (st *ServerType) listenerAddrsForServerBlockKey(sblock serverBlock, key string,
-	options map[string]any) ([]string, error) {
-	addr, err := ParseAddress(key)
-	if err != nil {
-		return nil, fmt.Errorf("parsing key: %v", err)
-	}
-	addr = addr.Normalize()
-
-	// figure out the HTTP and HTTPS ports; either
-	// use defaults, or override with user config
-	httpPort, httpsPort := strconv.Itoa(caddyhttp.DefaultHTTPPort), strconv.Itoa(caddyhttp.DefaultHTTPSPort)
-	if hport, ok := options["http_port"]; ok {
-		httpPort = strconv.Itoa(hport.(int))
-	}
-	if hsport, ok := options["https_port"]; ok {
-		httpsPort = strconv.Itoa(hsport.(int))
-	}
-
-	// default port is the HTTPS port
-	lnPort := httpsPort
-	if addr.Port != "" {
-		// port explicitly defined
-		lnPort = addr.Port
-	} else if addr.Scheme == "http" {
-		// port inferred from scheme
-		lnPort = httpPort
-	}
-
-	// error if scheme and port combination violate convention
-	if (addr.Scheme == "http" && lnPort == httpsPort) || (addr.Scheme == "https" && lnPort == httpPort) {
-		return nil, fmt.Errorf("[%s] scheme and port violate convention", key)
-	}
-
-	// the bind directive specifies hosts (and potentially network), but is optional
-	lnHosts := make([]string, 0, len(sblock.pile["bind"]))
-	for _, cfgVal := range sblock.pile["bind"] {
-		lnHosts = append(lnHosts, cfgVal.Value.([]string)...)
-	}
-	if len(lnHosts) == 0 {
-		if defaultBind, ok := options["default_bind"].([]string); ok {
-			lnHosts = defaultBind
-		} else {
-			lnHosts = []string{""}
-		}
-	}
-
-	// use a map to prevent duplication
-	listeners := make(map[string]struct{})
-	for _, lnHost := range lnHosts {
-		// normally we would simply append the port,
-		// but if lnHost is IPv6, we need to ensure it
-		// is enclosed in [ ]; net.JoinHostPort does
-		// this for us, but lnHost might also have a
-		// network type in front (e.g. "tcp/") leading
-		// to "[tcp/::1]" which causes parsing failures
-		// later; what we need is "tcp/[::1]", so we have
-		// to split the network and host, then re-combine
-		network, host, ok := strings.Cut(lnHost, "/")
-		if !ok {
-			host = network
-			network = ""
-		}
-		host = strings.Trim(host, "[]") // IPv6
-		networkAddr := caddy.JoinNetworkAddress(network, host, lnPort)
-		addr, err := caddy.ParseNetworkAddress(networkAddr)
-		if err != nil {
-			return nil, fmt.Errorf("parsing network address: %v", err)
-		}
-		listeners[addr.String()] = struct{}{}
-	}
-
-	// now turn map into list
-	listenersList := make([]string, 0, len(listeners))
-	for lnStr := range listeners {
-		listenersList = append(listenersList, lnStr)
-	}
-	sort.Strings(listenersList)
-
-	return listenersList, nil
-}
-
-// Address represents a site address. It contains
-// the original input value, and the component
-// parts of an address. The component parts may be
-// updated to the correct values as setup proceeds,
-// but the original value should never be changed.
-//
-// The Host field must be in a normalized form.
-type Address struct {
-	Original, Scheme, Host, Port, Path string
-}
-
-// ParseAddress parses an address string into a structured format with separate
-// scheme, host, port, and path portions, as well as the original input string.
-func ParseAddress(str string) (Address, error) {
-	const maxLen = 4096
-	if len(str) > maxLen {
-		str = str[:maxLen]
-	}
-	remaining := strings.TrimSpace(str)
-	a := Address{Original: remaining}
-
-	// extract scheme
-	splitScheme := strings.SplitN(remaining, "://", 2)
-	switch len(splitScheme) {
-	case 0:
-		return a, nil
-	case 1:
-		remaining = splitScheme[0]
-	case 2:
-		a.Scheme = splitScheme[0]
-		remaining = splitScheme[1]
-	}
-
-	// extract host and port
-	hostSplit := strings.SplitN(remaining, "/", 2)
-	if len(hostSplit) > 0 {
-		host, port, err := net.SplitHostPort(hostSplit[0])
-		if err != nil {
-			host, port, err = net.SplitHostPort(hostSplit[0] + ":")
-			if err != nil {
-				host = hostSplit[0]
-			}
-		}
-		a.Host = host
-		a.Port = port
-	}
-	if len(hostSplit) == 2 {
-		// all that remains is the path
-		a.Path = "/" + hostSplit[1]
-	}
-
-	// make sure port is valid
-	if a.Port != "" {
-		if portNum, err := strconv.Atoi(a.Port); err != nil {
-			return Address{}, fmt.Errorf("invalid port '%s': %v", a.Port, err)
-		} else if portNum < 0 || portNum > 65535 {
-			return Address{}, fmt.Errorf("port %d is out of range", portNum)
-		}
-	}
-
-	return a, nil
-}
-
-// String returns a human-readable form of a. It will
-// be a cleaned-up and filled-out URL string.
-func (a Address) String() string {
-	if a.Host == "" && a.Port == "" {
-		return ""
-	}
-	scheme := a.Scheme
-	if scheme == "" {
-		if a.Port == strconv.Itoa(certmagic.HTTPSPort) {
-			scheme = "https"
-		} else {
-			scheme = "http"
-		}
-	}
-	s := scheme
-	if s != "" {
-		s += "://"
-	}
-	if a.Port != "" &&
-		((scheme == "https" && a.Port != strconv.Itoa(caddyhttp.DefaultHTTPSPort)) ||
-			(scheme == "http" && a.Port != strconv.Itoa(caddyhttp.DefaultHTTPPort))) {
-		s += net.JoinHostPort(a.Host, a.Port)
-	} else {
-		s += a.Host
-	}
-	if a.Path != "" {
-		s += a.Path
-	}
-	return s
-}
-
-// Normalize returns a normalized version of a.
-func (a Address) Normalize() Address {
-	path := a.Path
-
-	// ensure host is normalized if it's an IP address
-	host := strings.TrimSpace(a.Host)
-	if ip, err := netip.ParseAddr(host); err == nil {
-		if ip.Is6() && !ip.Is4() && !ip.Is4In6() {
-			host = ip.String()
-		}
-	}
-
-	return Address{
-		Original: a.Original,
-		Scheme:   lowerExceptPlaceholders(a.Scheme),
-		Host:     lowerExceptPlaceholders(host),
-		Port:     a.Port,
-		Path:     path,
-	}
-}
-
-// lowerExceptPlaceholders lowercases s except within
-// placeholders (substrings in non-escaped '{ }' spans).
-// See https://github.com/caddyserver/caddy/issues/3264
-func lowerExceptPlaceholders(s string) string {
-	var sb strings.Builder
-	var escaped, inPlaceholder bool
-	for _, ch := range s {
-		if ch == '\\' && !escaped {
-			escaped = true
-			sb.WriteRune(ch)
-			continue
-		}
-		if ch == '{' && !escaped {
-			inPlaceholder = true
-		}
-		if ch == '}' && inPlaceholder && !escaped {
-			inPlaceholder = false
-		}
-		if inPlaceholder {
-			sb.WriteRune(ch)
-		} else {
-			sb.WriteRune(unicode.ToLower(ch))
-		}
-		escaped = false
-	}
-	return sb.String()
-}

caddyconfig/httpcaddyfile/addresses_fuzz.go

@@ -1,28 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-//go:build gofuzz
-
-package httpcaddyfile
-
-func FuzzParseAddress(data []byte) int {
-	addr, err := ParseAddress(string(data))
-	if err != nil {
-		if addr == (Address{}) {
-			return 1
-		}
-		return 0
-	}
-	return 1
-}

caddyconfig/httpcaddyfile/addresses_test.go

@@ -1,253 +0,0 @@
-package httpcaddyfile
-
-import (
-	"testing"
-)
-
-func TestParseAddress(t *testing.T) {
-	for i, test := range []struct {
-		input                    string
-		scheme, host, port, path string
-		shouldErr                bool
-	}{
-		{``, "", "", "", "", false},
-		{`localhost`, "", "localhost", "", "", false},
-		{`localhost:1234`, "", "localhost", "1234", "", false},
-		{`localhost:`, "", "localhost", "", "", false},
-		{`0.0.0.0`, "", "0.0.0.0", "", "", false},
-		{`127.0.0.1:1234`, "", "127.0.0.1", "1234", "", false},
-		{`:1234`, "", "", "1234", "", false},
-		{`[::1]`, "", "::1", "", "", false},
-		{`[::1]:1234`, "", "::1", "1234", "", false},
-		{`:`, "", "", "", "", false},
-		{`:http`, "", "", "", "", true},
-		{`:https`, "", "", "", "", true},
-		{`localhost:http`, "", "", "", "", true}, // using service name in port is verboten, as of Go 1.12.8
-		{`localhost:https`, "", "", "", "", true},
-		{`http://localhost:https`, "", "", "", "", true}, // conflict
-		{`http://localhost:http`, "", "", "", "", true},  // repeated scheme
-		{`host:https/path`, "", "", "", "", true},
-		{`http://localhost:443`, "http", "localhost", "443", "", false}, // NOTE: not conventional
-		{`https://localhost:80`, "https", "localhost", "80", "", false}, // NOTE: not conventional
-		{`http://localhost`, "http", "localhost", "", "", false},
-		{`https://localhost`, "https", "localhost", "", "", false},
-		{`http://{env.APP_DOMAIN}`, "http", "{env.APP_DOMAIN}", "", "", false},
-		{`{env.APP_DOMAIN}:80`, "", "{env.APP_DOMAIN}", "80", "", false},
-		{`{env.APP_DOMAIN}/path`, "", "{env.APP_DOMAIN}", "", "/path", false},
-		{`example.com/{env.APP_PATH}`, "", "example.com", "", "/{env.APP_PATH}", false},
-		{`http://127.0.0.1`, "http", "127.0.0.1", "", "", false},
-		{`https://127.0.0.1`, "https", "127.0.0.1", "", "", false},
-		{`http://[::1]`, "http", "::1", "", "", false},
-		{`http://localhost:1234`, "http", "localhost", "1234", "", false},
-		{`https://127.0.0.1:1234`, "https", "127.0.0.1", "1234", "", false},
-		{`http://[::1]:1234`, "http", "::1", "1234", "", false},
-		{``, "", "", "", "", false},
-		{`::1`, "", "::1", "", "", false},
-		{`localhost::`, "", "localhost::", "", "", false},
-		{`#$%@`, "", "#$%@", "", "", false}, // don't want to presume what the hostname could be
-		{`host/path`, "", "host", "", "/path", false},
-		{`http://host/`, "http", "host", "", "/", false},
-		{`//asdf`, "", "", "", "//asdf", false},
-		{`:1234/asdf`, "", "", "1234", "/asdf", false},
-		{`http://host/path`, "http", "host", "", "/path", false},
-		{`https://host:443/path/foo`, "https", "host", "443", "/path/foo", false},
-		{`host:80/path`, "", "host", "80", "/path", false},
-		{`/path`, "", "", "", "/path", false},
-	} {
-		actual, err := ParseAddress(test.input)
-
-		if err != nil && !test.shouldErr {
-			t.Errorf("Test %d (%s): Expected no error, but had error: %v", i, test.input, err)
-		}
-		if err == nil && test.shouldErr {
-			t.Errorf("Test %d (%s): Expected error, but had none (%#v)", i, test.input, actual)
-		}
-
-		if !test.shouldErr && actual.Original != test.input {
-			t.Errorf("Test %d (%s): Expected original '%s', got '%s'", i, test.input, test.input, actual.Original)
-		}
-		if actual.Scheme != test.scheme {
-			t.Errorf("Test %d (%s): Expected scheme '%s', got '%s'", i, test.input, test.scheme, actual.Scheme)
-		}
-		if actual.Host != test.host {
-			t.Errorf("Test %d (%s): Expected host '%s', got '%s'", i, test.input, test.host, actual.Host)
-		}
-		if actual.Port != test.port {
-			t.Errorf("Test %d (%s): Expected port '%s', got '%s'", i, test.input, test.port, actual.Port)
-		}
-		if actual.Path != test.path {
-			t.Errorf("Test %d (%s): Expected path '%s', got '%s'", i, test.input, test.path, actual.Path)
-		}
-	}
-}
-
-func TestAddressString(t *testing.T) {
-	for i, test := range []struct {
-		addr     Address
-		expected string
-	}{
-		{Address{Scheme: "http", Host: "host", Port: "1234", Path: "/path"}, "http://host:1234/path"},
-		{Address{Scheme: "", Host: "host", Port: "", Path: ""}, "http://host"},
-		{Address{Scheme: "", Host: "host", Port: "80", Path: ""}, "http://host"},
-		{Address{Scheme: "", Host: "host", Port: "443", Path: ""}, "https://host"},
-		{Address{Scheme: "https", Host: "host", Port: "443", Path: ""}, "https://host"},
-		{Address{Scheme: "https", Host: "host", Port: "", Path: ""}, "https://host"},
-		{Address{Scheme: "", Host: "host", Port: "80", Path: "/path"}, "http://host/path"},
-		{Address{Scheme: "http", Host: "", Port: "1234", Path: ""}, "http://:1234"},
-		{Address{Scheme: "", Host: "", Port: "", Path: ""}, ""},
-	} {
-		actual := test.addr.String()
-		if actual != test.expected {
-			t.Errorf("Test %d: expected '%s' but got '%s'", i, test.expected, actual)
-		}
-	}
-}
-
-func TestKeyNormalization(t *testing.T) {
-	testCases := []struct {
-		input  string
-		expect Address
-	}{
-		{
-			input: "example.com",
-			expect: Address{
-				Host: "example.com",
-			},
-		},
-		{
-			input: "http://host:1234/path",
-			expect: Address{
-				Scheme: "http",
-				Host:   "host",
-				Port:   "1234",
-				Path:   "/path",
-			},
-		},
-		{
-			input: "HTTP://A/ABCDEF",
-			expect: Address{
-				Scheme: "http",
-				Host:   "a",
-				Path:   "/ABCDEF",
-			},
-		},
-		{
-			input: "A/ABCDEF",
-			expect: Address{
-				Host: "a",
-				Path: "/ABCDEF",
-			},
-		},
-		{
-			input: "A:2015/Path",
-			expect: Address{
-				Host: "a",
-				Port: "2015",
-				Path: "/Path",
-			},
-		},
-		{
-			input: "sub.{env.MY_DOMAIN}",
-			expect: Address{
-				Host: "sub.{env.MY_DOMAIN}",
-			},
-		},
-		{
-			input: "sub.ExAmPle",
-			expect: Address{
-				Host: "sub.example",
-			},
-		},
-		{
-			input: "sub.\\{env.MY_DOMAIN\\}",
-			expect: Address{
-				Host: "sub.\\{env.my_domain\\}",
-			},
-		},
-		{
-			input: "sub.{env.MY_DOMAIN}.com",
-			expect: Address{
-				Host: "sub.{env.MY_DOMAIN}.com",
-			},
-		},
-		{
-			input: ":80",
-			expect: Address{
-				Port: "80",
-			},
-		},
-		{
-			input: ":443",
-			expect: Address{
-				Port: "443",
-			},
-		},
-		{
-			input: ":1234",
-			expect: Address{
-				Port: "1234",
-			},
-		},
-		{
-			input:  "",
-			expect: Address{},
-		},
-		{
-			input:  ":",
-			expect: Address{},
-		},
-		{
-			input: "[::]",
-			expect: Address{
-				Host: "::",
-			},
-		},
-		{
-			input: "127.0.0.1",
-			expect: Address{
-				Host: "127.0.0.1",
-			},
-		},
-		{
-			input: "[2001:db8:85a3:8d3:1319:8a2e:370:7348]:1234",
-			expect: Address{
-				Host: "2001:db8:85a3:8d3:1319:8a2e:370:7348",
-				Port: "1234",
-			},
-		},
-		{
-			// IPv4 address in IPv6 form (#4381)
-			input: "[::ffff:cff4:e77d]:1234",
-			expect: Address{
-				Host: "::ffff:cff4:e77d",
-				Port: "1234",
-			},
-		},
-		{
-			input: "::ffff:cff4:e77d",
-			expect: Address{
-				Host: "::ffff:cff4:e77d",
-			},
-		},
-	}
-	for i, tc := range testCases {
-		addr, err := ParseAddress(tc.input)
-		if err != nil {
-			t.Errorf("Test %d: Parsing address '%s': %v", i, tc.input, err)
-			continue
-		}
-		actual := addr.Normalize()
-		if actual.Scheme != tc.expect.Scheme {
-			t.Errorf("Test %d: Input '%s': Expected Scheme='%s' but got Scheme='%s'", i, tc.input, tc.expect.Scheme, actual.Scheme)
-		}
-		if actual.Host != tc.expect.Host {
-			t.Errorf("Test %d: Input '%s': Expected Host='%s' but got Host='%s'", i, tc.input, tc.expect.Host, actual.Host)
-		}
-		if actual.Port != tc.expect.Port {
-			t.Errorf("Test %d: Input '%s': Expected Port='%s' but got Port='%s'", i, tc.input, tc.expect.Port, actual.Port)
-		}
-		if actual.Path != tc.expect.Path {
-			t.Errorf("Test %d: Input '%s': Expected Path='%s' but got Path='%s'", i, tc.input, tc.expect.Path, actual.Path)
-		}
-	}
-}

caddyconfig/httpcaddyfile/builtins.go

@@ -1,876 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package httpcaddyfile
-
-import (
-	"encoding/base64"
-	"encoding/pem"
-	"fmt"
-	"html"
-	"net/http"
-	"os"
-	"reflect"
-	"strconv"
-	"strings"
-
-	"github.com/caddyserver/caddy/v2"
-	"github.com/caddyserver/caddy/v2/caddyconfig"
-	"github.com/caddyserver/caddy/v2/caddyconfig/caddyfile"
-	"github.com/caddyserver/caddy/v2/modules/caddyhttp"
-	"github.com/caddyserver/caddy/v2/modules/caddytls"
-	"github.com/caddyserver/certmagic"
-	"github.com/mholt/acmez/acme"
-	"go.uber.org/zap/zapcore"
-)
-
-func init() {
-	RegisterDirective("bind", parseBind)
-	RegisterDirective("tls", parseTLS)
-	RegisterHandlerDirective("root", parseRoot)
-	RegisterHandlerDirective("vars", parseVars)
-	RegisterHandlerDirective("redir", parseRedir)
-	RegisterHandlerDirective("respond", parseRespond)
-	RegisterHandlerDirective("abort", parseAbort)
-	RegisterHandlerDirective("error", parseError)
-	RegisterHandlerDirective("route", parseRoute)
-	RegisterHandlerDirective("handle", parseHandle)
-	RegisterDirective("handle_errors", parseHandleErrors)
-	RegisterDirective("log", parseLog)
-	RegisterHandlerDirective("skip_log", parseSkipLog)
-}
-
-// parseBind parses the bind directive. Syntax:
-//
-//	bind <addresses...>
-func parseBind(h Helper) ([]ConfigValue, error) {
-	var lnHosts []string
-	for h.Next() {
-		lnHosts = append(lnHosts, h.RemainingArgs()...)
-	}
-	return h.NewBindAddresses(lnHosts), nil
-}
-
-// parseTLS parses the tls directive. Syntax:
-//
-//	tls [<email>|internal]|[<cert_file> <key_file>] {
-//	    protocols <min> [<max>]
-//	    ciphers   <cipher_suites...>
-//	    curves    <curves...>
-//	    client_auth {
-//	        mode                   [request|require|verify_if_given|require_and_verify]
-//	        trusted_ca_cert        <base64_der>
-//	        trusted_ca_cert_file   <filename>
-//	        trusted_leaf_cert      <base64_der>
-//	        trusted_leaf_cert_file <filename>
-//	    }
-//	    alpn      <values...>
-//	    load      <paths...>
-//	    ca        <acme_ca_endpoint>
-//	    ca_root   <pem_file>
-//	    dns       <provider_name> [...]
-//	    on_demand
-//	    eab    <key_id> <mac_key>
-//	    issuer <module_name> [...]
-//	    get_certificate <module_name> [...]
-//	    insecure_secrets_log <log_file>
-//	}
-func parseTLS(h Helper) ([]ConfigValue, error) {
-	cp := new(caddytls.ConnectionPolicy)
-	var fileLoader caddytls.FileLoader
-	var folderLoader caddytls.FolderLoader
-	var certSelector caddytls.CustomCertSelectionPolicy
-	var acmeIssuer *caddytls.ACMEIssuer
-	var keyType string
-	var internalIssuer *caddytls.InternalIssuer
-	var issuers []certmagic.Issuer
-	var certManagers []certmagic.Manager
-	var onDemand bool
-
-	for h.Next() {
-		// file certificate loader
-		firstLine := h.RemainingArgs()
-		switch len(firstLine) {
-		case 0:
-		case 1:
-			if firstLine[0] == "internal" {
-				internalIssuer = new(caddytls.InternalIssuer)
-			} else if !strings.Contains(firstLine[0], "@") {
-				return nil, h.Err("single argument must either be 'internal' or an email address")
-			} else {
-				if acmeIssuer == nil {
-					acmeIssuer = new(caddytls.ACMEIssuer)
-				}
-				acmeIssuer.Email = firstLine[0]
-			}
-
-		case 2:
-			certFilename := firstLine[0]
-			keyFilename := firstLine[1]
-
-			// tag this certificate so if multiple certs match, specifically
-			// this one that the user has provided will be used, see #2588:
-			// https://github.com/caddyserver/caddy/issues/2588 ... but we
-			// must be careful about how we do this; being careless will
-			// lead to failed handshakes
-			//
-			// we need to remember which cert files we've seen, since we
-			// must load each cert only once; otherwise, they each get a
-			// different tag... since a cert loaded twice has the same
-			// bytes, it will overwrite the first one in the cache, and
-			// only the last cert (and its tag) will survive, so any conn
-			// policy that is looking for any tag other than the last one
-			// to be loaded won't find it, and TLS handshakes will fail
-			// (see end of issue #3004)
-			//
-			// tlsCertTags maps certificate filenames to their tag.
-			// This is used to remember which tag is used for each
-			// certificate files, since we need to avoid loading
-			// the same certificate files more than once, overwriting
-			// previous tags
-			tlsCertTags, ok := h.State["tlsCertTags"].(map[string]string)
-			if !ok {
-				tlsCertTags = make(map[string]string)
-				h.State["tlsCertTags"] = tlsCertTags
-			}
-
-			tag, ok := tlsCertTags[certFilename]
-			if !ok {
-				// haven't seen this cert file yet, let's give it a tag
-				// and add a loader for it
-				tag = fmt.Sprintf("cert%d", len(tlsCertTags))
-				fileLoader = append(fileLoader, caddytls.CertKeyFilePair{
-					Certificate: certFilename,
-					Key:         keyFilename,
-					Tags:        []string{tag},
-				})
-				// remember this for next time we see this cert file
-				tlsCertTags[certFilename] = tag
-			}
-			certSelector.AnyTag = append(certSelector.AnyTag, tag)
-
-		default:
-			return nil, h.ArgErr()
-		}
-
-		var hasBlock bool
-		for nesting := h.Nesting(); h.NextBlock(nesting); {
-			hasBlock = true
-
-			switch h.Val() {
-			case "protocols":
-				args := h.RemainingArgs()
-				if len(args) == 0 {
-					return nil, h.SyntaxErr("one or two protocols")
-				}
-				if len(args) > 0 {
-					if _, ok := caddytls.SupportedProtocols[args[0]]; !ok {
-						return nil, h.Errf("Wrong protocol name or protocol not supported: '%s'", args[0])
-					}
-					cp.ProtocolMin = args[0]
-				}
-				if len(args) > 1 {
-					if _, ok := caddytls.SupportedProtocols[args[1]]; !ok {
-						return nil, h.Errf("Wrong protocol name or protocol not supported: '%s'", args[1])
-					}
-					cp.ProtocolMax = args[1]
-				}
-
-			case "ciphers":
-				for h.NextArg() {
-					if !caddytls.CipherSuiteNameSupported(h.Val()) {
-						return nil, h.Errf("Wrong cipher suite name or cipher suite not supported: '%s'", h.Val())
-					}
-					cp.CipherSuites = append(cp.CipherSuites, h.Val())
-				}
-
-			case "curves":
-				for h.NextArg() {
-					if _, ok := caddytls.SupportedCurves[h.Val()]; !ok {
-						return nil, h.Errf("Wrong curve name or curve not supported: '%s'", h.Val())
-					}
-					cp.Curves = append(cp.Curves, h.Val())
-				}
-
-			case "client_auth":
-				cp.ClientAuthentication = &caddytls.ClientAuthentication{}
-				for nesting := h.Nesting(); h.NextBlock(nesting); {
-					subdir := h.Val()
-					switch subdir {
-					case "mode":
-						if !h.Args(&cp.ClientAuthentication.Mode) {
-							return nil, h.ArgErr()
-						}
-						if h.NextArg() {
-							return nil, h.ArgErr()
-						}
-
-					case "trusted_ca_cert",
-						"trusted_leaf_cert":
-						if !h.NextArg() {
-							return nil, h.ArgErr()
-						}
-						if subdir == "trusted_ca_cert" {
-							cp.ClientAuthentication.TrustedCACerts = append(cp.ClientAuthentication.TrustedCACerts, h.Val())
-						} else {
-							cp.ClientAuthentication.TrustedLeafCerts = append(cp.ClientAuthentication.TrustedLeafCerts, h.Val())
-						}
-
-					case "trusted_ca_cert_file",
-						"trusted_leaf_cert_file":
-						if !h.NextArg() {
-							return nil, h.ArgErr()
-						}
-						filename := h.Val()
-						certDataPEM, err := os.ReadFile(filename)
-						if err != nil {
-							return nil, err
-						}
-						block, _ := pem.Decode(certDataPEM)
-						if block == nil || block.Type != "CERTIFICATE" {
-							return nil, h.Errf("no CERTIFICATE pem block found in %s", h.Val())
-						}
-						if subdir == "trusted_ca_cert_file" {
-							cp.ClientAuthentication.TrustedCACerts = append(cp.ClientAuthentication.TrustedCACerts,
-								base64.StdEncoding.EncodeToString(block.Bytes))
-						} else {
-							cp.ClientAuthentication.TrustedLeafCerts = append(cp.ClientAuthentication.TrustedLeafCerts,
-								base64.StdEncoding.EncodeToString(block.Bytes))
-						}
-
-					default:
-						return nil, h.Errf("unknown subdirective for client_auth: %s", subdir)
-					}
-				}
-
-			case "alpn":
-				args := h.RemainingArgs()
-				if len(args) == 0 {
-					return nil, h.ArgErr()
-				}
-				cp.ALPN = args
-
-			case "load":
-				folderLoader = append(folderLoader, h.RemainingArgs()...)
-
-			case "ca":
-				arg := h.RemainingArgs()
-				if len(arg) != 1 {
-					return nil, h.ArgErr()
-				}
-				if acmeIssuer == nil {
-					acmeIssuer = new(caddytls.ACMEIssuer)
-				}
-				acmeIssuer.CA = arg[0]
-
-			case "key_type":
-				arg := h.RemainingArgs()
-				if len(arg) != 1 {
-					return nil, h.ArgErr()
-				}
-				keyType = arg[0]
-
-			case "eab":
-				arg := h.RemainingArgs()
-				if len(arg) != 2 {
-					return nil, h.ArgErr()
-				}
-				if acmeIssuer == nil {
-					acmeIssuer = new(caddytls.ACMEIssuer)
-				}
-				acmeIssuer.ExternalAccount = &acme.EAB{
-					KeyID:  arg[0],
-					MACKey: arg[1],
-				}
-
-			case "issuer":
-				if !h.NextArg() {
-					return nil, h.ArgErr()
-				}
-				modName := h.Val()
-				modID := "tls.issuance." + modName
-				unm, err := caddyfile.UnmarshalModule(h.Dispenser, modID)
-				if err != nil {
-					return nil, err
-				}
-				issuer, ok := unm.(certmagic.Issuer)
-				if !ok {
-					return nil, h.Errf("module %s (%T) is not a certmagic.Issuer", modID, unm)
-				}
-				issuers = append(issuers, issuer)
-
-			case "get_certificate":
-				if !h.NextArg() {
-					return nil, h.ArgErr()
-				}
-				modName := h.Val()
-				modID := "tls.get_certificate." + modName
-				unm, err := caddyfile.UnmarshalModule(h.Dispenser, modID)
-				if err != nil {
-					return nil, err
-				}
-				certManager, ok := unm.(certmagic.Manager)
-				if !ok {
-					return nil, h.Errf("module %s (%T) is not a certmagic.CertificateManager", modID, unm)
-				}
-				certManagers = append(certManagers, certManager)
-
-			case "dns":
-				if !h.NextArg() {
-					return nil, h.ArgErr()
-				}
-				provName := h.Val()
-				if acmeIssuer == nil {
-					acmeIssuer = new(caddytls.ACMEIssuer)
-				}
-				if acmeIssuer.Challenges == nil {
-					acmeIssuer.Challenges = new(caddytls.ChallengesConfig)
-				}
-				if acmeIssuer.Challenges.DNS == nil {
-					acmeIssuer.Challenges.DNS = new(caddytls.DNSChallengeConfig)
-				}
-				modID := "dns.providers." + provName
-				unm, err := caddyfile.UnmarshalModule(h.Dispenser, modID)
-				if err != nil {
-					return nil, err
-				}
-				acmeIssuer.Challenges.DNS.ProviderRaw = caddyconfig.JSONModuleObject(unm, "name", provName, h.warnings)
-
-			case "resolvers":
-				args := h.RemainingArgs()
-				if len(args) == 0 {
-					return nil, h.ArgErr()
-				}
-				if acmeIssuer == nil {
-					acmeIssuer = new(caddytls.ACMEIssuer)
-				}
-				if acmeIssuer.Challenges == nil {
-					acmeIssuer.Challenges = new(caddytls.ChallengesConfig)
-				}
-				if acmeIssuer.Challenges.DNS == nil {
-					acmeIssuer.Challenges.DNS = new(caddytls.DNSChallengeConfig)
-				}
-				acmeIssuer.Challenges.DNS.Resolvers = args
-
-			case "dns_challenge_override_domain":
-				arg := h.RemainingArgs()
-				if len(arg) != 1 {
-					return nil, h.ArgErr()
-				}
-				if acmeIssuer == nil {
-					acmeIssuer = new(caddytls.ACMEIssuer)
-				}
-				if acmeIssuer.Challenges == nil {
-					acmeIssuer.Challenges = new(caddytls.ChallengesConfig)
-				}
-				if acmeIssuer.Challenges.DNS == nil {
-					acmeIssuer.Challenges.DNS = new(caddytls.DNSChallengeConfig)
-				}
-				acmeIssuer.Challenges.DNS.OverrideDomain = arg[0]
-
-			case "ca_root":
-				arg := h.RemainingArgs()
-				if len(arg) != 1 {
-					return nil, h.ArgErr()
-				}
-				if acmeIssuer == nil {
-					acmeIssuer = new(caddytls.ACMEIssuer)
-				}
-				acmeIssuer.TrustedRootsPEMFiles = append(acmeIssuer.TrustedRootsPEMFiles, arg[0])
-
-			case "on_demand":
-				if h.NextArg() {
-					return nil, h.ArgErr()
-				}
-				onDemand = true
-
-			case "insecure_secrets_log":
-				if !h.NextArg() {
-					return nil, h.ArgErr()
-				}
-				cp.InsecureSecretsLog = h.Val()
-
-			default:
-				return nil, h.Errf("unknown subdirective: %s", h.Val())
-			}
-		}
-
-		// a naked tls directive is not allowed
-		if len(firstLine) == 0 && !hasBlock {
-			return nil, h.ArgErr()
-		}
-	}
-
-	// begin building the final config values
-	configVals := []ConfigValue{}
-
-	// certificate loaders
-	if len(fileLoader) > 0 {
-		configVals = append(configVals, ConfigValue{
-			Class: "tls.cert_loader",
-			Value: fileLoader,
-		})
-	}
-	if len(folderLoader) > 0 {
-		configVals = append(configVals, ConfigValue{
-			Class: "tls.cert_loader",
-			Value: folderLoader,
-		})
-	}
-
-	// some tls subdirectives are shortcuts that implicitly configure issuers, and the
-	// user can also configure issuers explicitly using the issuer subdirective; the
-	// logic to support both would likely be complex, or at least unintuitive
-	if len(issuers) > 0 && (acmeIssuer != nil || internalIssuer != nil) {
-		return nil, h.Err("cannot mix issuer subdirective (explicit issuers) with other issuer-specific subdirectives (implicit issuers)")
-	}
-	if acmeIssuer != nil && internalIssuer != nil {
-		return nil, h.Err("cannot create both ACME and internal certificate issuers")
-	}
-
-	// now we should either have: explicitly-created issuers, or an implicitly-created
-	// ACME or internal issuer, or no issuers at all
-	switch {
-	case len(issuers) > 0:
-		for _, issuer := range issuers {
-			configVals = append(configVals, ConfigValue{
-				Class: "tls.cert_issuer",
-				Value: issuer,
-			})
-		}
-
-	case acmeIssuer != nil:
-		// implicit ACME issuers (from various subdirectives) - use defaults; there might be more than one
-		defaultIssuers := caddytls.DefaultIssuers()
-
-		// if a CA endpoint was set, override multiple implicit issuers since it's a specific one
-		if acmeIssuer.CA != "" {
-			defaultIssuers = []certmagic.Issuer{acmeIssuer}
-		}
-
-		for _, issuer := range defaultIssuers {
-			switch iss := issuer.(type) {
-			case *caddytls.ACMEIssuer:
-				issuer = acmeIssuer
-			case *caddytls.ZeroSSLIssuer:
-				iss.ACMEIssuer = acmeIssuer
-			}
-			configVals = append(configVals, ConfigValue{
-				Class: "tls.cert_issuer",
-				Value: issuer,
-			})
-		}
-
-	case internalIssuer != nil:
-		configVals = append(configVals, ConfigValue{
-			Class: "tls.cert_issuer",
-			Value: internalIssuer,
-		})
-	}
-
-	// certificate key type
-	if keyType != "" {
-		configVals = append(configVals, ConfigValue{
-			Class: "tls.key_type",
-			Value: keyType,
-		})
-	}
-
-	// on-demand TLS
-	if onDemand {
-		configVals = append(configVals, ConfigValue{
-			Class: "tls.on_demand",
-			Value: true,
-		})
-	}
-	for _, certManager := range certManagers {
-		configVals = append(configVals, ConfigValue{
-			Class: "tls.cert_manager",
-			Value: certManager,
-		})
-	}
-
-	// custom certificate selection
-	if len(certSelector.AnyTag) > 0 {
-		cp.CertSelection = &certSelector
-	}
-
-	// connection policy -- always add one, to ensure that TLS
-	// is enabled, because this directive was used (this is
-	// needed, for instance, when a site block has a key of
-	// just ":5000" - i.e. no hostname, and only on-demand TLS
-	// is enabled)
-	configVals = append(configVals, ConfigValue{
-		Class: "tls.connection_policy",
-		Value: cp,
-	})
-
-	return configVals, nil
-}
-
-// parseRoot parses the root directive. Syntax:
-//
-//	root [<matcher>] <path>
-func parseRoot(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	var root string
-	for h.Next() {
-		if !h.NextArg() {
-			return nil, h.ArgErr()
-		}
-		root = h.Val()
-		if h.NextArg() {
-			return nil, h.ArgErr()
-		}
-	}
-	return caddyhttp.VarsMiddleware{"root": root}, nil
-}
-
-// parseVars parses the vars directive. See its UnmarshalCaddyfile method for syntax.
-func parseVars(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	v := new(caddyhttp.VarsMiddleware)
-	err := v.UnmarshalCaddyfile(h.Dispenser)
-	return v, err
-}
-
-// parseRedir parses the redir directive. Syntax:
-//
-//	redir [<matcher>] <to> [<code>]
-//
-// <code> can be "permanent" for 301, "temporary" for 302 (default),
-// a placeholder, or any number in the 3xx range or 401. The special
-// code "html" can be used to redirect only browser clients (will
-// respond with HTTP 200 and no Location header; redirect is performed
-// with JS and a meta tag).
-func parseRedir(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	if !h.Next() {
-		return nil, h.ArgErr()
-	}
-
-	if !h.NextArg() {
-		return nil, h.ArgErr()
-	}
-	to := h.Val()
-
-	var code string
-	if h.NextArg() {
-		code = h.Val()
-	}
-
-	var body string
-	var hdr http.Header
-	switch code {
-	case "permanent":
-		code = "301"
-	case "temporary", "":
-		code = "302"
-	case "html":
-		// Script tag comes first since that will better imitate a redirect in the browser's
-		// history, but the meta tag is a fallback for most non-JS clients.
-		const metaRedir = `<!DOCTYPE html>
-<html>
-	<head>
-		<title>Redirecting...</title>
-		<script>window.location.replace("%s");</script>
-		<meta http-equiv="refresh" content="0; URL='%s'">
-	</head>
-	<body>Redirecting to <a href="%s">%s</a>...</body>
-</html>
-`
-		safeTo := html.EscapeString(to)
-		body = fmt.Sprintf(metaRedir, safeTo, safeTo, safeTo, safeTo)
-		code = "200" // don't redirect non-browser clients
-	default:
-		// Allow placeholders for the code
-		if strings.HasPrefix(code, "{") {
-			break
-		}
-		// Try to validate as an integer otherwise
-		codeInt, err := strconv.Atoi(code)
-		if err != nil {
-			return nil, h.Errf("Not a supported redir code type or not valid integer: '%s'", code)
-		}
-		// Sometimes, a 401 with Location header is desirable because
-		// requests made with XHR will "eat" the 3xx redirect; so if
-		// the intent was to redirect to an auth page, a 3xx won't
-		// work. Responding with 401 allows JS code to read the
-		// Location header and do a window.location redirect manually.
-		// see https://stackoverflow.com/a/2573589/846934
-		// see https://github.com/oauth2-proxy/oauth2-proxy/issues/1522
-		if codeInt < 300 || (codeInt > 399 && codeInt != 401) {
-			return nil, h.Errf("Redir code not in the 3xx range or 401: '%v'", codeInt)
-		}
-	}
-
-	// don't redirect non-browser clients
-	if code != "200" {
-		hdr = http.Header{"Location": []string{to}}
-	}
-
-	return caddyhttp.StaticResponse{
-		StatusCode: caddyhttp.WeakString(code),
-		Headers:    hdr,
-		Body:       body,
-	}, nil
-}
-
-// parseRespond parses the respond directive.
-func parseRespond(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	sr := new(caddyhttp.StaticResponse)
-	err := sr.UnmarshalCaddyfile(h.Dispenser)
-	if err != nil {
-		return nil, err
-	}
-	return sr, nil
-}
-
-// parseAbort parses the abort directive.
-func parseAbort(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	h.Next() // consume directive
-	for h.Next() || h.NextBlock(0) {
-		return nil, h.ArgErr()
-	}
-	return &caddyhttp.StaticResponse{Abort: true}, nil
-}
-
-// parseError parses the error directive.
-func parseError(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	se := new(caddyhttp.StaticError)
-	err := se.UnmarshalCaddyfile(h.Dispenser)
-	if err != nil {
-		return nil, err
-	}
-	return se, nil
-}
-
-// parseRoute parses the route directive.
-func parseRoute(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	sr := new(caddyhttp.Subroute)
-
-	allResults, err := parseSegmentAsConfig(h)
-	if err != nil {
-		return nil, err
-	}
-
-	for _, result := range allResults {
-		switch handler := result.Value.(type) {
-		case caddyhttp.Route:
-			sr.Routes = append(sr.Routes, handler)
-		case caddyhttp.Subroute:
-			// directives which return a literal subroute instead of a route
-			// means they intend to keep those handlers together without
-			// them being reordered; we're doing that anyway since we're in
-			// the route directive, so just append its handlers
-			sr.Routes = append(sr.Routes, handler.Routes...)
-		default:
-			return nil, h.Errf("%s directive returned something other than an HTTP route or subroute: %#v (only handler directives can be used in routes)", result.directive, result.Value)
-		}
-	}
-
-	return sr, nil
-}
-
-func parseHandle(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	return ParseSegmentAsSubroute(h)
-}
-
-func parseHandleErrors(h Helper) ([]ConfigValue, error) {
-	subroute, err := ParseSegmentAsSubroute(h)
-	if err != nil {
-		return nil, err
-	}
-	return []ConfigValue{
-		{
-			Class: "error_route",
-			Value: subroute,
-		},
-	}, nil
-}
-
-// parseLog parses the log directive. Syntax:
-//
-//	log {
-//	    output <writer_module> ...
-//	    format <encoder_module> ...
-//	    level  <level>
-//	}
-func parseLog(h Helper) ([]ConfigValue, error) {
-	return parseLogHelper(h, nil)
-}
-
-// parseLogHelper is used both for the parseLog directive within Server Blocks,
-// as well as the global "log" option for configuring loggers at the global
-// level. The parseAsGlobalOption parameter is used to distinguish any differing logic
-// between the two.
-func parseLogHelper(h Helper, globalLogNames map[string]struct{}) ([]ConfigValue, error) {
-	// When the globalLogNames parameter is passed in, we make
-	// modifications to the parsing behavior.
-	parseAsGlobalOption := globalLogNames != nil
-
-	var configValues []ConfigValue
-	for h.Next() {
-		// Logic below expects that a name is always present when a
-		// global option is being parsed.
-		var globalLogName string
-		if parseAsGlobalOption {
-			if h.NextArg() {
-				globalLogName = h.Val()
-
-				// Only a single argument is supported.
-				if h.NextArg() {
-					return nil, h.ArgErr()
-				}
-			} else {
-				// If there is no log name specified, we
-				// reference the default logger. See the
-				// setupNewDefault function in the logging
-				// package for where this is configured.
-				globalLogName = "default"
-			}
-
-			// Verify this name is unused.
-			_, used := globalLogNames[globalLogName]
-			if used {
-				return nil, h.Err("duplicate global log option for: " + globalLogName)
-			}
-			globalLogNames[globalLogName] = struct{}{}
-		} else {
-			// No arguments are supported for the server block log directive
-			if h.NextArg() {
-				return nil, h.ArgErr()
-			}
-		}
-
-		cl := new(caddy.CustomLog)
-
-		for h.NextBlock(0) {
-			switch h.Val() {
-			case "output":
-				if !h.NextArg() {
-					return nil, h.ArgErr()
-				}
-				moduleName := h.Val()
-
-				// can't use the usual caddyfile.Unmarshaler flow with the
-				// standard writers because they are in the caddy package
-				// (because they are the default) and implementing that
-				// interface there would unfortunately create circular import
-				var wo caddy.WriterOpener
-				switch moduleName {
-				case "stdout":
-					wo = caddy.StdoutWriter{}
-				case "stderr":
-					wo = caddy.StderrWriter{}
-				case "discard":
-					wo = caddy.DiscardWriter{}
-				default:
-					modID := "caddy.logging.writers." + moduleName
-					unm, err := caddyfile.UnmarshalModule(h.Dispenser, modID)
-					if err != nil {
-						return nil, err
-					}
-					var ok bool
-					wo, ok = unm.(caddy.WriterOpener)
-					if !ok {
-						return nil, h.Errf("module %s (%T) is not a WriterOpener", modID, unm)
-					}
-				}
-				cl.WriterRaw = caddyconfig.JSONModuleObject(wo, "output", moduleName, h.warnings)
-
-			case "format":
-				if !h.NextArg() {
-					return nil, h.ArgErr()
-				}
-				moduleName := h.Val()
-				moduleID := "caddy.logging.encoders." + moduleName
-				unm, err := caddyfile.UnmarshalModule(h.Dispenser, moduleID)
-				if err != nil {
-					return nil, err
-				}
-				enc, ok := unm.(zapcore.Encoder)
-				if !ok {
-					return nil, h.Errf("module %s (%T) is not a zapcore.Encoder", moduleID, unm)
-				}
-				cl.EncoderRaw = caddyconfig.JSONModuleObject(enc, "format", moduleName, h.warnings)
-
-			case "level":
-				if !h.NextArg() {
-					return nil, h.ArgErr()
-				}
-				cl.Level = h.Val()
-				if h.NextArg() {
-					return nil, h.ArgErr()
-				}
-
-			case "include":
-				// This configuration is only allowed in the global options
-				if !parseAsGlobalOption {
-					return nil, h.ArgErr()
-				}
-				for h.NextArg() {
-					cl.Include = append(cl.Include, h.Val())
-				}
-
-			case "exclude":
-				// This configuration is only allowed in the global options
-				if !parseAsGlobalOption {
-					return nil, h.ArgErr()
-				}
-				for h.NextArg() {
-					cl.Exclude = append(cl.Exclude, h.Val())
-				}
-
-			default:
-				return nil, h.Errf("unrecognized subdirective: %s", h.Val())
-			}
-		}
-
-		var val namedCustomLog
-		// Skip handling of empty logging configs
-		if !reflect.DeepEqual(cl, new(caddy.CustomLog)) {
-			if parseAsGlobalOption {
-				// Use indicated name for global log options
-				val.name = globalLogName
-				val.log = cl
-			} else {
-				// Construct a log name for server log streams
-				logCounter, ok := h.State["logCounter"].(int)
-				if !ok {
-					logCounter = 0
-				}
-				val.name = fmt.Sprintf("log%d", logCounter)
-				cl.Include = []string{"http.log.access." + val.name}
-				val.log = cl
-				logCounter++
-				h.State["logCounter"] = logCounter
-			}
-		}
-		configValues = append(configValues, ConfigValue{
-			Class: "custom_log",
-			Value: val,
-		})
-	}
-	return configValues, nil
-}
-
-// parseSkipLog parses the skip_log directive. Syntax:
-//
-//	skip_log [<matcher>]
-func parseSkipLog(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	for h.Next() {
-		if h.NextArg() {
-			return nil, h.ArgErr()
-		}
-	}
-	return caddyhttp.VarsMiddleware{"skip_log": true}, nil
-}

caddyconfig/httpcaddyfile/builtins_test.go

@@ -1,215 +0,0 @@
-package httpcaddyfile
-
-import (
-	"testing"
-
-	"github.com/caddyserver/caddy/v2/caddyconfig/caddyfile"
-	_ "github.com/caddyserver/caddy/v2/modules/logging"
-)
-
-func TestLogDirectiveSyntax(t *testing.T) {
-	for i, tc := range []struct {
-		input       string
-		output      string
-		expectError bool
-	}{
-		{
-			input: `:8080 {
-				log
-			}
-			`,
-			output:      `{"apps":{"http":{"servers":{"srv0":{"listen":[":8080"],"logs":{}}}}}}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				log {
-					output file foo.log
-				}
-			}
-			`,
-			output:      `{"logging":{"logs":{"default":{"exclude":["http.log.access.log0"]},"log0":{"writer":{"filename":"foo.log","output":"file"},"include":["http.log.access.log0"]}}},"apps":{"http":{"servers":{"srv0":{"listen":[":8080"],"logs":{"default_logger_name":"log0"}}}}}}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				log {
-					format filter {
-						wrap console
-						fields {
-							request>remote_ip ip_mask {
-								ipv4 24
-								ipv6 32
-							}
-						}
-					}
-				}
-			}
-			`,
-			output:      `{"logging":{"logs":{"default":{"exclude":["http.log.access.log0"]},"log0":{"encoder":{"fields":{"request\u003eremote_ip":{"filter":"ip_mask","ipv4_cidr":24,"ipv6_cidr":32}},"format":"filter","wrap":{"format":"console"}},"include":["http.log.access.log0"]}}},"apps":{"http":{"servers":{"srv0":{"listen":[":8080"],"logs":{"default_logger_name":"log0"}}}}}}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				log invalid {
-					output file foo.log
-				}
-			}
-			`,
-			expectError: true,
-		},
-	} {
-
-		adapter := caddyfile.Adapter{
-			ServerType: ServerType{},
-		}
-
-		out, _, err := adapter.Adapt([]byte(tc.input), nil)
-
-		if err != nil != tc.expectError {
-			t.Errorf("Test %d error expectation failed Expected: %v, got %s", i, tc.expectError, err)
-			continue
-		}
-
-		if string(out) != tc.output {
-			t.Errorf("Test %d error output mismatch Expected: %s, got %s", i, tc.output, out)
-		}
-	}
-}
-
-func TestRedirDirectiveSyntax(t *testing.T) {
-	for i, tc := range []struct {
-		input       string
-		expectError bool
-	}{
-		{
-			input: `:8080 {
-				redir :8081
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir * :8081
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir /api/* :8081 300
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir :8081 300
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir /api/* :8081 399
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir :8081 399
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir /old.html /new.html
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir /old.html /new.html temporary
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir https://example.com{uri} permanent
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir /old.html /new.html permanent
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir /old.html /new.html html
-			}`,
-			expectError: false,
-		},
-		{
-			// this is now allowed so a Location header
-			// can be written and consumed by JS
-			// in the case of XHR requests
-			input: `:8080 {
-				redir * :8081 401
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir * :8081 402
-			}`,
-			expectError: true,
-		},
-		{
-			input: `:8080 {
-				redir * :8081 {http.reverse_proxy.status_code}
-			}`,
-			expectError: false,
-		},
-		{
-			input: `:8080 {
-				redir /old.html /new.html htlm
-			}`,
-			expectError: true,
-		},
-		{
-			input: `:8080 {
-				redir * :8081 200
-			}`,
-			expectError: true,
-		},
-		{
-			input: `:8080 {
-				redir * :8081 temp
-			}`,
-			expectError: true,
-		},
-		{
-			input: `:8080 {
-				redir * :8081 perm
-			}`,
-			expectError: true,
-		},
-		{
-			input: `:8080 {
-				redir * :8081 php
-			}`,
-			expectError: true,
-		},
-	} {
-
-		adapter := caddyfile.Adapter{
-			ServerType: ServerType{},
-		}
-
-		_, _, err := adapter.Adapt([]byte(tc.input), nil)
-
-		if err != nil != tc.expectError {
-			t.Errorf("Test %d error expectation failed Expected: %v, got %s", i, tc.expectError, err)
-			continue
-		}
-	}
-}

caddyconfig/httpcaddyfile/directives.go

@@ -1,583 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package httpcaddyfile
-
-import (
-	"encoding/json"
-	"net"
-	"sort"
-	"strconv"
-	"strings"
-
-	"github.com/caddyserver/caddy/v2"
-	"github.com/caddyserver/caddy/v2/caddyconfig"
-	"github.com/caddyserver/caddy/v2/caddyconfig/caddyfile"
-	"github.com/caddyserver/caddy/v2/modules/caddyhttp"
-)
-
-// directiveOrder specifies the order
-// to apply directives in HTTP routes.
-//
-// The root directive goes first in case rewrites or
-// redirects depend on existence of files, i.e. the
-// file matcher, which must know the root first.
-//
-// The header directive goes second so that headers
-// can be manipulated before doing redirects.
-var directiveOrder = []string{
-	"tracing",
-
-	"map",
-	"vars",
-	"root",
-	"skip_log",
-
-	"header",
-	"copy_response_headers", // only in reverse_proxy's handle_response
-	"request_body",
-
-	"redir",
-
-	// incoming request manipulation
-	"method",
-	"rewrite",
-	"uri",
-	"try_files",
-
-	// middleware handlers; some wrap responses
-	"basicauth",
-	"forward_auth",
-	"request_header",
-	"encode",
-	"push",
-	"templates",
-
-	// special routing & dispatching directives
-	"handle",
-	"handle_path",
-	"route",
-
-	// handlers that typically respond to requests
-	"abort",
-	"error",
-	"copy_response", // only in reverse_proxy's handle_response
-	"respond",
-	"metrics",
-	"reverse_proxy",
-	"php_fastcgi",
-	"file_server",
-	"acme_server",
-}
-
-// directiveIsOrdered returns true if dir is
-// a known, ordered (sorted) directive.
-func directiveIsOrdered(dir string) bool {
-	for _, d := range directiveOrder {
-		if d == dir {
-			return true
-		}
-	}
-	return false
-}
-
-// RegisterDirective registers a unique directive dir with an
-// associated unmarshaling (setup) function. When directive dir
-// is encountered in a Caddyfile, setupFunc will be called to
-// unmarshal its tokens.
-func RegisterDirective(dir string, setupFunc UnmarshalFunc) {
-	if _, ok := registeredDirectives[dir]; ok {
-		panic("directive " + dir + " already registered")
-	}
-	registeredDirectives[dir] = setupFunc
-}
-
-// RegisterHandlerDirective is like RegisterDirective, but for
-// directives which specifically output only an HTTP handler.
-// Directives registered with this function will always have
-// an optional matcher token as the first argument.
-func RegisterHandlerDirective(dir string, setupFunc UnmarshalHandlerFunc) {
-	RegisterDirective(dir, func(h Helper) ([]ConfigValue, error) {
-		if !h.Next() {
-			return nil, h.ArgErr()
-		}
-
-		matcherSet, err := h.ExtractMatcherSet()
-		if err != nil {
-			return nil, err
-		}
-
-		val, err := setupFunc(h)
-		if err != nil {
-			return nil, err
-		}
-
-		return h.NewRoute(matcherSet, val), nil
-	})
-}
-
-// RegisterGlobalOption registers a unique global option opt with
-// an associated unmarshaling (setup) function. When the global
-// option opt is encountered in a Caddyfile, setupFunc will be
-// called to unmarshal its tokens.
-func RegisterGlobalOption(opt string, setupFunc UnmarshalGlobalFunc) {
-	if _, ok := registeredGlobalOptions[opt]; ok {
-		panic("global option " + opt + " already registered")
-	}
-	registeredGlobalOptions[opt] = setupFunc
-}
-
-// Helper is a type which helps setup a value from
-// Caddyfile tokens.
-type Helper struct {
-	*caddyfile.Dispenser
-	// State stores intermediate variables during caddyfile adaptation.
-	State        map[string]any
-	options      map[string]any
-	warnings     *[]caddyconfig.Warning
-	matcherDefs  map[string]caddy.ModuleMap
-	parentBlock  caddyfile.ServerBlock
-	groupCounter counter
-}
-
-// Option gets the option keyed by name.
-func (h Helper) Option(name string) any {
-	return h.options[name]
-}
-
-// Caddyfiles returns the list of config files from
-// which tokens in the current server block were loaded.
-func (h Helper) Caddyfiles() []string {
-	// first obtain set of names of files involved
-	// in this server block, without duplicates
-	files := make(map[string]struct{})
-	for _, segment := range h.parentBlock.Segments {
-		for _, token := range segment {
-			files[token.File] = struct{}{}
-		}
-	}
-	// then convert the set into a slice
-	filesSlice := make([]string, 0, len(files))
-	for file := range files {
-		filesSlice = append(filesSlice, file)
-	}
-	return filesSlice
-}
-
-// JSON converts val into JSON. Any errors are added to warnings.
-func (h Helper) JSON(val any) json.RawMessage {
-	return caddyconfig.JSON(val, h.warnings)
-}
-
-// MatcherToken assumes the next argument token is (possibly) a matcher,
-// and if so, returns the matcher set along with a true value. If the next
-// token is not a matcher, nil and false is returned. Note that a true
-// value may be returned with a nil matcher set if it is a catch-all.
-func (h Helper) MatcherToken() (caddy.ModuleMap, bool, error) {
-	if !h.NextArg() {
-		return nil, false, nil
-	}
-	return matcherSetFromMatcherToken(h.Dispenser.Token(), h.matcherDefs, h.warnings)
-}
-
-// ExtractMatcherSet is like MatcherToken, except this is a higher-level
-// method that returns the matcher set described by the matcher token,
-// or nil if there is none, and deletes the matcher token from the
-// dispenser and resets it as if this look-ahead never happened. Useful
-// when wrapping a route (one or more handlers) in a user-defined matcher.
-func (h Helper) ExtractMatcherSet() (caddy.ModuleMap, error) {
-	matcherSet, hasMatcher, err := h.MatcherToken()
-	if err != nil {
-		return nil, err
-	}
-	if hasMatcher {
-		// strip matcher token; we don't need to
-		// use the return value here because a
-		// new dispenser should have been made
-		// solely for this directive's tokens,
-		// with no other uses of same slice
-		h.Dispenser.Delete()
-	}
-	h.Dispenser.Reset() // pretend this lookahead never happened
-	return matcherSet, nil
-}
-
-// NewRoute returns config values relevant to creating a new HTTP route.
-func (h Helper) NewRoute(matcherSet caddy.ModuleMap,
-	handler caddyhttp.MiddlewareHandler) []ConfigValue {
-	mod, err := caddy.GetModule(caddy.GetModuleID(handler))
-	if err != nil {
-		*h.warnings = append(*h.warnings, caddyconfig.Warning{
-			File:    h.File(),
-			Line:    h.Line(),
-			Message: err.Error(),
-		})
-		return nil
-	}
-	var matcherSetsRaw []caddy.ModuleMap
-	if matcherSet != nil {
-		matcherSetsRaw = append(matcherSetsRaw, matcherSet)
-	}
-	return []ConfigValue{
-		{
-			Class: "route",
-			Value: caddyhttp.Route{
-				MatcherSetsRaw: matcherSetsRaw,
-				HandlersRaw:    []json.RawMessage{caddyconfig.JSONModuleObject(handler, "handler", mod.ID.Name(), h.warnings)},
-			},
-		},
-	}
-}
-
-// GroupRoutes adds the routes (caddyhttp.Route type) in vals to the
-// same group, if there is more than one route in vals.
-func (h Helper) GroupRoutes(vals []ConfigValue) {
-	// ensure there's at least two routes; group of one is pointless
-	var count int
-	for _, v := range vals {
-		if _, ok := v.Value.(caddyhttp.Route); ok {
-			count++
-			if count > 1 {
-				break
-			}
-		}
-	}
-	if count < 2 {
-		return
-	}
-
-	// now that we know the group will have some effect, do it
-	groupName := h.groupCounter.nextGroup()
-	for i := range vals {
-		if route, ok := vals[i].Value.(caddyhttp.Route); ok {
-			route.Group = groupName
-			vals[i].Value = route
-		}
-	}
-}
-
-// NewBindAddresses returns config values relevant to adding
-// listener bind addresses to the config.
-func (h Helper) NewBindAddresses(addrs []string) []ConfigValue {
-	return []ConfigValue{{Class: "bind", Value: addrs}}
-}
-
-// WithDispenser returns a new instance based on d. All others Helper
-// fields are copied, so typically maps are shared with this new instance.
-func (h Helper) WithDispenser(d *caddyfile.Dispenser) Helper {
-	h.Dispenser = d
-	return h
-}
-
-// ParseSegmentAsSubroute parses the segment such that its subdirectives
-// are themselves treated as directives, from which a subroute is built
-// and returned.
-func ParseSegmentAsSubroute(h Helper) (caddyhttp.MiddlewareHandler, error) {
-	allResults, err := parseSegmentAsConfig(h)
-	if err != nil {
-		return nil, err
-	}
-
-	return buildSubroute(allResults, h.groupCounter)
-}
-
-// parseSegmentAsConfig parses the segment such that its subdirectives
-// are themselves treated as directives, including named matcher definitions,
-// and the raw Config structs are returned.
-func parseSegmentAsConfig(h Helper) ([]ConfigValue, error) {
-	var allResults []ConfigValue
-
-	for h.Next() {
-		// don't allow non-matcher args on the first line
-		if h.NextArg() {
-			return nil, h.ArgErr()
-		}
-
-		// slice the linear list of tokens into top-level segments
-		var segments []caddyfile.Segment
-		for nesting := h.Nesting(); h.NextBlock(nesting); {
-			segments = append(segments, h.NextSegment())
-		}
-
-		// copy existing matcher definitions so we can augment
-		// new ones that are defined only in this scope
-		matcherDefs := make(map[string]caddy.ModuleMap, len(h.matcherDefs))
-		for key, val := range h.matcherDefs {
-			matcherDefs[key] = val
-		}
-
-		// find and extract any embedded matcher definitions in this scope
-		for i := 0; i < len(segments); i++ {
-			seg := segments[i]
-			if strings.HasPrefix(seg.Directive(), matcherPrefix) {
-				// parse, then add the matcher to matcherDefs
-				err := parseMatcherDefinitions(caddyfile.NewDispenser(seg), matcherDefs)
-				if err != nil {
-					return nil, err
-				}
-				// remove the matcher segment (consumed), then step back the loop
-				segments = append(segments[:i], segments[i+1:]...)
-				i--
-			}
-		}
-
-		// with matchers ready to go, evaluate each directive's segment
-		for _, seg := range segments {
-			dir := seg.Directive()
-			dirFunc, ok := registeredDirectives[dir]
-			if !ok {
-				return nil, h.Errf("unrecognized directive: %s - are you sure your Caddyfile structure (nesting and braces) is correct?", dir)
-			}
-
-			subHelper := h
-			subHelper.Dispenser = caddyfile.NewDispenser(seg)
-			subHelper.matcherDefs = matcherDefs
-
-			results, err := dirFunc(subHelper)
-			if err != nil {
-				return nil, h.Errf("parsing caddyfile tokens for '%s': %v", dir, err)
-			}
-
-			dir = normalizeDirectiveName(dir)
-
-			for _, result := range results {
-				result.directive = dir
-				allResults = append(allResults, result)
-			}
-		}
-	}
-
-	return allResults, nil
-}
-
-// ConfigValue represents a value to be added to the final
-// configuration, or a value to be consulted when building
-// the final configuration.
-type ConfigValue struct {
-	// The kind of value this is. As the config is
-	// being built, the adapter will look in the
-	// "pile" for values belonging to a certain
-	// class when it is setting up a certain part
-	// of the config. The associated value will be
-	// type-asserted and placed accordingly.
-	Class string
-
-	// The value to be used when building the config.
-	// Generally its type is associated with the
-	// name of the Class.
-	Value any
-
-	directive string
-}
-
-func sortRoutes(routes []ConfigValue) {
-	dirPositions := make(map[string]int)
-	for i, dir := range directiveOrder {
-		dirPositions[dir] = i
-	}
-
-	sort.SliceStable(routes, func(i, j int) bool {
-		// if the directives are different, just use the established directive order
-		iDir, jDir := routes[i].directive, routes[j].directive
-		if iDir != jDir {
-			return dirPositions[iDir] < dirPositions[jDir]
-		}
-
-		// directives are the same; sub-sort by path matcher length if there's
-		// only one matcher set and one path (this is a very common case and
-		// usually -- but not always -- helpful/expected, oh well; user can
-		// always take manual control of order using handler or route blocks)
-		iRoute, ok := routes[i].Value.(caddyhttp.Route)
-		if !ok {
-			return false
-		}
-		jRoute, ok := routes[j].Value.(caddyhttp.Route)
-		if !ok {
-			return false
-		}
-
-		// decode the path matchers if there is just one matcher set
-		var iPM, jPM caddyhttp.MatchPath
-		if len(iRoute.MatcherSetsRaw) == 1 {
-			_ = json.Unmarshal(iRoute.MatcherSetsRaw[0]["path"], &iPM)
-		}
-		if len(jRoute.MatcherSetsRaw) == 1 {
-			_ = json.Unmarshal(jRoute.MatcherSetsRaw[0]["path"], &jPM)
-		}
-
-		// if there is only one path in the path matcher, sort by longer path
-		// (more specific) first; missing path matchers or multi-matchers are
-		// treated as zero-length paths
-		var iPathLen, jPathLen int
-		if len(iPM) == 1 {
-			iPathLen = len(iPM[0])
-		}
-		if len(jPM) == 1 {
-			jPathLen = len(jPM[0])
-		}
-
-		// some directives involve setting values which can overwrite
-		// each other, so it makes most sense to reverse the order so
-		// that the lease specific matcher is first; everything else
-		// has most-specific matcher first
-		if iDir == "vars" {
-			// we can only confidently compare path lengths if both
-			// directives have a single path to match (issue #5037)
-			if iPathLen > 0 && jPathLen > 0 {
-				// sort least-specific (shortest) path first
-				return iPathLen < jPathLen
-			}
-
-			// if both directives don't have a single path to compare,
-			// sort whichever one has no matcher first; if both have
-			// no matcher, sort equally (stable sort preserves order)
-			return len(iRoute.MatcherSetsRaw) == 0 && len(jRoute.MatcherSetsRaw) > 0
-		} else {
-			// we can only confidently compare path lengths if both
-			// directives have a single path to match (issue #5037)
-			if iPathLen > 0 && jPathLen > 0 {
-				// sort most-specific (longest) path first
-				return iPathLen > jPathLen
-			}
-
-			// if both directives don't have a single path to compare,
-			// sort whichever one has a matcher first; if both have
-			// a matcher, sort equally (stable sort preserves order)
-			return len(iRoute.MatcherSetsRaw) > 0 && len(jRoute.MatcherSetsRaw) == 0
-		}
-	})
-}
-
-// serverBlock pairs a Caddyfile server block with
-// a "pile" of config values, keyed by class name,
-// as well as its parsed keys for convenience.
-type serverBlock struct {
-	block caddyfile.ServerBlock
-	pile  map[string][]ConfigValue // config values obtained from directives
-	keys  []Address
-}
-
-// hostsFromKeys returns a list of all the non-empty hostnames found in
-// the keys of the server block sb. If logger mode is false, a key with
-// an empty hostname portion will return an empty slice, since that
-// server block is interpreted to effectively match all hosts. An empty
-// string is never added to the slice.
-//
-// If loggerMode is true, then the non-standard ports of keys will be
-// joined to the hostnames. This is to effectively match the Host
-// header of requests that come in for that key.
-//
-// The resulting slice is not sorted but will never have duplicates.
-func (sb serverBlock) hostsFromKeys(loggerMode bool) []string {
-	// ensure each entry in our list is unique
-	hostMap := make(map[string]struct{})
-	for _, addr := range sb.keys {
-		if addr.Host == "" {
-			if !loggerMode {
-				// server block contains a key like ":443", i.e. the host portion
-				// is empty / catch-all, which means to match all hosts
-				return []string{}
-			}
-			// never append an empty string
-			continue
-		}
-		if loggerMode &&
-			addr.Port != "" &&
-			addr.Port != strconv.Itoa(caddyhttp.DefaultHTTPPort) &&
-			addr.Port != strconv.Itoa(caddyhttp.DefaultHTTPSPort) {
-			hostMap[net.JoinHostPort(addr.Host, addr.Port)] = struct{}{}
-		} else {
-			hostMap[addr.Host] = struct{}{}
-		}
-	}
-
-	// convert map to slice
-	sblockHosts := make([]string, 0, len(hostMap))
-	for host := range hostMap {
-		sblockHosts = append(sblockHosts, host)
-	}
-
-	return sblockHosts
-}
-
-func (sb serverBlock) hostsFromKeysNotHTTP(httpPort string) []string {
-	// ensure each entry in our list is unique
-	hostMap := make(map[string]struct{})
-	for _, addr := range sb.keys {
-		if addr.Host == "" {
-			continue
-		}
-		if addr.Scheme != "http" && addr.Port != httpPort {
-			hostMap[addr.Host] = struct{}{}
-		}
-	}
-
-	// convert map to slice
-	sblockHosts := make([]string, 0, len(hostMap))
-	for host := range hostMap {
-		sblockHosts = append(sblockHosts, host)
-	}
-
-	return sblockHosts
-}
-
-// hasHostCatchAllKey returns true if sb has a key that
-// omits a host portion, i.e. it "catches all" hosts.
-func (sb serverBlock) hasHostCatchAllKey() bool {
-	for _, addr := range sb.keys {
-		if addr.Host == "" {
-			return true
-		}
-	}
-	return false
-}
-
-// isAllHTTP returns true if all sb keys explicitly specify
-// the http:// scheme
-func (sb serverBlock) isAllHTTP() bool {
-	for _, addr := range sb.keys {
-		if addr.Scheme != "http" {
-			return false
-		}
-	}
-	return true
-}
-
-type (
-	// UnmarshalFunc is a function which can unmarshal Caddyfile
-	// tokens into zero or more config values using a Helper type.
-	// These are passed in a call to RegisterDirective.
-	UnmarshalFunc func(h Helper) ([]ConfigValue, error)
-
-	// UnmarshalHandlerFunc is like UnmarshalFunc, except the
-	// output of the unmarshaling is an HTTP handler. This
-	// function does not need to deal with HTTP request matching
-	// which is abstracted away. Since writing HTTP handlers
-	// with Caddyfile support is very common, this is a more
-	// convenient way to add a handler to the chain since a lot
-	// of the details common to HTTP handlers are taken care of
-	// for you. These are passed to a call to
-	// RegisterHandlerDirective.
-	UnmarshalHandlerFunc func(h Helper) (caddyhttp.MiddlewareHandler, error)
-
-	// UnmarshalGlobalFunc is a function which can unmarshal Caddyfile
-	// tokens from a global option. It is passed the tokens to parse and
-	// existing value from the previous instance of this global option
-	// (if any). It returns the value to associate with this global option.
-	UnmarshalGlobalFunc func(d *caddyfile.Dispenser, existingVal any) (any, error)
-)
-
-var registeredDirectives = make(map[string]UnmarshalFunc)
-
-var registeredGlobalOptions = make(map[string]UnmarshalGlobalFunc)

caddyconfig/httpcaddyfile/directives_test.go

@@ -1,94 +0,0 @@
-package httpcaddyfile
-
-import (
-	"reflect"
-	"sort"
-	"testing"
-)
-
-func TestHostsFromKeys(t *testing.T) {
-	for i, tc := range []struct {
-		keys             []Address
-		expectNormalMode []string
-		expectLoggerMode []string
-	}{
-		{
-			[]Address{
-				{Original: "foo", Host: "foo"},
-			},
-			[]string{"foo"},
-			[]string{"foo"},
-		},
-		{
-			[]Address{
-				{Original: "foo", Host: "foo"},
-				{Original: "bar", Host: "bar"},
-			},
-			[]string{"bar", "foo"},
-			[]string{"bar", "foo"},
-		},
-		{
-			[]Address{
-				{Original: ":2015", Port: "2015"},
-			},
-			[]string{}, []string{},
-		},
-		{
-			[]Address{
-				{Original: ":443", Port: "443"},
-			},
-			[]string{}, []string{},
-		},
-		{
-			[]Address{
-				{Original: "foo", Host: "foo"},
-				{Original: ":2015", Port: "2015"},
-			},
-			[]string{}, []string{"foo"},
-		},
-		{
-			[]Address{
-				{Original: "example.com:2015", Host: "example.com", Port: "2015"},
-			},
-			[]string{"example.com"},
-			[]string{"example.com:2015"},
-		},
-		{
-			[]Address{
-				{Original: "example.com:80", Host: "example.com", Port: "80"},
-			},
-			[]string{"example.com"},
-			[]string{"example.com"},
-		},
-		{
-			[]Address{
-				{Original: "https://:2015/foo", Scheme: "https", Port: "2015", Path: "/foo"},
-			},
-			[]string{},
-			[]string{},
-		},
-		{
-			[]Address{
-				{Original: "https://example.com:2015/foo", Scheme: "https", Host: "example.com", Port: "2015", Path: "/foo"},
-			},
-			[]string{"example.com"},
-			[]string{"example.com:2015"},
-		},
-	} {
-		sb := serverBlock{keys: tc.keys}
-
-		// test in normal mode
-		actual := sb.hostsFromKeys(false)
-		sort.Strings(actual)
-		if !reflect.DeepEqual(tc.expectNormalMode, actual) {
-			t.Errorf("Test %d (loggerMode=false): Expected: %v Actual: %v", i, tc.expectNormalMode, actual)
-		}
-
-		// test in logger mode
-		actual = sb.hostsFromKeys(true)
-		sort.Strings(actual)
-		if !reflect.DeepEqual(tc.expectLoggerMode, actual) {
-			t.Errorf("Test %d (loggerMode=true): Expected: %v Actual: %v", i, tc.expectLoggerMode, actual)
-		}
-	}
-}

caddyconfig/httpcaddyfile/httptype_test.go

@@ -1,211 +0,0 @@
-package httpcaddyfile
-
-import (
-	"testing"
-
-	"github.com/caddyserver/caddy/v2/caddyconfig/caddyfile"
-)
-
-func TestMatcherSyntax(t *testing.T) {
-	for i, tc := range []struct {
-		input       string
-		expectError bool
-	}{
-		{
-			input: `http://localhost
-			@debug {
-				query showdebug=1
-			}
-			`,
-			expectError: false,
-		},
-		{
-			input: `http://localhost
-			@debug {
-				query bad format
-			}
-			`,
-			expectError: true,
-		},
-		{
-			input: `http://localhost
-			@debug {
-				not {
-					path /somepath*
-				}
-			}
-			`,
-			expectError: false,
-		},
-		{
-			input: `http://localhost
-			@debug {
-				not path /somepath*
-			}
-			`,
-			expectError: false,
-		},
-		{
-			input: `http://localhost
-			@debug not path /somepath*
-			`,
-			expectError: false,
-		},
-		{
-			input: `@matcher {
-				path /matcher-not-allowed/outside-of-site-block/*
-			}
-			http://localhost
-			`,
-			expectError: true,
-		},
-	} {
-
-		adapter := caddyfile.Adapter{
-			ServerType: ServerType{},
-		}
-
-		_, _, err := adapter.Adapt([]byte(tc.input), nil)
-
-		if err != nil != tc.expectError {
-			t.Errorf("Test %d error expectation failed Expected: %v, got %s", i, tc.expectError, err)
-			continue
-		}
-	}
-}
-
-func TestSpecificity(t *testing.T) {
-	for i, tc := range []struct {
-		input  string
-		expect int
-	}{
-		{"", 0},
-		{"*", 0},
-		{"*.*", 1},
-		{"{placeholder}", 0},
-		{"/{placeholder}", 1},
-		{"foo", 3},
-		{"example.com", 11},
-		{"a.example.com", 13},
-		{"*.example.com", 12},
-		{"/foo", 4},
-		{"/foo*", 4},
-		{"{placeholder}.example.com", 12},
-		{"{placeholder.example.com", 24},
-		{"}.", 2},
-		{"}{", 2},
-		{"{}", 0},
-		{"{{{}}", 1},
-	} {
-		actual := specificity(tc.input)
-		if actual != tc.expect {
-			t.Errorf("Test %d (%s): Expected %d but got %d", i, tc.input, tc.expect, actual)
-		}
-	}
-}
-
-func TestGlobalOptions(t *testing.T) {
-	for i, tc := range []struct {
-		input       string
-		expectError bool
-	}{
-		{
-			input: `
-				{
-					email test@example.com
-				}
-				:80
-			`,
-			expectError: false,
-		},
-		{
-			input: `
-				{
-					admin off
-				}
-				:80
-			`,
-			expectError: false,
-		},
-		{
-			input: `
-				{
-					admin 127.0.0.1:2020
-				}
-				:80
-			`,
-			expectError: false,
-		},
-		{
-			input: `
-				{
-					admin {
-						disabled false
-					}
-				}
-				:80
-			`,
-			expectError: true,
-		},
-		{
-			input: `
-				{
-					admin {
-						enforce_origin
-						origins 192.168.1.1:2020 127.0.0.1:2020
-					}
-				}
-				:80
-			`,
-			expectError: false,
-		},
-		{
-			input: `
-				{
-					admin 127.0.0.1:2020 {
-						enforce_origin
-						origins 192.168.1.1:2020 127.0.0.1:2020
-					}
-				}
-				:80
-			`,
-			expectError: false,
-		},
-		{
-			input: `
-				{
-					admin 192.168.1.1:2020 127.0.0.1:2020 {
-						enforce_origin
-						origins 192.168.1.1:2020 127.0.0.1:2020
-					}
-				}
-				:80
-			`,
-			expectError: true,
-		},
-		{
-			input: `
-				{
-					admin off {
-						enforce_origin
-						origins 192.168.1.1:2020 127.0.0.1:2020
-					}
-				}
-				:80
-			`,
-			expectError: true,
-		},
-	} {
-
-		adapter := caddyfile.Adapter{
-			ServerType: ServerType{},
-		}
-
-		_, _, err := adapter.Adapt([]byte(tc.input), nil)
-
-		if err != nil != tc.expectError {
-			t.Errorf("Test %d error expectation failed Expected: %v, got %s", i, tc.expectError, err)
-			continue
-		}
-	}
-}

caddyconfig/httpcaddyfile/options.go

@@ -1,472 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package httpcaddyfile
-
-import (
-	"strconv"
-
-	"github.com/caddyserver/caddy/v2"
-	"github.com/caddyserver/caddy/v2/caddyconfig"
-	"github.com/caddyserver/caddy/v2/caddyconfig/caddyfile"
-	"github.com/caddyserver/caddy/v2/modules/caddytls"
-	"github.com/caddyserver/certmagic"
-	"github.com/mholt/acmez/acme"
-)
-
-func init() {
-	RegisterGlobalOption("debug", parseOptTrue)
-	RegisterGlobalOption("http_port", parseOptHTTPPort)
-	RegisterGlobalOption("https_port", parseOptHTTPSPort)
-	RegisterGlobalOption("default_bind", parseOptStringList)
-	RegisterGlobalOption("grace_period", parseOptDuration)
-	RegisterGlobalOption("shutdown_delay", parseOptDuration)
-	RegisterGlobalOption("default_sni", parseOptSingleString)
-	RegisterGlobalOption("order", parseOptOrder)
-	RegisterGlobalOption("storage", parseOptStorage)
-	RegisterGlobalOption("storage_clean_interval", parseOptDuration)
-	RegisterGlobalOption("renew_interval", parseOptDuration)
-	RegisterGlobalOption("ocsp_interval", parseOptDuration)
-	RegisterGlobalOption("acme_ca", parseOptSingleString)
-	RegisterGlobalOption("acme_ca_root", parseOptSingleString)
-	RegisterGlobalOption("acme_dns", parseOptACMEDNS)
-	RegisterGlobalOption("acme_eab", parseOptACMEEAB)
-	RegisterGlobalOption("cert_issuer", parseOptCertIssuer)
-	RegisterGlobalOption("skip_install_trust", parseOptTrue)
-	RegisterGlobalOption("email", parseOptSingleString)
-	RegisterGlobalOption("admin", parseOptAdmin)
-	RegisterGlobalOption("on_demand_tls", parseOptOnDemand)
-	RegisterGlobalOption("local_certs", parseOptTrue)
-	RegisterGlobalOption("key_type", parseOptSingleString)
-	RegisterGlobalOption("auto_https", parseOptAutoHTTPS)
-	RegisterGlobalOption("servers", parseServerOptions)
-	RegisterGlobalOption("ocsp_stapling", parseOCSPStaplingOptions)
-	RegisterGlobalOption("log", parseLogOptions)
-	RegisterGlobalOption("preferred_chains", parseOptPreferredChains)
-}
-
-func parseOptTrue(d *caddyfile.Dispenser, _ any) (any, error) { return true, nil }
-
-func parseOptHTTPPort(d *caddyfile.Dispenser, _ any) (any, error) {
-	var httpPort int
-	for d.Next() {
-		var httpPortStr string
-		if !d.AllArgs(&httpPortStr) {
-			return 0, d.ArgErr()
-		}
-		var err error
-		httpPort, err = strconv.Atoi(httpPortStr)
-		if err != nil {
-			return 0, d.Errf("converting port '%s' to integer value: %v", httpPortStr, err)
-		}
-	}
-	return httpPort, nil
-}
-
-func parseOptHTTPSPort(d *caddyfile.Dispenser, _ any) (any, error) {
-	var httpsPort int
-	for d.Next() {
-		var httpsPortStr string
-		if !d.AllArgs(&httpsPortStr) {
-			return 0, d.ArgErr()
-		}
-		var err error
-		httpsPort, err = strconv.Atoi(httpsPortStr)
-		if err != nil {
-			return 0, d.Errf("converting port '%s' to integer value: %v", httpsPortStr, err)
-		}
-	}
-	return httpsPort, nil
-}
-
-func parseOptOrder(d *caddyfile.Dispenser, _ any) (any, error) {
-	newOrder := directiveOrder
-
-	for d.Next() {
-		// get directive name
-		if !d.Next() {
-			return nil, d.ArgErr()
-		}
-		dirName := d.Val()
-		if _, ok := registeredDirectives[dirName]; !ok {
-			return nil, d.Errf("%s is not a registered directive", dirName)
-		}
-
-		// get positional token
-		if !d.Next() {
-			return nil, d.ArgErr()
-		}
-		pos := d.Val()
-
-		// if directive exists, first remove it
-		for i, d := range newOrder {
-			if d == dirName {
-				newOrder = append(newOrder[:i], newOrder[i+1:]...)
-				break
-			}
-		}
-
-		// act on the positional
-		switch pos {
-		case "first":
-			newOrder = append([]string{dirName}, newOrder...)
-			if d.NextArg() {
-				return nil, d.ArgErr()
-			}
-			directiveOrder = newOrder
-			return newOrder, nil
-		case "last":
-			newOrder = append(newOrder, dirName)
-			if d.NextArg() {
-				return nil, d.ArgErr()
-			}
-			directiveOrder = newOrder
-			return newOrder, nil
-		case "before":
-		case "after":
-		default:
-			return nil, d.Errf("unknown positional '%s'", pos)
-		}
-
-		// get name of other directive
-		if !d.NextArg() {
-			return nil, d.ArgErr()
-		}
-		otherDir := d.Val()
-		if d.NextArg() {
-			return nil, d.ArgErr()
-		}
-
-		// insert directive into proper position
-		for i, d := range newOrder {
-			if d == otherDir {
-				if pos == "before" {
-					newOrder = append(newOrder[:i], append([]string{dirName}, newOrder[i:]...)...)
-				} else if pos == "after" {
-					newOrder = append(newOrder[:i+1], append([]string{dirName}, newOrder[i+1:]...)...)
-				}
-				break
-			}
-		}
-	}
-
-	directiveOrder = newOrder
-
-	return newOrder, nil
-}
-
-func parseOptStorage(d *caddyfile.Dispenser, _ any) (any, error) {
-	if !d.Next() { // consume option name
-		return nil, d.ArgErr()
-	}
-	if !d.Next() { // get storage module name
-		return nil, d.ArgErr()
-	}
-	modID := "caddy.storage." + d.Val()
-	unm, err := caddyfile.UnmarshalModule(d, modID)
-	if err != nil {
-		return nil, err
-	}
-	storage, ok := unm.(caddy.StorageConverter)
-	if !ok {
-		return nil, d.Errf("module %s is not a caddy.StorageConverter", modID)
-	}
-	return storage, nil
-}
-
-func parseOptDuration(d *caddyfile.Dispenser, _ any) (any, error) {
-	if !d.Next() { // consume option name
-		return nil, d.ArgErr()
-	}
-	if !d.Next() { // get duration value
-		return nil, d.ArgErr()
-	}
-	dur, err := caddy.ParseDuration(d.Val())
-	if err != nil {
-		return nil, err
-	}
-	return caddy.Duration(dur), nil
-}
-
-func parseOptACMEDNS(d *caddyfile.Dispenser, _ any) (any, error) {
-	if !d.Next() { // consume option name
-		return nil, d.ArgErr()
-	}
-	if !d.Next() { // get DNS module name
-		return nil, d.ArgErr()
-	}
-	modID := "dns.providers." + d.Val()
-	unm, err := caddyfile.UnmarshalModule(d, modID)
-	if err != nil {
-		return nil, err
-	}
-	prov, ok := unm.(certmagic.ACMEDNSProvider)
-	if !ok {
-		return nil, d.Errf("module %s (%T) is not a certmagic.ACMEDNSProvider", modID, unm)
-	}
-	return prov, nil
-}
-
-func parseOptACMEEAB(d *caddyfile.Dispenser, _ any) (any, error) {
-	eab := new(acme.EAB)
-	for d.Next() {
-		if d.NextArg() {
-			return nil, d.ArgErr()
-		}
-		for nesting := d.Nesting(); d.NextBlock(nesting); {
-			switch d.Val() {
-			case "key_id":
-				if !d.NextArg() {
-					return nil, d.ArgErr()
-				}
-				eab.KeyID = d.Val()
-
-			case "mac_key":
-				if !d.NextArg() {
-					return nil, d.ArgErr()
-				}
-				eab.MACKey = d.Val()
-
-			default:
-				return nil, d.Errf("unrecognized parameter '%s'", d.Val())
-			}
-		}
-	}
-	return eab, nil
-}
-
-func parseOptCertIssuer(d *caddyfile.Dispenser, existing any) (any, error) {
-	var issuers []certmagic.Issuer
-	if existing != nil {
-		issuers = existing.([]certmagic.Issuer)
-	}
-	for d.Next() { // consume option name
-		if !d.Next() { // get issuer module name
-			return nil, d.ArgErr()
-		}
-		modID := "tls.issuance." + d.Val()
-		unm, err := caddyfile.UnmarshalModule(d, modID)
-		if err != nil {
-			return nil, err
-		}
-		iss, ok := unm.(certmagic.Issuer)
-		if !ok {
-			return nil, d.Errf("module %s (%T) is not a certmagic.Issuer", modID, unm)
-		}
-		issuers = append(issuers, iss)
-	}
-	return issuers, nil
-}
-
-func parseOptSingleString(d *caddyfile.Dispenser, _ any) (any, error) {
-	d.Next() // consume parameter name
-	if !d.Next() {
-		return "", d.ArgErr()
-	}
-	val := d.Val()
-	if d.Next() {
-		return "", d.ArgErr()
-	}
-	return val, nil
-}
-
-func parseOptStringList(d *caddyfile.Dispenser, _ any) (any, error) {
-	d.Next() // consume parameter name
-	val := d.RemainingArgs()
-	if len(val) == 0 {
-		return "", d.ArgErr()
-	}
-	return val, nil
-}
-
-func parseOptAdmin(d *caddyfile.Dispenser, _ any) (any, error) {
-	adminCfg := new(caddy.AdminConfig)
-	for d.Next() {
-		if d.NextArg() {
-			listenAddress := d.Val()
-			if listenAddress == "off" {
-				adminCfg.Disabled = true
-				if d.Next() { // Do not accept any remaining options including block
-					return nil, d.Err("No more option is allowed after turning off admin config")
-				}
-			} else {
-				adminCfg.Listen = listenAddress
-				if d.NextArg() { // At most 1 arg is allowed
-					return nil, d.ArgErr()
-				}
-			}
-		}
-		for nesting := d.Nesting(); d.NextBlock(nesting); {
-			switch d.Val() {
-			case "enforce_origin":
-				adminCfg.EnforceOrigin = true
-
-			case "origins":
-				adminCfg.Origins = d.RemainingArgs()
-
-			default:
-				return nil, d.Errf("unrecognized parameter '%s'", d.Val())
-			}
-		}
-	}
-	if adminCfg.Listen == "" && !adminCfg.Disabled {
-		adminCfg.Listen = caddy.DefaultAdminListen
-	}
-	return adminCfg, nil
-}
-
-func parseOptOnDemand(d *caddyfile.Dispenser, _ any) (any, error) {
-	var ond *caddytls.OnDemandConfig
-	for d.Next() {
-		if d.NextArg() {
-			return nil, d.ArgErr()
-		}
-		for nesting := d.Nesting(); d.NextBlock(nesting); {
-			switch d.Val() {
-			case "ask":
-				if !d.NextArg() {
-					return nil, d.ArgErr()
-				}
-				if ond == nil {
-					ond = new(caddytls.OnDemandConfig)
-				}
-				ond.Ask = d.Val()
-
-			case "interval":
-				if !d.NextArg() {
-					return nil, d.ArgErr()
-				}
-				dur, err := caddy.ParseDuration(d.Val())
-				if err != nil {
-					return nil, err
-				}
-				if ond == nil {
-					ond = new(caddytls.OnDemandConfig)
-				}
-				if ond.RateLimit == nil {
-					ond.RateLimit = new(caddytls.RateLimit)
-				}
-				ond.RateLimit.Interval = caddy.Duration(dur)
-
-			case "burst":
-				if !d.NextArg() {
-					return nil, d.ArgErr()
-				}
-				burst, err := strconv.Atoi(d.Val())
-				if err != nil {
-					return nil, err
-				}
-				if ond == nil {
-					ond = new(caddytls.OnDemandConfig)
-				}
-				if ond.RateLimit == nil {
-					ond.RateLimit = new(caddytls.RateLimit)
-				}
-				ond.RateLimit.Burst = burst
-
-			default:
-				return nil, d.Errf("unrecognized parameter '%s'", d.Val())
-			}
-		}
-	}
-	if ond == nil {
-		return nil, d.Err("expected at least one config parameter for on_demand_tls")
-	}
-	return ond, nil
-}
-
-func parseOptAutoHTTPS(d *caddyfile.Dispenser, _ any) (any, error) {
-	d.Next() // consume parameter name
-	if !d.Next() {
-		return "", d.ArgErr()
-	}
-	val := d.Val()
-	if d.Next() {
-		return "", d.ArgErr()
-	}
-	if val != "off" && val != "disable_redirects" && val != "disable_certs" && val != "ignore_loaded_certs" {
-		return "", d.Errf("auto_https must be one of 'off', 'disable_redirects', 'disable_certs', or 'ignore_loaded_certs'")
-	}
-	return val, nil
-}
-
-func parseServerOptions(d *caddyfile.Dispenser, _ any) (any, error) {
-	return unmarshalCaddyfileServerOptions(d)
-}
-
-func parseOCSPStaplingOptions(d *caddyfile.Dispenser, _ any) (any, error) {
-	d.Next() // consume option name
-	var val string
-	if !d.AllArgs(&val) {
-		return nil, d.ArgErr()
-	}
-	if val != "off" {
-		return nil, d.Errf("invalid argument '%s'", val)
-	}
-	return certmagic.OCSPConfig{
-		DisableStapling: val == "off",
-	}, nil
-}
-
-// parseLogOptions parses the global log option. Syntax:
-//
-//	log [name] {
-//	    output  <writer_module> ...
-//	    format  <encoder_module> ...
-//	    level   <level>
-//	    include <namespaces...>
-//	    exclude <namespaces...>
-//	}
-//
-// When the name argument is unspecified, this directive modifies the default
-// logger.
-func parseLogOptions(d *caddyfile.Dispenser, existingVal any) (any, error) {
-	currentNames := make(map[string]struct{})
-	if existingVal != nil {
-		innerVals, ok := existingVal.([]ConfigValue)
-		if !ok {
-			return nil, d.Errf("existing log values of unexpected type: %T", existingVal)
-		}
-		for _, rawVal := range innerVals {
-			val, ok := rawVal.Value.(namedCustomLog)
-			if !ok {
-				return nil, d.Errf("existing log value of unexpected type: %T", existingVal)
-			}
-			currentNames[val.name] = struct{}{}
-		}
-	}
-
-	var warnings []caddyconfig.Warning
-	// Call out the same parser that handles server-specific log configuration.
-	configValues, err := parseLogHelper(
-		Helper{
-			Dispenser: d,
-			warnings:  &warnings,
-		},
-		currentNames,
-	)
-	if err != nil {
-		return nil, err
-	}
-	if len(warnings) > 0 {
-		return nil, d.Errf("warnings found in parsing global log options: %+v", warnings)
-	}
-
-	return configValues, nil
-}
-
-func parseOptPreferredChains(d *caddyfile.Dispenser, _ any) (any, error) {
-	d.Next()
-	return caddytls.ParseCaddyfilePreferredChainsOptions(d)
-}

caddyconfig/httpcaddyfile/options_test.go

@@ -1,64 +0,0 @@
-package httpcaddyfile
-
-import (
-	"testing"
-
-	"github.com/caddyserver/caddy/v2/caddyconfig/caddyfile"
-	_ "github.com/caddyserver/caddy/v2/modules/logging"
-)
-
-func TestGlobalLogOptionSyntax(t *testing.T) {
-	for i, tc := range []struct {
-		input       string
-		output      string
-		expectError bool
-	}{
-		// NOTE: Additional test cases of successful Caddyfile parsing
-		// are present in: caddytest/integration/caddyfile_adapt/
-		{
-			input: `{
-				log default
-			}
-			`,
-			output:      `{}`,
-			expectError: false,
-		},
-		{
-			input: `{
-				log example {
-					output file foo.log
-				}
-				log example {
-					format json
-				}
-			}
-			`,
-			expectError: true,
-		},
-		{
-			input: `{
-				log example /foo {
-					output file foo.log
-				}
-			}
-			`,
-			expectError: true,
-		},
-	} {
-
-		adapter := caddyfile.Adapter{
-			ServerType: ServerType{},
-		}
-
-		out, _, err := adapter.Adapt([]byte(tc.input), nil)
-
-		if err != nil != tc.expectError {
-			t.Errorf("Test %d error expectation failed Expected: %v, got %v", i, tc.expectError, err)
-			continue
-		}
-
-		if string(out) != tc.output {
-			t.Errorf("Test %d error output mismatch Expected: %s, got %s", i, tc.output, out)
-		}
-	}
-}

caddyconfig/httpcaddyfile/pkiapp.go

@@ -1,218 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package httpcaddyfile
-
-import (
-	"github.com/caddyserver/caddy/v2/caddyconfig"
-	"github.com/caddyserver/caddy/v2/caddyconfig/caddyfile"
-	"github.com/caddyserver/caddy/v2/modules/caddypki"
-)
-
-func init() {
-	RegisterGlobalOption("pki", parsePKIApp)
-}
-
-// parsePKIApp parses the global log option. Syntax:
-//
-//     pki {
-//         ca [<id>] {
-//             name            <name>
-//             root_cn         <name>
-//             intermediate_cn <name>
-//             root {
-//                 cert   <path>
-//                 key    <path>
-//                 format <format>
-//             }
-//             intermediate {
-//                 cert   <path>
-//                 key    <path>
-//                 format <format>
-//             }
-//         }
-//     }
-//
-// When the CA ID is unspecified, 'local' is assumed.
-func parsePKIApp(d *caddyfile.Dispenser, existingVal any) (any, error) {
-	pki := &caddypki.PKI{CAs: make(map[string]*caddypki.CA)}
-
-	for d.Next() {
-		for nesting := d.Nesting(); d.NextBlock(nesting); {
-			switch d.Val() {
-			case "ca":
-				pkiCa := new(caddypki.CA)
-				if d.NextArg() {
-					pkiCa.ID = d.Val()
-					if d.NextArg() {
-						return nil, d.ArgErr()
-					}
-				}
-				if pkiCa.ID == "" {
-					pkiCa.ID = caddypki.DefaultCAID
-				}
-
-				for nesting := d.Nesting(); d.NextBlock(nesting); {
-					switch d.Val() {
-					case "name":
-						if !d.NextArg() {
-							return nil, d.ArgErr()
-						}
-						pkiCa.Name = d.Val()
-
-					case "root_cn":
-						if !d.NextArg() {
-							return nil, d.ArgErr()
-						}
-						pkiCa.RootCommonName = d.Val()
-
-					case "intermediate_cn":
-						if !d.NextArg() {
-							return nil, d.ArgErr()
-						}
-						pkiCa.IntermediateCommonName = d.Val()
-
-					case "root":
-						if pkiCa.Root == nil {
-							pkiCa.Root = new(caddypki.KeyPair)
-						}
-						for nesting := d.Nesting(); d.NextBlock(nesting); {
-							switch d.Val() {
-							case "cert":
-								if !d.NextArg() {
-									return nil, d.ArgErr()
-								}
-								pkiCa.Root.Certificate = d.Val()
-
-							case "key":
-								if !d.NextArg() {
-									return nil, d.ArgErr()
-								}
-								pkiCa.Root.PrivateKey = d.Val()
-
-							case "format":
-								if !d.NextArg() {
-									return nil, d.ArgErr()
-								}
-								pkiCa.Root.Format = d.Val()
-
-							default:
-								return nil, d.Errf("unrecognized pki ca root option '%s'", d.Val())
-							}
-						}
-
-					case "intermediate":
-						if pkiCa.Intermediate == nil {
-							pkiCa.Intermediate = new(caddypki.KeyPair)
-						}
-						for nesting := d.Nesting(); d.NextBlock(nesting); {
-							switch d.Val() {
-							case "cert":
-								if !d.NextArg() {
-									return nil, d.ArgErr()
-								}
-								pkiCa.Intermediate.Certificate = d.Val()
-
-							case "key":
-								if !d.NextArg() {
-									return nil, d.ArgErr()
-								}
-								pkiCa.Intermediate.PrivateKey = d.Val()
-
-							case "format":
-								if !d.NextArg() {
-									return nil, d.ArgErr()
-								}
-								pkiCa.Intermediate.Format = d.Val()
-
-							default:
-								return nil, d.Errf("unrecognized pki ca intermediate option '%s'", d.Val())
-							}
-						}
-
-					default:
-						return nil, d.Errf("unrecognized pki ca option '%s'", d.Val())
-					}
-				}
-
-				pki.CAs[pkiCa.ID] = pkiCa
-
-			default:
-				return nil, d.Errf("unrecognized pki option '%s'", d.Val())
-			}
-		}
-	}
-
-	return pki, nil
-}
-
-func (st ServerType) buildPKIApp(
-	pairings []sbAddrAssociation,
-	options map[string]any,
-	warnings []caddyconfig.Warning,
-) (*caddypki.PKI, []caddyconfig.Warning, error) {
-
-	skipInstallTrust := false
-	if _, ok := options["skip_install_trust"]; ok {
-		skipInstallTrust = true
-	}
-	falseBool := false
-
-	// Load the PKI app configured via global options
-	var pkiApp *caddypki.PKI
-	unwrappedPki, ok := options["pki"].(*caddypki.PKI)
-	if ok {
-		pkiApp = unwrappedPki
-	} else {
-		pkiApp = &caddypki.PKI{CAs: make(map[string]*caddypki.CA)}
-	}
-	for _, ca := range pkiApp.CAs {
-		if skipInstallTrust {
-			ca.InstallTrust = &falseBool
-		}
-		pkiApp.CAs[ca.ID] = ca
-	}
-
-	// Add in the CAs configured via directives
-	for _, p := range pairings {
-		for _, sblock := range p.serverBlocks {
-			// find all the CAs that were defined and add them to the app config
-			// i.e. from any "acme_server" directives
-			for _, caCfgValue := range sblock.pile["pki.ca"] {
-				ca := caCfgValue.Value.(*caddypki.CA)
-				if skipInstallTrust {
-					ca.InstallTrust = &falseBool
-				}
-
-				// the CA might already exist from global options, so
-				// don't overwrite it in that case
-				if _, ok := pkiApp.CAs[ca.ID]; !ok {
-					pkiApp.CAs[ca.ID] = ca
-				}
-			}
-		}
-	}
-
-	// if there was no CAs defined in any of the servers,
-	// and we were requested to not install trust, then
-	// add one for the default/local CA to do so
-	if len(pkiApp.CAs) == 0 && skipInstallTrust {
-		ca := new(caddypki.CA)
-		ca.ID = caddypki.DefaultCAID
-		ca.InstallTrust = &falseBool
-		pkiApp.CAs[ca.ID] = ca
-	}
-
-	return pkiApp, warnings, nil
-}

caddyconfig/httpcaddyfile/serveroptions.go

@@ -1,282 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package httpcaddyfile
-
-import (
-	"encoding/json"
-	"fmt"
-
-	"github.com/caddyserver/caddy/v2"
-	"github.com/caddyserver/caddy/v2/caddyconfig"
-	"github.com/caddyserver/caddy/v2/caddyconfig/caddyfile"
-	"github.com/caddyserver/caddy/v2/modules/caddyhttp"
-	"github.com/dustin/go-humanize"
-)
-
-// serverOptions collects server config overrides parsed from Caddyfile global options
-type serverOptions struct {
-	// If set, will only apply these options to servers that contain a
-	// listener address that matches exactly. If empty, will apply to all
-	// servers that were not already matched by another serverOptions.
-	ListenerAddress string
-
-	// These will all map 1:1 to the caddyhttp.Server struct
-	ListenerWrappersRaw  []json.RawMessage
-	ReadTimeout          caddy.Duration
-	ReadHeaderTimeout    caddy.Duration
-	WriteTimeout         caddy.Duration
-	IdleTimeout          caddy.Duration
-	KeepAliveInterval    caddy.Duration
-	MaxHeaderBytes       int
-	Protocols            []string
-	StrictSNIHost        *bool
-	ShouldLogCredentials bool
-	Metrics              *caddyhttp.Metrics
-}
-
-func unmarshalCaddyfileServerOptions(d *caddyfile.Dispenser) (any, error) {
-	serverOpts := serverOptions{}
-	for d.Next() {
-		if d.NextArg() {
-			serverOpts.ListenerAddress = d.Val()
-			if d.NextArg() {
-				return nil, d.ArgErr()
-			}
-		}
-		for nesting := d.Nesting(); d.NextBlock(nesting); {
-			switch d.Val() {
-			case "listener_wrappers":
-				for nesting := d.Nesting(); d.NextBlock(nesting); {
-					modID := "caddy.listeners." + d.Val()
-					unm, err := caddyfile.UnmarshalModule(d, modID)
-					if err != nil {
-						return nil, err
-					}
-					listenerWrapper, ok := unm.(caddy.ListenerWrapper)
-					if !ok {
-						return nil, fmt.Errorf("module %s (%T) is not a listener wrapper", modID, unm)
-					}
-					jsonListenerWrapper := caddyconfig.JSONModuleObject(
-						listenerWrapper,
-						"wrapper",
-						listenerWrapper.(caddy.Module).CaddyModule().ID.Name(),
-						nil,
-					)
-					serverOpts.ListenerWrappersRaw = append(serverOpts.ListenerWrappersRaw, jsonListenerWrapper)
-				}
-
-			case "timeouts":
-				for nesting := d.Nesting(); d.NextBlock(nesting); {
-					switch d.Val() {
-					case "read_body":
-						if !d.NextArg() {
-							return nil, d.ArgErr()
-						}
-						dur, err := caddy.ParseDuration(d.Val())
-						if err != nil {
-							return nil, d.Errf("parsing read_body timeout duration: %v", err)
-						}
-						serverOpts.ReadTimeout = caddy.Duration(dur)
-
-					case "read_header":
-						if !d.NextArg() {
-							return nil, d.ArgErr()
-						}
-						dur, err := caddy.ParseDuration(d.Val())
-						if err != nil {
-							return nil, d.Errf("parsing read_header timeout duration: %v", err)
-						}
-						serverOpts.ReadHeaderTimeout = caddy.Duration(dur)
-
-					case "write":
-						if !d.NextArg() {
-							return nil, d.ArgErr()
-						}
-						dur, err := caddy.ParseDuration(d.Val())
-						if err != nil {
-							return nil, d.Errf("parsing write timeout duration: %v", err)
-						}
-						serverOpts.WriteTimeout = caddy.Duration(dur)
-
-					case "idle":
-						if !d.NextArg() {
-							return nil, d.ArgErr()
-						}
-						dur, err := caddy.ParseDuration(d.Val())
-						if err != nil {
-							return nil, d.Errf("parsing idle timeout duration: %v", err)
-						}
-						serverOpts.IdleTimeout = caddy.Duration(dur)
-
-					default:
-						return nil, d.Errf("unrecognized timeouts option '%s'", d.Val())
-					}
-				}
-			case "keepalive_interval":
-				if !d.NextArg() {
-					return nil, d.ArgErr()
-				}
-				dur, err := caddy.ParseDuration(d.Val())
-				if err != nil {
-					return nil, d.Errf("parsing keepalive interval duration: %v", err)
-				}
-				serverOpts.KeepAliveInterval = caddy.Duration(dur)
-
-			case "max_header_size":
-				var sizeStr string
-				if !d.AllArgs(&sizeStr) {
-					return nil, d.ArgErr()
-				}
-				size, err := humanize.ParseBytes(sizeStr)
-				if err != nil {
-					return nil, d.Errf("parsing max_header_size: %v", err)
-				}
-				serverOpts.MaxHeaderBytes = int(size)
-
-			case "log_credentials":
-				if d.NextArg() {
-					return nil, d.ArgErr()
-				}
-				serverOpts.ShouldLogCredentials = true
-
-			case "protocols":
-				protos := d.RemainingArgs()
-				for _, proto := range protos {
-					if proto != "h1" && proto != "h2" && proto != "h2c" && proto != "h3" {
-						return nil, d.Errf("unknown protocol '%s': expected h1, h2, h2c, or h3", proto)
-					}
-					if sliceContains(serverOpts.Protocols, proto) {
-						return nil, d.Errf("protocol %s specified more than once", proto)
-					}
-					serverOpts.Protocols = append(serverOpts.Protocols, proto)
-				}
-				if d.NextBlock(0) {
-					return nil, d.ArgErr()
-				}
-
-			case "strict_sni_host":
-				if d.NextArg() && d.Val() != "insecure_off" && d.Val() != "on" {
-					return nil, d.Errf("strict_sni_host only supports 'on' or 'insecure_off', got '%s'", d.Val())
-				}
-				boolVal := true
-				if d.Val() == "insecure_off" {
-					boolVal = false
-				}
-				serverOpts.StrictSNIHost = &boolVal
-
-			case "metrics":
-				if d.NextArg() {
-					return nil, d.ArgErr()
-				}
-				if d.NextBlock(0) {
-					return nil, d.ArgErr()
-				}
-				serverOpts.Metrics = new(caddyhttp.Metrics)
-
-			// TODO: DEPRECATED. (August 2022)
-			case "protocol":
-				caddy.Log().Named("caddyfile").Warn("DEPRECATED: protocol sub-option will be removed soon")
-
-				for nesting := d.Nesting(); d.NextBlock(nesting); {
-					switch d.Val() {
-					case "allow_h2c":
-						caddy.Log().Named("caddyfile").Warn("DEPRECATED: allow_h2c will be removed soon; use protocols option instead")
-
-						if d.NextArg() {
-							return nil, d.ArgErr()
-						}
-						if sliceContains(serverOpts.Protocols, "h2c") {
-							return nil, d.Errf("protocol h2c already specified")
-						}
-						serverOpts.Protocols = append(serverOpts.Protocols, "h2c")
-
-					case "strict_sni_host":
-						caddy.Log().Named("caddyfile").Warn("DEPRECATED: protocol > strict_sni_host in this position will be removed soon; move up to the servers block instead")
-
-						if d.NextArg() && d.Val() != "insecure_off" && d.Val() != "on" {
-							return nil, d.Errf("strict_sni_host only supports 'on' or 'insecure_off', got '%s'", d.Val())
-						}
-						boolVal := true
-						if d.Val() == "insecure_off" {
-							boolVal = false
-						}
-						serverOpts.StrictSNIHost = &boolVal
-
-					default:
-						return nil, d.Errf("unrecognized protocol option '%s'", d.Val())
-					}
-				}
-
-			default:
-				return nil, d.Errf("unrecognized servers option '%s'", d.Val())
-			}
-		}
-	}
-	return serverOpts, nil
-}
-
-// applyServerOptions sets the server options on the appropriate servers
-func applyServerOptions(
-	servers map[string]*caddyhttp.Server,
-	options map[string]any,
-	warnings *[]caddyconfig.Warning,
-) error {
-	serverOpts, ok := options["servers"].([]serverOptions)
-	if !ok {
-		return nil
-	}
-
-	for _, server := range servers {
-		// find the options that apply to this server
-		opts := func() *serverOptions {
-			for _, entry := range serverOpts {
-				if entry.ListenerAddress == "" {
-					return &entry
-				}
-				for _, listener := range server.Listen {
-					if entry.ListenerAddress == listener {
-						return &entry
-					}
-				}
-			}
-			return nil
-		}()
-
-		// if none apply, then move to the next server
-		if opts == nil {
-			continue
-		}
-
-		// set all the options
-		server.ListenerWrappersRaw = opts.ListenerWrappersRaw
-		server.ReadTimeout = opts.ReadTimeout
-		server.ReadHeaderTimeout = opts.ReadHeaderTimeout
-		server.WriteTimeout = opts.WriteTimeout
-		server.IdleTimeout = opts.IdleTimeout
-		server.KeepAliveInterval = opts.KeepAliveInterval
-		server.MaxHeaderBytes = opts.MaxHeaderBytes
-		server.Protocols = opts.Protocols
-		server.StrictSNIHost = opts.StrictSNIHost
-		server.Metrics = opts.Metrics
-		if opts.ShouldLogCredentials {
-			if server.Logs == nil {
-				server.Logs = &caddyhttp.ServerLogConfig{}
-			}
-			server.Logs.ShouldLogCredentials = opts.ShouldLogCredentials
-		}
-	}
-
-	return nil
-}

caddyconfig/httpcaddyfile/tlsapp.go

@@ -1,665 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package httpcaddyfile
-
-import (
-	"bytes"
-	"encoding/json"
-	"fmt"
-	"reflect"
-	"sort"
-	"strconv"
-	"strings"
-
-	"github.com/caddyserver/caddy/v2"
-	"github.com/caddyserver/caddy/v2/caddyconfig"
-	"github.com/caddyserver/caddy/v2/modules/caddyhttp"
-	"github.com/caddyserver/caddy/v2/modules/caddytls"
-	"github.com/caddyserver/certmagic"
-	"github.com/mholt/acmez/acme"
-)
-
-func (st ServerType) buildTLSApp(
-	pairings []sbAddrAssociation,
-	options map[string]any,
-	warnings []caddyconfig.Warning,
-) (*caddytls.TLS, []caddyconfig.Warning, error) {
-
-	tlsApp := &caddytls.TLS{CertificatesRaw: make(caddy.ModuleMap)}
-	var certLoaders []caddytls.CertificateLoader
-
-	httpPort := strconv.Itoa(caddyhttp.DefaultHTTPPort)
-	if hp, ok := options["http_port"].(int); ok {
-		httpPort = strconv.Itoa(hp)
-	}
-	httpsPort := strconv.Itoa(caddyhttp.DefaultHTTPSPort)
-	if hsp, ok := options["https_port"].(int); ok {
-		httpsPort = strconv.Itoa(hsp)
-	}
-
-	// count how many server blocks have a TLS-enabled key with
-	// no host, and find all hosts that share a server block with
-	// a hostless key, so that they don't get forgotten/omitted
-	// by auto-HTTPS (since they won't appear in route matchers)
-	var serverBlocksWithTLSHostlessKey int
-	httpsHostsSharedWithHostlessKey := make(map[string]struct{})
-	for _, pair := range pairings {
-		for _, sb := range pair.serverBlocks {
-			for _, addr := range sb.keys {
-				if addr.Host == "" {
-					// this address has no hostname, but if it's explicitly set
-					// to HTTPS, then we need to count it as being TLS-enabled
-					if addr.Scheme == "https" || addr.Port == httpsPort {
-						serverBlocksWithTLSHostlessKey++
-					}
-					// this server block has a hostless key, now
-					// go through and add all the hosts to the set
-					for _, otherAddr := range sb.keys {
-						if otherAddr.Original == addr.Original {
-							continue
-						}
-						if otherAddr.Host != "" && otherAddr.Scheme != "http" && otherAddr.Port != httpPort {
-							httpsHostsSharedWithHostlessKey[otherAddr.Host] = struct{}{}
-						}
-					}
-					break
-				}
-			}
-		}
-	}
-
-	// a catch-all automation policy is used as a "default" for all subjects that
-	// don't have custom configuration explicitly associated with them; this
-	// is only to add if the global settings or defaults are non-empty
-	catchAllAP, err := newBaseAutomationPolicy(options, warnings, false)
-	if err != nil {
-		return nil, warnings, err
-	}
-	if catchAllAP != nil {
-		if tlsApp.Automation == nil {
-			tlsApp.Automation = new(caddytls.AutomationConfig)
-		}
-		tlsApp.Automation.Policies = append(tlsApp.Automation.Policies, catchAllAP)
-	}
-
-	for _, p := range pairings {
-		// avoid setting up TLS automation policies for a server that is HTTP-only
-		if !listenersUseAnyPortOtherThan(p.addresses, httpPort) {
-			continue
-		}
-
-		for _, sblock := range p.serverBlocks {
-			// check the scheme of all the site addresses,
-			// skip building AP if they all had http://
-			if sblock.isAllHTTP() {
-				continue
-			}
-
-			// get values that populate an automation policy for this block
-			ap, err := newBaseAutomationPolicy(options, warnings, true)
-			if err != nil {
-				return nil, warnings, err
-			}
-
-			sblockHosts := sblock.hostsFromKeys(false)
-			if len(sblockHosts) == 0 && catchAllAP != nil {
-				ap = catchAllAP
-			}
-
-			// on-demand tls
-			if _, ok := sblock.pile["tls.on_demand"]; ok {
-				ap.OnDemand = true
-			}
-
-			if keyTypeVals, ok := sblock.pile["tls.key_type"]; ok {
-				ap.KeyType = keyTypeVals[0].Value.(string)
-			}
-
-			// certificate issuers
-			if issuerVals, ok := sblock.pile["tls.cert_issuer"]; ok {
-				var issuers []certmagic.Issuer
-				for _, issuerVal := range issuerVals {
-					issuers = append(issuers, issuerVal.Value.(certmagic.Issuer))
-				}
-				if ap == catchAllAP && !reflect.DeepEqual(ap.Issuers, issuers) {
-					return nil, warnings, fmt.Errorf("automation policy from site block is also default/catch-all policy because of key without hostname, and the two are in conflict: %#v != %#v", ap.Issuers, issuers)
-				}
-				ap.Issuers = issuers
-			}
-
-			// certificate managers
-			if certManagerVals, ok := sblock.pile["tls.cert_manager"]; ok {
-				for _, certManager := range certManagerVals {
-					certGetterName := certManager.Value.(caddy.Module).CaddyModule().ID.Name()
-					ap.ManagersRaw = append(ap.ManagersRaw, caddyconfig.JSONModuleObject(certManager.Value, "via", certGetterName, &warnings))
-				}
-			}
-			// custom bind host
-			for _, cfgVal := range sblock.pile["bind"] {
-				for _, iss := range ap.Issuers {
-					// if an issuer was already configured and it is NOT an ACME issuer,
-					// skip, since we intend to adjust only ACME issuers; ensure we
-					// include any issuer that embeds/wraps an underlying ACME issuer
-					var acmeIssuer *caddytls.ACMEIssuer
-					if acmeWrapper, ok := iss.(acmeCapable); ok {
-						acmeIssuer = acmeWrapper.GetACMEIssuer()
-					}
-					if acmeIssuer == nil {
-						continue
-					}
-
-					// proceed to configure the ACME issuer's bind host, without
-					// overwriting any existing settings
-					if acmeIssuer.Challenges == nil {
-						acmeIssuer.Challenges = new(caddytls.ChallengesConfig)
-					}
-					if acmeIssuer.Challenges.BindHost == "" {
-						// only binding to one host is supported
-						var bindHost string
-						if bindHosts, ok := cfgVal.Value.([]string); ok && len(bindHosts) > 0 {
-							bindHost = bindHosts[0]
-						}
-						acmeIssuer.Challenges.BindHost = bindHost
-					}
-				}
-			}
-
-			// first make sure this block is allowed to create an automation policy;
-			// doing so is forbidden if it has a key with no host (i.e. ":443")
-			// and if there is a different server block that also has a key with no
-			// host -- since a key with no host matches any host, we need its
-			// associated automation policy to have an empty Subjects list, i.e. no
-			// host filter, which is indistinguishable between the two server blocks
-			// because automation is not done in the context of a particular server...
-			// this is an example of a poor mapping from Caddyfile to JSON but that's
-			// the least-leaky abstraction I could figure out
-			if len(sblockHosts) == 0 {
-				if serverBlocksWithTLSHostlessKey > 1 {
-					// this server block and at least one other has a key with no host,
-					// making the two indistinguishable; it is misleading to define such
-					// a policy within one server block since it actually will apply to
-					// others as well
-					return nil, warnings, fmt.Errorf("cannot make a TLS automation policy from a server block that has a host-less address when there are other TLS-enabled server block addresses lacking a host")
-				}
-				if catchAllAP == nil {
-					// this server block has a key with no hosts, but there is not yet
-					// a catch-all automation policy (probably because no global options
-					// were set), so this one becomes it
-					catchAllAP = ap
-				}
-			}
-
-			// associate our new automation policy with this server block's hosts
-			ap.Subjects = sblock.hostsFromKeysNotHTTP(httpPort)
-			sort.Strings(ap.Subjects) // solely for deterministic test results
-
-			// if a combination of public and internal names were given
-			// for this same server block and no issuer was specified, we
-			// need to separate them out in the automation policies so
-			// that the internal names can use the internal issuer and
-			// the other names can use the default/public/ACME issuer
-			var ap2 *caddytls.AutomationPolicy
-			if len(ap.Issuers) == 0 {
-				var internal, external []string
-				for _, s := range ap.Subjects {
-					if !certmagic.SubjectQualifiesForCert(s) {
-						return nil, warnings, fmt.Errorf("subject does not qualify for certificate: '%s'", s)
-					}
-					// we don't use certmagic.SubjectQualifiesForPublicCert() because of one nuance:
-					// names like *.*.tld that may not qualify for a public certificate are actually
-					// fine when used with OnDemand, since OnDemand (currently) does not obtain
-					// wildcards (if it ever does, there will be a separate config option to enable
-					// it that we would need to check here) since the hostname is known at handshake;
-					// and it is unexpected to switch to internal issuer when the user wants to get
-					// regular certificates on-demand for a class of certs like *.*.tld.
-					if subjectQualifiesForPublicCert(ap, s) {
-						external = append(external, s)
-					} else {
-						internal = append(internal, s)
-					}
-				}
-				if len(external) > 0 && len(internal) > 0 {
-					ap.Subjects = external
-					apCopy := *ap
-					ap2 = &apCopy
-					ap2.Subjects = internal
-					ap2.IssuersRaw = []json.RawMessage{caddyconfig.JSONModuleObject(caddytls.InternalIssuer{}, "module", "internal", &warnings)}
-				}
-			}
-			if tlsApp.Automation == nil {
-				tlsApp.Automation = new(caddytls.AutomationConfig)
-			}
-			tlsApp.Automation.Policies = append(tlsApp.Automation.Policies, ap)
-			if ap2 != nil {
-				tlsApp.Automation.Policies = append(tlsApp.Automation.Policies, ap2)
-			}
-
-			// certificate loaders
-			if clVals, ok := sblock.pile["tls.cert_loader"]; ok {
-				for _, clVal := range clVals {
-					certLoaders = append(certLoaders, clVal.Value.(caddytls.CertificateLoader))
-				}
-			}
-		}
-	}
-
-	// group certificate loaders by module name, then add to config
-	if len(certLoaders) > 0 {
-		loadersByName := make(map[string]caddytls.CertificateLoader)
-		for _, cl := range certLoaders {
-			name := caddy.GetModuleName(cl)
-			// ugh... technically, we may have multiple FileLoader and FolderLoader
-			// modules (because the tls directive returns one per occurrence), but
-			// the config structure expects only one instance of each kind of loader
-			// module, so we have to combine them... instead of enumerating each
-			// possible cert loader module in a type switch, we can use reflection,
-			// which works on any cert loaders that are slice types
-			if reflect.TypeOf(cl).Kind() == reflect.Slice {
-				combined := reflect.ValueOf(loadersByName[name])
-				if !combined.IsValid() {
-					combined = reflect.New(reflect.TypeOf(cl)).Elem()
-				}
-				clVal := reflect.ValueOf(cl)
-				for i := 0; i < clVal.Len(); i++ {
-					combined = reflect.Append(combined, clVal.Index(i))
-				}
-				loadersByName[name] = combined.Interface().(caddytls.CertificateLoader)
-			}
-		}
-		for certLoaderName, loaders := range loadersByName {
-			tlsApp.CertificatesRaw[certLoaderName] = caddyconfig.JSON(loaders, &warnings)
-		}
-	}
-
-	// set any of the on-demand options, for if/when on-demand TLS is enabled
-	if onDemand, ok := options["on_demand_tls"].(*caddytls.OnDemandConfig); ok {
-		if tlsApp.Automation == nil {
-			tlsApp.Automation = new(caddytls.AutomationConfig)
-		}
-		tlsApp.Automation.OnDemand = onDemand
-	}
-
-	// set the storage clean interval if configured
-	if storageCleanInterval, ok := options["storage_clean_interval"].(caddy.Duration); ok {
-		if tlsApp.Automation == nil {
-			tlsApp.Automation = new(caddytls.AutomationConfig)
-		}
-		tlsApp.Automation.StorageCleanInterval = storageCleanInterval
-	}
-
-	// set the expired certificates renew interval if configured
-	if renewCheckInterval, ok := options["renew_interval"].(caddy.Duration); ok {
-		if tlsApp.Automation == nil {
-			tlsApp.Automation = new(caddytls.AutomationConfig)
-		}
-		tlsApp.Automation.RenewCheckInterval = renewCheckInterval
-	}
-
-	// set the OCSP check interval if configured
-	if ocspCheckInterval, ok := options["ocsp_interval"].(caddy.Duration); ok {
-		if tlsApp.Automation == nil {
-			tlsApp.Automation = new(caddytls.AutomationConfig)
-		}
-		tlsApp.Automation.OCSPCheckInterval = ocspCheckInterval
-	}
-
-	// set whether OCSP stapling should be disabled for manually-managed certificates
-	if ocspConfig, ok := options["ocsp_stapling"].(certmagic.OCSPConfig); ok {
-		tlsApp.DisableOCSPStapling = ocspConfig.DisableStapling
-	}
-
-	// if any hostnames appear on the same server block as a key with
-	// no host, they will not be used with route matchers because the
-	// hostless key matches all hosts, therefore, it wouldn't be
-	// considered for auto-HTTPS, so we need to make sure those hosts
-	// are manually considered for managed certificates; we also need
-	// to make sure that any of these names which are internal-only
-	// get internal certificates by default rather than ACME
-	var al caddytls.AutomateLoader
-	internalAP := &caddytls.AutomationPolicy{
-		IssuersRaw: []json.RawMessage{json.RawMessage(`{"module":"internal"}`)},
-	}
-	for h := range httpsHostsSharedWithHostlessKey {
-		al = append(al, h)
-		if !certmagic.SubjectQualifiesForPublicCert(h) {
-			internalAP.Subjects = append(internalAP.Subjects, h)
-		}
-	}
-	if len(al) > 0 {
-		tlsApp.CertificatesRaw["automate"] = caddyconfig.JSON(al, &warnings)
-	}
-	if len(internalAP.Subjects) > 0 {
-		if tlsApp.Automation == nil {
-			tlsApp.Automation = new(caddytls.AutomationConfig)
-		}
-		tlsApp.Automation.Policies = append(tlsApp.Automation.Policies, internalAP)
-	}
-
-	// if there are any global options set for issuers (ACME ones in particular), make sure they
-	// take effect in every automation policy that does not have any issuers
-	if tlsApp.Automation != nil {
-		globalEmail := options["email"]
-		globalACMECA := options["acme_ca"]
-		globalACMECARoot := options["acme_ca_root"]
-		globalACMEDNS := options["acme_dns"]
-		globalACMEEAB := options["acme_eab"]
-		globalPreferredChains := options["preferred_chains"]
-		hasGlobalACMEDefaults := globalEmail != nil || globalACMECA != nil || globalACMECARoot != nil || globalACMEDNS != nil || globalACMEEAB != nil || globalPreferredChains != nil
-		if hasGlobalACMEDefaults {
-			for i := 0; i < len(tlsApp.Automation.Policies); i++ {
-				ap := tlsApp.Automation.Policies[i]
-				if len(ap.Issuers) == 0 && automationPolicyHasAllPublicNames(ap) {
-					// for public names, create default issuers which will later be filled in with configured global defaults
-					// (internal names will implicitly use the internal issuer at auto-https time)
-					ap.Issuers = caddytls.DefaultIssuers()
-
-					// if a specific endpoint is configured, can't use multiple default issuers
-					if globalACMECA != nil {
-						if strings.Contains(globalACMECA.(string), "zerossl") {
-							ap.Issuers = []certmagic.Issuer{&caddytls.ZeroSSLIssuer{ACMEIssuer: new(caddytls.ACMEIssuer)}}
-						} else {
-							ap.Issuers = []certmagic.Issuer{new(caddytls.ACMEIssuer)}
-						}
-					}
-				}
-			}
-		}
-	}
-
-	// finalize and verify policies; do cleanup
-	if tlsApp.Automation != nil {
-		for i, ap := range tlsApp.Automation.Policies {
-			// ensure all issuers have global defaults filled in
-			for j, issuer := range ap.Issuers {
-				err := fillInGlobalACMEDefaults(issuer, options)
-				if err != nil {
-					return nil, warnings, fmt.Errorf("filling in global issuer defaults for AP %d, issuer %d: %v", i, j, err)
-				}
-			}
-
-			// encode all issuer values we created, so they will be rendered in the output
-			if len(ap.Issuers) > 0 && ap.IssuersRaw == nil {
-				for _, iss := range ap.Issuers {
-					issuerName := iss.(caddy.Module).CaddyModule().ID.Name()
-					ap.IssuersRaw = append(ap.IssuersRaw, caddyconfig.JSONModuleObject(iss, "module", issuerName, &warnings))
-				}
-			}
-		}
-
-		// consolidate automation policies that are the exact same
-		tlsApp.Automation.Policies = consolidateAutomationPolicies(tlsApp.Automation.Policies)
-
-		// ensure automation policies don't overlap subjects (this should be
-		// an error at provision-time as well, but catch it in the adapt phase
-		// for convenience)
-		automationHostSet := make(map[string]struct{})
-		for _, ap := range tlsApp.Automation.Policies {
-			for _, s := range ap.Subjects {
-				if _, ok := automationHostSet[s]; ok {
-					return nil, warnings, fmt.Errorf("hostname appears in more than one automation policy, making certificate management ambiguous: %s", s)
-				}
-				automationHostSet[s] = struct{}{}
-			}
-		}
-
-		// if nothing remains, remove any excess values to clean up the resulting config
-		if len(tlsApp.Automation.Policies) == 0 {
-			tlsApp.Automation.Policies = nil
-		}
-		if reflect.DeepEqual(tlsApp.Automation, new(caddytls.AutomationConfig)) {
-			tlsApp.Automation = nil
-		}
-	}
-
-	return tlsApp, warnings, nil
-}
-
-type acmeCapable interface{ GetACMEIssuer() *caddytls.ACMEIssuer }
-
-func fillInGlobalACMEDefaults(issuer certmagic.Issuer, options map[string]any) error {
-	acmeWrapper, ok := issuer.(acmeCapable)
-	if !ok {
-		return nil
-	}
-	acmeIssuer := acmeWrapper.GetACMEIssuer()
-	if acmeIssuer == nil {
-		return nil
-	}
-
-	globalEmail := options["email"]
-	globalACMECA := options["acme_ca"]
-	globalACMECARoot := options["acme_ca_root"]
-	globalACMEDNS := options["acme_dns"]
-	globalACMEEAB := options["acme_eab"]
-	globalPreferredChains := options["preferred_chains"]
-
-	if globalEmail != nil && acmeIssuer.Email == "" {
-		acmeIssuer.Email = globalEmail.(string)
-	}
-	if globalACMECA != nil && acmeIssuer.CA == "" {
-		acmeIssuer.CA = globalACMECA.(string)
-	}
-	if globalACMECARoot != nil && !sliceContains(acmeIssuer.TrustedRootsPEMFiles, globalACMECARoot.(string)) {
-		acmeIssuer.TrustedRootsPEMFiles = append(acmeIssuer.TrustedRootsPEMFiles, globalACMECARoot.(string))
-	}
-	if globalACMEDNS != nil && (acmeIssuer.Challenges == nil || acmeIssuer.Challenges.DNS == nil) {
-		acmeIssuer.Challenges = &caddytls.ChallengesConfig{
-			DNS: &caddytls.DNSChallengeConfig{
-				ProviderRaw: caddyconfig.JSONModuleObject(globalACMEDNS, "name", globalACMEDNS.(caddy.Module).CaddyModule().ID.Name(), nil),
-			},
-		}
-	}
-	if globalACMEEAB != nil && acmeIssuer.ExternalAccount == nil {
-		acmeIssuer.ExternalAccount = globalACMEEAB.(*acme.EAB)
-	}
-	if globalPreferredChains != nil && acmeIssuer.PreferredChains == nil {
-		acmeIssuer.PreferredChains = globalPreferredChains.(*caddytls.ChainPreference)
-	}
-	return nil
-}
-
-// newBaseAutomationPolicy returns a new TLS automation policy that gets
-// its values from the global options map. It should be used as the base
-// for any other automation policies. A nil policy (and no error) will be
-// returned if there are no default/global options. However, if always is
-// true, a non-nil value will always be returned (unless there is an error).
-func newBaseAutomationPolicy(options map[string]any, warnings []caddyconfig.Warning, always bool) (*caddytls.AutomationPolicy, error) {
-	issuers, hasIssuers := options["cert_issuer"]
-	_, hasLocalCerts := options["local_certs"]
-	keyType, hasKeyType := options["key_type"]
-	ocspStapling, hasOCSPStapling := options["ocsp_stapling"]
-
-	hasGlobalAutomationOpts := hasIssuers || hasLocalCerts || hasKeyType || hasOCSPStapling
-
-	// if there are no global options related to automation policies
-	// set, then we can just return right away
-	if !hasGlobalAutomationOpts {
-		if always {
-			return new(caddytls.AutomationPolicy), nil
-		}
-		return nil, nil
-	}
-
-	ap := new(caddytls.AutomationPolicy)
-	if hasKeyType {
-		ap.KeyType = keyType.(string)
-	}
-
-	if hasIssuers && hasLocalCerts {
-		return nil, fmt.Errorf("global options are ambiguous: local_certs is confusing when combined with cert_issuer, because local_certs is also a specific kind of issuer")
-	}
-
-	if hasIssuers {
-		ap.Issuers = issuers.([]certmagic.Issuer)
-	} else if hasLocalCerts {
-		ap.Issuers = []certmagic.Issuer{new(caddytls.InternalIssuer)}
-	}
-
-	if hasOCSPStapling {
-		ocspConfig := ocspStapling.(certmagic.OCSPConfig)
-		ap.DisableOCSPStapling = ocspConfig.DisableStapling
-		ap.OCSPOverrides = ocspConfig.ResponderOverrides
-	}
-
-	return ap, nil
-}
-
-// consolidateAutomationPolicies combines automation policies that are the same,
-// for a cleaner overall output.
-func consolidateAutomationPolicies(aps []*caddytls.AutomationPolicy) []*caddytls.AutomationPolicy {
-	// sort from most specific to least specific; we depend on this ordering
-	sort.SliceStable(aps, func(i, j int) bool {
-		if automationPolicyIsSubset(aps[i], aps[j]) {
-			return true
-		}
-		if automationPolicyIsSubset(aps[j], aps[i]) {
-			return false
-		}
-		return len(aps[i].Subjects) > len(aps[j].Subjects)
-	})
-
-	emptyAPCount := 0
-	origLenAPs := len(aps)
-	// compute the number of empty policies (disregarding subjects) - see #4128
-	emptyAP := new(caddytls.AutomationPolicy)
-	for i := 0; i < len(aps); i++ {
-		emptyAP.Subjects = aps[i].Subjects
-		if reflect.DeepEqual(aps[i], emptyAP) {
-			emptyAPCount++
-			if !automationPolicyHasAllPublicNames(aps[i]) {
-				// if this automation policy has internal names, we might as well remove it
-				// so auto-https can implicitly use the internal issuer
-				aps = append(aps[:i], aps[i+1:]...)
-				i--
-			}
-		}
-	}
-	// If all policies are empty, we can return nil, as there is no need to set any policy
-	if emptyAPCount == origLenAPs {
-		return nil
-	}
-
-	// remove or combine duplicate policies
-outer:
-	for i := 0; i < len(aps); i++ {
-		// compare only with next policies; we sorted by specificity so we must not delete earlier policies
-		for j := i + 1; j < len(aps); j++ {
-			// if they're exactly equal in every way, just keep one of them
-			if reflect.DeepEqual(aps[i], aps[j]) {
-				aps = append(aps[:j], aps[j+1:]...)
-				// must re-evaluate current i against next j; can't skip it!
-				// even if i decrements to -1, will be incremented to 0 immediately
-				i--
-				continue outer
-			}
-
-			// if the policy is the same, we can keep just one, but we have
-			// to be careful which one we keep; if only one has any hostnames
-			// defined, then we need to keep the one without any hostnames,
-			// otherwise the one without any subjects (a catch-all) would be
-			// eaten up by the one with subjects; and if both have subjects, we
-			// need to combine their lists
-			if reflect.DeepEqual(aps[i].IssuersRaw, aps[j].IssuersRaw) &&
-				bytes.Equal(aps[i].StorageRaw, aps[j].StorageRaw) &&
-				aps[i].MustStaple == aps[j].MustStaple &&
-				aps[i].KeyType == aps[j].KeyType &&
-				aps[i].OnDemand == aps[j].OnDemand &&
-				aps[i].RenewalWindowRatio == aps[j].RenewalWindowRatio {
-				if len(aps[i].Subjects) > 0 && len(aps[j].Subjects) == 0 {
-					// later policy (at j) has no subjects ("catch-all"), so we can
-					// remove the identical-but-more-specific policy that comes first
-					// AS LONG AS it is not shadowed by another policy before it; e.g.
-					// if policy i is for example.com, policy i+1 is '*.com', and policy
-					// j is catch-all, we cannot remove policy i because that would
-					// cause example.com to be served by the less specific policy for
-					// '*.com', which might be different (yes we've seen this happen)
-					if automationPolicyShadows(i, aps) >= j {
-						aps = append(aps[:i], aps[i+1:]...)
-						i--
-						continue outer
-					}
-				} else {
-					// avoid repeated subjects
-					for _, subj := range aps[j].Subjects {
-						if !sliceContains(aps[i].Subjects, subj) {
-							aps[i].Subjects = append(aps[i].Subjects, subj)
-						}
-					}
-					aps = append(aps[:j], aps[j+1:]...)
-					j--
-				}
-			}
-		}
-	}
-
-	return aps
-}
-
-// automationPolicyIsSubset returns true if a's subjects are a subset
-// of b's subjects.
-func automationPolicyIsSubset(a, b *caddytls.AutomationPolicy) bool {
-	if len(b.Subjects) == 0 {
-		return true
-	}
-	if len(a.Subjects) == 0 {
-		return false
-	}
-	for _, aSubj := range a.Subjects {
-		var inSuperset bool
-		for _, bSubj := range b.Subjects {
-			if certmagic.MatchWildcard(aSubj, bSubj) {
-				inSuperset = true
-				break
-			}
-		}
-		if !inSuperset {
-			return false
-		}
-	}
-	return true
-}
-
-// automationPolicyShadows returns the index of a policy that aps[i] shadows;
-// in other words, for all policies after position i, if that policy covers
-// the same subjects but is less specific, that policy's position is returned,
-// or -1 if no shadowing is found. For example, if policy i is for
-// "foo.example.com" and policy i+2 is for "*.example.com", then i+2 will be
-// returned, since that policy is shadowed by i, which is in front.
-func automationPolicyShadows(i int, aps []*caddytls.AutomationPolicy) int {
-	for j := i + 1; j < len(aps); j++ {
-		if automationPolicyIsSubset(aps[i], aps[j]) {
-			return j
-		}
-	}
-	return -1
-}
-
-// subjectQualifiesForPublicCert is like certmagic.SubjectQualifiesForPublicCert() except
-// that this allows domains with multiple wildcard levels like '*.*.example.com' to qualify
-// if the automation policy has OnDemand enabled (i.e. this function is more lenient).
-func subjectQualifiesForPublicCert(ap *caddytls.AutomationPolicy, subj string) bool {
-	return !certmagic.SubjectIsIP(subj) &&
-		!certmagic.SubjectIsInternal(subj) &&
-		(strings.Count(subj, "*.") < 2 || ap.OnDemand)
-}
-
-func automationPolicyHasAllPublicNames(ap *caddytls.AutomationPolicy) bool {
-	for _, subj := range ap.Subjects {
-		if !subjectQualifiesForPublicCert(ap, subj) {
-			return false
-		}
-	}
-	return true
-}

caddyconfig/httpcaddyfile/tlsapp_test.go

@@ -1,56 +0,0 @@
-package httpcaddyfile
-
-import (
-	"testing"
-
-	"github.com/caddyserver/caddy/v2/modules/caddytls"
-)
-
-func TestAutomationPolicyIsSubset(t *testing.T) {
-	for i, test := range []struct {
-		a, b   []string
-		expect bool
-	}{
-		{
-			a:      []string{"example.com"},
-			b:      []string{},
-			expect: true,
-		},
-		{
-			a:      []string{},
-			b:      []string{"example.com"},
-			expect: false,
-		},
-		{
-			a:      []string{"foo.example.com"},
-			b:      []string{"*.example.com"},
-			expect: true,
-		},
-		{
-			a:      []string{"foo.example.com"},
-			b:      []string{"foo.example.com"},
-			expect: true,
-		},
-		{
-			a:      []string{"foo.example.com"},
-			b:      []string{"example.com"},
-			expect: false,
-		},
-		{
-			a:      []string{"example.com", "foo.example.com"},
-			b:      []string{"*.com", "*.*.com"},
-			expect: true,
-		},
-		{
-			a:      []string{"example.com", "foo.example.com"},
-			b:      []string{"*.com"},
-			expect: false,
-		},
-	} {
-		apA := &caddytls.AutomationPolicy{Subjects: test.a}
-		apB := &caddytls.AutomationPolicy{Subjects: test.b}
-		if actual := automationPolicyIsSubset(apA, apB); actual != test.expect {
-			t.Errorf("Test %d: Expected %t but got %t (A: %v  B: %v)", i, test.expect, actual, test.a, test.b)
-		}
-	}
-}

caddyconfig/httploader.go

@@ -1,173 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyconfig
-
-import (
-	"crypto/tls"
-	"crypto/x509"
-	"fmt"
-	"io"
-	"net/http"
-	"os"
-	"time"
-
-	"github.com/caddyserver/caddy/v2"
-)
-
-func init() {
-	caddy.RegisterModule(HTTPLoader{})
-}
-
-// HTTPLoader can load Caddy configs over HTTP(S). It can adapt the config
-// based on the Content-Type header of the HTTP response.
-type HTTPLoader struct {
-	// The method for the request. Default: GET
-	Method string `json:"method,omitempty"`
-
-	// The URL of the request.
-	URL string `json:"url,omitempty"`
-
-	// HTTP headers to add to the request.
-	Headers http.Header `json:"header,omitempty"`
-
-	// Maximum time allowed for a complete connection and request.
-	Timeout caddy.Duration `json:"timeout,omitempty"`
-
-	TLS *struct {
-		// Present this instance's managed remote identity credentials to the server.
-		UseServerIdentity bool `json:"use_server_identity,omitempty"`
-
-		// PEM-encoded client certificate filename to present to the server.
-		ClientCertificateFile string `json:"client_certificate_file,omitempty"`
-
-		// PEM-encoded key to use with the client certificate.
-		ClientCertificateKeyFile string `json:"client_certificate_key_file,omitempty"`
-
-		// List of PEM-encoded CA certificate files to add to the same trust
-		// store as RootCAPool (or root_ca_pool in the JSON).
-		RootCAPEMFiles []string `json:"root_ca_pem_files,omitempty"`
-	} `json:"tls,omitempty"`
-}
-
-// CaddyModule returns the Caddy module information.
-func (HTTPLoader) CaddyModule() caddy.ModuleInfo {
-	return caddy.ModuleInfo{
-		ID:  "caddy.config_loaders.http",
-		New: func() caddy.Module { return new(HTTPLoader) },
-	}
-}
-
-// LoadConfig loads a Caddy config.
-func (hl HTTPLoader) LoadConfig(ctx caddy.Context) ([]byte, error) {
-	repl := caddy.NewReplacer()
-
-	client, err := hl.makeClient(ctx)
-	if err != nil {
-		return nil, err
-	}
-
-	method := repl.ReplaceAll(hl.Method, "")
-	if method == "" {
-		method = http.MethodGet
-	}
-
-	url := repl.ReplaceAll(hl.URL, "")
-	req, err := http.NewRequest(method, url, nil)
-	if err != nil {
-		return nil, err
-	}
-	for key, vals := range hl.Headers {
-		for _, val := range vals {
-			req.Header.Add(repl.ReplaceAll(key, ""), repl.ReplaceKnown(val, ""))
-		}
-	}
-
-	resp, err := client.Do(req)
-	if err != nil {
-		return nil, err
-	}
-	defer resp.Body.Close()
-	if resp.StatusCode >= 400 {
-		return nil, fmt.Errorf("server responded with HTTP %d", resp.StatusCode)
-	}
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return nil, err
-	}
-
-	result, warnings, err := adaptByContentType(resp.Header.Get("Content-Type"), body)
-	if err != nil {
-		return nil, err
-	}
-	for _, warn := range warnings {
-		ctx.Logger().Warn(warn.String())
-	}
-
-	return result, nil
-}
-
-func (hl HTTPLoader) makeClient(ctx caddy.Context) (*http.Client, error) {
-	client := &http.Client{
-		Timeout: time.Duration(hl.Timeout),
-	}
-
-	if hl.TLS != nil {
-		var tlsConfig *tls.Config
-
-		// client authentication
-		if hl.TLS.UseServerIdentity {
-			certs, err := ctx.IdentityCredentials(ctx.Logger())
-			if err != nil {
-				return nil, fmt.Errorf("getting server identity credentials: %v", err)
-			}
-			if tlsConfig == nil {
-				tlsConfig = new(tls.Config)
-			}
-			tlsConfig.Certificates = certs
-		} else if hl.TLS.ClientCertificateFile != "" && hl.TLS.ClientCertificateKeyFile != "" {
-			cert, err := tls.LoadX509KeyPair(hl.TLS.ClientCertificateFile, hl.TLS.ClientCertificateKeyFile)
-			if err != nil {
-				return nil, err
-			}
-			if tlsConfig == nil {
-				tlsConfig = new(tls.Config)
-			}
-			tlsConfig.Certificates = []tls.Certificate{cert}
-		}
-
-		// trusted server certs
-		if len(hl.TLS.RootCAPEMFiles) > 0 {
-			rootPool := x509.NewCertPool()
-			for _, pemFile := range hl.TLS.RootCAPEMFiles {
-				pemData, err := os.ReadFile(pemFile)
-				if err != nil {
-					return nil, fmt.Errorf("failed reading ca cert: %v", err)
-				}
-				rootPool.AppendCertsFromPEM(pemData)
-			}
-			if tlsConfig == nil {
-				tlsConfig = new(tls.Config)
-			}
-			tlsConfig.RootCAs = rootPool
-		}
-
-		client.Transport = &http.Transport{TLSClientConfig: tlsConfig}
-	}
-
-	return client, nil
-}
-
-var _ caddy.ConfigLoader = (*HTTPLoader)(nil)

caddyconfig/load.go

@@ -1,214 +0,0 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-package caddyconfig
-
-import (
-	"bytes"
-	"encoding/json"
-	"fmt"
-	"io"
-	"mime"
-	"net/http"
-	"strings"
-	"sync"
-
-	"github.com/caddyserver/caddy/v2"
-)
-
-func init() {
-	caddy.RegisterModule(adminLoad{})
-}
-
-// adminLoad is a module that provides the /load endpoint
-// for the Caddy admin API. The only reason it's not baked
-// into the caddy package directly is because of the import
-// of the caddyconfig package for its GetAdapter function.
-// If the caddy package depends on the caddyconfig package,
-// then the caddyconfig package will not be able to import
-// the caddy package, and it can more easily cause backward
-// edges in the dependency tree (i.e. import cycle).
-// Fortunately, the admin API has first-class support for
-// adding endpoints from modules.
-type adminLoad struct{}
-
-// CaddyModule returns the Caddy module information.
-func (adminLoad) CaddyModule() caddy.ModuleInfo {
-	return caddy.ModuleInfo{
-		ID:  "admin.api.load",
-		New: func() caddy.Module { return new(adminLoad) },
-	}
-}
-
-// Routes returns a route for the /load endpoint.
-func (al adminLoad) Routes() []caddy.AdminRoute {
-	return []caddy.AdminRoute{
-		{
-			Pattern: "/load",
-			Handler: caddy.AdminHandlerFunc(al.handleLoad),
-		},
-		{
-			Pattern: "/adapt",
-			Handler: caddy.AdminHandlerFunc(al.handleAdapt),
-		},
-	}
-}
-
-// handleLoad replaces the entire current configuration with
-// a new one provided in the response body. It supports config
-// adapters through the use of the Content-Type header. A
-// config that is identical to the currently-running config
-// will be a no-op unless Cache-Control: must-revalidate is set.
-func (adminLoad) handleLoad(w http.ResponseWriter, r *http.Request) error {
-	if r.Method != http.MethodPost {
-		return caddy.APIError{
-			HTTPStatus: http.StatusMethodNotAllowed,
-			Err:        fmt.Errorf("method not allowed"),
-		}
-	}
-
-	buf := bufPool.Get().(*bytes.Buffer)
-	buf.Reset()
-	defer bufPool.Put(buf)
-
-	_, err := io.Copy(buf, r.Body)
-	if err != nil {
-		return caddy.APIError{
-			HTTPStatus: http.StatusBadRequest,
-			Err:        fmt.Errorf("reading request body: %v", err),
-		}
-	}
-	body := buf.Bytes()
-
-	// if the config is formatted other than Caddy's native
-	// JSON, we need to adapt it before loading it
-	if ctHeader := r.Header.Get("Content-Type"); ctHeader != "" {
-		result, warnings, err := adaptByContentType(ctHeader, body)
-		if err != nil {
-			return caddy.APIError{
-				HTTPStatus: http.StatusBadRequest,
-				Err:        err,
-			}
-		}
-		if len(warnings) > 0 {
-			respBody, err := json.Marshal(warnings)
-			if err != nil {
-				caddy.Log().Named("admin.api.load").Error(err.Error())
-			}
-			_, _ = w.Write(respBody)
-		}
-		body = result
-	}
-
-	forceReload := r.Header.Get("Cache-Control") == "must-revalidate"
-
-	err = caddy.Load(body, forceReload)
-	if err != nil {
-		return caddy.APIError{
-			HTTPStatus: http.StatusBadRequest,
-			Err:        fmt.Errorf("loading config: %v", err),
-		}
-	}
-
-	caddy.Log().Named("admin.api").Info("load complete")
-
-	return nil
-}
-
-// handleAdapt adapts the given Caddy config to JSON and responds with the result.
-func (adminLoad) handleAdapt(w http.ResponseWriter, r *http.Request) error {
-	if r.Method != http.MethodPost {
-		return caddy.APIError{
-			HTTPStatus: http.StatusMethodNotAllowed,
-			Err:        fmt.Errorf("method not allowed"),
-		}
-	}
-
-	buf := bufPool.Get().(*bytes.Buffer)
-	buf.Reset()
-	defer bufPool.Put(buf)
-
-	_, err := io.Copy(buf, r.Body)
-	if err != nil {
-		return caddy.APIError{
-			HTTPStatus: http.StatusBadRequest,
-			Err:        fmt.Errorf("reading request body: %v", err),
-		}
-	}
-
-	result, warnings, err := adaptByContentType(r.Header.Get("Content-Type"), buf.Bytes())
-	if err != nil {
-		return caddy.APIError{
-			HTTPStatus: http.StatusBadRequest,
-			Err:        err,
-		}
-	}
-
-	out := struct {
-		Warnings []Warning       `json:"warnings,omitempty"`
-		Result   json.RawMessage `json:"result"`
-	}{
-		Warnings: warnings,
-		Result:   result,
-	}
-
-	w.Header().Set("Content-Type", "application/json")
-	return json.NewEncoder(w).Encode(out)
-}
-
-// adaptByContentType adapts body to Caddy JSON using the adapter specified by contentType.
-// If contentType is empty or ends with "/json", the input will be returned, as a no-op.
-func adaptByContentType(contentType string, body []byte) ([]byte, []Warning, error) {
-	// assume JSON as the default
-	if contentType == "" {
-		return body, nil, nil
-	}
-
-	ct, _, err := mime.ParseMediaType(contentType)
-	if err != nil {
-		return nil, nil, caddy.APIError{
-			HTTPStatus: http.StatusBadRequest,
-			Err:        fmt.Errorf("invalid Content-Type: %v", err),
-		}
-	}
-
-	// if already JSON, no need to adapt
-	if strings.HasSuffix(ct, "/json") {
-		return body, nil, nil
-	}
-
-	// adapter name should be suffix of MIME type
-	_, adapterName, slashFound := strings.Cut(ct, "/")
-	if !slashFound {
-		return nil, nil, fmt.Errorf("malformed Content-Type")
-	}
-
-	cfgAdapter := GetAdapter(adapterName)
-	if cfgAdapter == nil {
-		return nil, nil, fmt.Errorf("unrecognized config adapter '%s'", adapterName)
-	}
-
-	result, warnings, err := cfgAdapter.Adapt(body, nil)
-	if err != nil {
-		return nil, nil, fmt.Errorf("adapting config using %s adapter: %v", adapterName, err)
-	}
-
-	return result, warnings, nil
-}
-
-var bufPool = sync.Pool{
-	New: func() any {
-		return new(bytes.Buffer)
-	},
-}

caddyfile/dispenser.go

@@ -0,0 +1,251 @@
+package caddyfile
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"strings"
+)
+
+// Dispenser is a type that dispenses tokens, similarly to a lexer,
+// except that it can do so with some notion of structure and has
+// some really convenient methods.
+type Dispenser struct {
+	filename string
+	tokens   []Token
+	cursor   int
+	nesting  int
+}
+
+// NewDispenser returns a Dispenser, ready to use for parsing the given input.
+func NewDispenser(filename string, input io.Reader) Dispenser {
+	return Dispenser{
+		filename: filename,
+		tokens:   allTokens(input),
+		cursor:   -1,
+	}
+}
+
+// NewDispenserTokens returns a Dispenser filled with the given tokens.
+func NewDispenserTokens(filename string, tokens []Token) Dispenser {
+	return Dispenser{
+		filename: filename,
+		tokens:   tokens,
+		cursor:   -1,
+	}
+}
+
+// Next loads the next token. Returns true if a token
+// was loaded; false otherwise. If false, all tokens
+// have been consumed.
+func (d *Dispenser) Next() bool {
+	if d.cursor < len(d.tokens)-1 {
+		d.cursor++
+		return true
+	}
+	return false
+}
+
+// NextArg loads the next token if it is on the same
+// line. Returns true if a token was loaded; false
+// otherwise. If false, all tokens on the line have
+// been consumed. It handles imported tokens correctly.
+func (d *Dispenser) NextArg() bool {
+	if d.cursor < 0 {
+		d.cursor++
+		return true
+	}
+	if d.cursor >= len(d.tokens) {
+		return false
+	}
+	if d.cursor < len(d.tokens)-1 &&
+		d.tokens[d.cursor].File == d.tokens[d.cursor+1].File &&
+		d.tokens[d.cursor].Line+d.numLineBreaks(d.cursor) == d.tokens[d.cursor+1].Line {
+		d.cursor++
+		return true
+	}
+	return false
+}
+
+// NextLine loads the next token only if it is not on the same
+// line as the current token, and returns true if a token was
+// loaded; false otherwise. If false, there is not another token
+// or it is on the same line. It handles imported tokens correctly.
+func (d *Dispenser) NextLine() bool {
+	if d.cursor < 0 {
+		d.cursor++
+		return true
+	}
+	if d.cursor >= len(d.tokens) {
+		return false
+	}
+	if d.cursor < len(d.tokens)-1 &&
+		(d.tokens[d.cursor].File != d.tokens[d.cursor+1].File ||
+			d.tokens[d.cursor].Line+d.numLineBreaks(d.cursor) < d.tokens[d.cursor+1].Line) {
+		d.cursor++
+		return true
+	}
+	return false
+}
+
+// NextBlock can be used as the condition of a for loop
+// to load the next token as long as it opens a block or
+// is already in a block. It returns true if a token was
+// loaded, or false when the block's closing curly brace
+// was loaded and thus the block ended. Nested blocks are
+// not supported.
+func (d *Dispenser) NextBlock() bool {
+	if d.nesting > 0 {
+		d.Next()
+		if d.Val() == "}" {
+			d.nesting--
+			return false
+		}
+		return true
+	}
+	if !d.NextArg() { // block must open on same line
+		return false
+	}
+	if d.Val() != "{" {
+		d.cursor-- // roll back if not opening brace
+		return false
+	}
+	d.Next()
+	if d.Val() == "}" {
+		// Open and then closed right away
+		return false
+	}
+	d.nesting++
+	return true
+}
+
+// IncrNest adds a level of nesting to the dispenser.
+func (d *Dispenser) IncrNest() {
+	d.nesting++
+	return
+}
+
+// Val gets the text of the current token. If there is no token
+// loaded, it returns empty string.
+func (d *Dispenser) Val() string {
+	if d.cursor < 0 || d.cursor >= len(d.tokens) {
+		return ""
+	}
+	return d.tokens[d.cursor].Text
+}
+
+// Line gets the line number of the current token. If there is no token
+// loaded, it returns 0.
+func (d *Dispenser) Line() int {
+	if d.cursor < 0 || d.cursor >= len(d.tokens) {
+		return 0
+	}
+	return d.tokens[d.cursor].Line
+}
+
+// File gets the filename of the current token. If there is no token loaded,
+// it returns the filename originally given when parsing started.
+func (d *Dispenser) File() string {
+	if d.cursor < 0 || d.cursor >= len(d.tokens) {
+		return d.filename
+	}
+	if tokenFilename := d.tokens[d.cursor].File; tokenFilename != "" {
+		return tokenFilename
+	}
+	return d.filename
+}
+
+// Args is a convenience function that loads the next arguments
+// (tokens on the same line) into an arbitrary number of strings
+// pointed to in targets. If there are fewer tokens available
+// than string pointers, the remaining strings will not be changed
+// and false will be returned. If there were enough tokens available
+// to fill the arguments, then true will be returned.
+func (d *Dispenser) Args(targets ...*string) bool {
+	enough := true
+	for i := 0; i < len(targets); i++ {
+		if !d.NextArg() {
+			enough = false
+			break
+		}
+		*targets[i] = d.Val()
+	}
+	return enough
+}
+
+// RemainingArgs loads any more arguments (tokens on the same line)
+// into a slice and returns them. Open curly brace tokens also indicate
+// the end of arguments, and the curly brace is not included in
+// the return value nor is it loaded.
+func (d *Dispenser) RemainingArgs() []string {
+	var args []string
+
+	for d.NextArg() {
+		if d.Val() == "{" {
+			d.cursor--
+			break
+		}
+		args = append(args, d.Val())
+	}
+
+	return args
+}
+
+// ArgErr returns an argument error, meaning that another
+// argument was expected but not found. In other words,
+// a line break or open curly brace was encountered instead of
+// an argument.
+func (d *Dispenser) ArgErr() error {
+	if d.Val() == "{" {
+		return d.Err("Unexpected token '{', expecting argument")
+	}
+	return d.Errf("Wrong argument count or unexpected line ending after '%s'", d.Val())
+}
+
+// SyntaxErr creates a generic syntax error which explains what was
+// found and what was expected.
+func (d *Dispenser) SyntaxErr(expected string) error {
+	msg := fmt.Sprintf("%s:%d - Syntax error: Unexpected token '%s', expecting '%s'", d.File(), d.Line(), d.Val(), expected)
+	return errors.New(msg)
+}
+
+// EOFErr returns an error indicating that the dispenser reached
+// the end of the input when searching for the next token.
+func (d *Dispenser) EOFErr() error {
+	return d.Errf("Unexpected EOF")
+}
+
+// Err generates a custom parse error with a message of msg.
+func (d *Dispenser) Err(msg string) error {
+	msg = fmt.Sprintf("%s:%d - Parse error: %s", d.File(), d.Line(), msg)
+	return errors.New(msg)
+}
+
+// Errf is like Err, but for formatted error messages
+func (d *Dispenser) Errf(format string, args ...interface{}) error {
+	return d.Err(fmt.Sprintf(format, args...))
+}
+
+// numLineBreaks counts how many line breaks are in the token
+// value given by the token index tknIdx. It returns 0 if the
+// token does not exist or there are no line breaks.
+func (d *Dispenser) numLineBreaks(tknIdx int) int {
+	if tknIdx < 0 || tknIdx >= len(d.tokens) {
+		return 0
+	}
+	return strings.Count(d.tokens[tknIdx].Text, "\n")
+}
+
+// isNewLine determines whether the current token is on a different
+// line (higher line number) than the previous token. It handles imported
+// tokens correctly. If there isn't a previous token, it returns true.
+func (d *Dispenser) isNewLine() bool {
+	if d.cursor < 1 {
+		return true
+	}
+	if d.cursor > len(d.tokens)-1 {
+		return false
+	}
+	return d.tokens[d.cursor-1].File != d.tokens[d.cursor].File ||
+		d.tokens[d.cursor-1].Line+d.numLineBreaks(d.cursor-1) < d.tokens[d.cursor].Line
+}

caddyfile/dispenser_test.go

@@ -1,21 +1,6 @@
-// Copyright 2015 Matthew Holt and The Caddy Authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
 package caddyfile
 
 import (
-	"errors"
 	"reflect"
 	"strings"
 	"testing"
@@ -26,7 +11,7 @@ func TestDispenser_Val_Next(t *testing.T) {
 			  dir1 arg1
 			  dir2 arg2 arg3
 			  dir3`
-	d := NewTestDispenser(input)
+	d := NewDispenser("Testfile", strings.NewReader(input))
 
 	if val := d.Val(); val != "" {
 		t.Fatalf("Val(): Should return empty string when no token loaded; got '%s'", val)
@@ -64,7 +49,7 @@ func TestDispenser_NextArg(t *testing.T) {
 	input := `dir1 arg1
 			  dir2 arg2 arg3
 			  dir3`
-	d := NewTestDispenser(input)
+	d := NewDispenser("Testfile", strings.NewReader(input))
 
 	assertNext := func(shouldLoad bool, expectedVal string, expectedCursor int) {
 		if d.Next() != shouldLoad {
@@ -79,7 +64,7 @@ func TestDispenser_NextArg(t *testing.T) {
 	}
 
 	assertNextArg := func(expectedVal string, loadAnother bool, expectedCursor int) {
-		if !d.NextArg() {
+		if d.NextArg() != true {
 			t.Error("NextArg(): Should load next argument but got false instead")
 		}
 		if d.cursor != expectedCursor {
@@ -89,7 +74,7 @@ func TestDispenser_NextArg(t *testing.T) {
 			t.Errorf("Val(): Expected '%s' but got '%s'", expectedVal, val)
 		}
 		if !loadAnother {
-			if d.NextArg() {
+			if d.NextArg() != false {
 				t.Fatalf("NextArg(): Should NOT load another argument, but got true instead (val: '%s')", d.Val())
 			}
 			if d.cursor != expectedCursor {
@@ -111,7 +96,7 @@ func TestDispenser_NextLine(t *testing.T) {
 	input := `host:port
 			  dir1 arg1
 			  dir2 arg2 arg3`
-	d := NewTestDispenser(input)
+	d := NewDispenser("Testfile", strings.NewReader(input))
 
 	assertNextLine := func(shouldLoad bool, expectedVal string, expectedCursor int) {
 		if d.NextLine() != shouldLoad {
@@ -144,10 +129,10 @@ func TestDispenser_NextBlock(t *testing.T) {
 			  }
 			  foobar2 {
 			  }`
-	d := NewTestDispenser(input)
+	d := NewDispenser("Testfile", strings.NewReader(input))
 
 	assertNextBlock := func(shouldLoad bool, expectedCursor, expectedNesting int) {
-		if loaded := d.NextBlock(0); loaded != shouldLoad {
+		if loaded := d.NextBlock(); loaded != shouldLoad {
 			t.Errorf("NextBlock(): Should return %v but got %v", shouldLoad, loaded)
 		}
 		if d.cursor != expectedCursor {
@@ -174,7 +159,7 @@ func TestDispenser_Args(t *testing.T) {
 			  dir2 arg4 arg5
 			  dir3 arg6 arg7
 			  dir4`
-	d := NewTestDispenser(input)
+	d := NewDispenser("Testfile", strings.NewReader(input))
 
 	d.Next() // dir1
 
@@ -241,7 +226,7 @@ func TestDispenser_RemainingArgs(t *testing.T) {
 			  dir2 arg4 arg5
 			  dir3 arg6 { arg7
 			  dir4`
-	d := NewTestDispenser(input)
+	d := NewDispenser("Testfile", strings.NewReader(input))
 
 	d.Next() // dir1
 
@@ -278,7 +263,7 @@ func TestDispenser_ArgErr_Err(t *testing.T) {
 	input := `dir1 {
 			  }
 			  dir2 arg1 arg2`
-	d := NewTestDispenser(input)
+	d := NewDispenser("Testfile", strings.NewReader(input))
 
 	d.cursor = 1 // {
 
@@ -304,10 +289,4 @@ func TestDispenser_ArgErr_Err(t *testing.T) {
 	if !strings.Contains(err.Error(), "foobar") {
 		t.Errorf("Expected error message with custom message in it ('foobar'); got '%v'", err)
 	}
-
-	var ErrBarIsFull = errors.New("bar is full")
-	bookingError := d.Errf("unable to reserve: %w", ErrBarIsFull)
-	if !errors.Is(bookingError, ErrBarIsFull) {
-		t.Errorf("Errf(): should be able to unwrap the error chain")
-	}
 }

caddyfile/json.go

@@ -0,0 +1,184 @@
+package caddyfile
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"sort"
+	"strconv"
+	"strings"
+)
+
+const filename = "Caddyfile"
+
+// ToJSON converts caddyfile to its JSON representation.
+func ToJSON(caddyfile []byte) ([]byte, error) {
+	var j EncodedCaddyfile
+
+	serverBlocks, err := Parse(filename, bytes.NewReader(caddyfile), nil)
+	if err != nil {
+		return nil, err
+	}
+
+	for _, sb := range serverBlocks {
+		block := EncodedServerBlock{
+			Keys: sb.Keys,
+			Body: [][]interface{}{},
+		}
+
+		// Extract directives deterministically by sorting them
+		var directives = make([]string, len(sb.Tokens))
+		for dir := range sb.Tokens {
+			directives = append(directives, dir)
+		}
+		sort.Strings(directives)
+
+		// Convert each directive's tokens into our JSON structure
+		for _, dir := range directives {
+			disp := NewDispenserTokens(filename, sb.Tokens[dir])
+			for disp.Next() {
+				block.Body = append(block.Body, constructLine(&disp))
+			}
+		}
+
+		// tack this block onto the end of the list
+		j = append(j, block)
+	}
+
+	result, err := json.Marshal(j)
+	if err != nil {
+		return nil, err
+	}
+
+	return result, nil
+}
+
+// constructLine transforms tokens into a JSON-encodable structure;
+// but only one line at a time, to be used at the top-level of
+// a server block only (where the first token on each line is a
+// directive) - not to be used at any other nesting level.
+func constructLine(d *Dispenser) []interface{} {
+	var args []interface{}
+
+	args = append(args, d.Val())
+
+	for d.NextArg() {
+		if d.Val() == "{" {
+			args = append(args, constructBlock(d))
+			continue
+		}
+		args = append(args, d.Val())
+	}
+
+	return args
+}
+
+// constructBlock recursively processes tokens into a
+// JSON-encodable structure. To be used in a directive's
+// block. Goes to end of block.
+func constructBlock(d *Dispenser) [][]interface{} {
+	block := [][]interface{}{}
+
+	for d.Next() {
+		if d.Val() == "}" {
+			break
+		}
+		block = append(block, constructLine(d))
+	}
+
+	return block
+}
+
+// FromJSON converts JSON-encoded jsonBytes to Caddyfile text
+func FromJSON(jsonBytes []byte) ([]byte, error) {
+	var j EncodedCaddyfile
+	var result string
+
+	err := json.Unmarshal(jsonBytes, &j)
+	if err != nil {
+		return nil, err
+	}
+
+	for sbPos, sb := range j {
+		if sbPos > 0 {
+			result += "\n\n"
+		}
+		for i, key := range sb.Keys {
+			if i > 0 {
+				result += ", "
+			}
+			//result += standardizeScheme(key)
+			result += key
+		}
+		result += jsonToText(sb.Body, 1)
+	}
+
+	return []byte(result), nil
+}
+
+// jsonToText recursively transforms a scope of JSON into plain
+// Caddyfile text.
+func jsonToText(scope interface{}, depth int) string {
+	var result string
+
+	switch val := scope.(type) {
+	case string:
+		if strings.ContainsAny(val, "\" \n\t\r") {
+			result += `"` + strings.Replace(val, "\"", "\\\"", -1) + `"`
+		} else {
+			result += val
+		}
+	case int:
+		result += strconv.Itoa(val)
+	case float64:
+		result += fmt.Sprintf("%v", val)
+	case bool:
+		result += fmt.Sprintf("%t", val)
+	case [][]interface{}:
+		result += " {\n"
+		for _, arg := range val {
+			result += strings.Repeat("\t", depth) + jsonToText(arg, depth+1) + "\n"
+		}
+		result += strings.Repeat("\t", depth-1) + "}"
+	case []interface{}:
+		for i, v := range val {
+			if block, ok := v.([]interface{}); ok {
+				result += "{\n"
+				for _, arg := range block {
+					result += strings.Repeat("\t", depth) + jsonToText(arg, depth+1) + "\n"
+				}
+				result += strings.Repeat("\t", depth-1) + "}"
+				continue
+			}
+			result += jsonToText(v, depth)
+			if i < len(val)-1 {
+				result += " "
+			}
+		}
+	}
+
+	return result
+}
+
+// TODO: Will this function come in handy somewhere else?
+/*
+// standardizeScheme turns an address like host:https into https://host,
+// or "host:" into "host".
+func standardizeScheme(addr string) string {
+	if hostname, port, err := net.SplitHostPort(addr); err == nil {
+		if port == "http" || port == "https" {
+			addr = port + "://" + hostname
+		}
+	}
+	return strings.TrimSuffix(addr, ":")
+}
+*/
+
+// EncodedCaddyfile encapsulates a slice of EncodedServerBlocks.
+type EncodedCaddyfile []EncodedServerBlock
+
+// EncodedServerBlock represents a server block ripe for encoding.
+type EncodedServerBlock struct {
+	Keys []string        `json:"keys"`
+	Body [][]interface{} `json:"body"`
+}

caddyfile/json_test.go

@@ -0,0 +1,164 @@
+package caddyfile
+
+import "testing"
+
+var tests = []struct {
+	caddyfile, json string
+}{
+	{ // 0
+		caddyfile: `foo {
+	root /bar
+}`,
+		json: `[{"keys":["foo"],"body":[["root","/bar"]]}]`,
+	},
+	{ // 1
+		caddyfile: `host1, host2 {
+	dir {
+		def
+	}
+}`,
+		json: `[{"keys":["host1","host2"],"body":[["dir",[["def"]]]]}]`,
+	},
+	{ // 2
+		caddyfile: `host1, host2 {
+	dir abc {
+		def ghi
+		jkl
+	}
+}`,
+		json: `[{"keys":["host1","host2"],"body":[["dir","abc",[["def","ghi"],["jkl"]]]]}]`,
+	},
+	{ // 3
+		caddyfile: `host1:1234, host2:5678 {
+	dir abc {
+	}
+}`,
+		json: `[{"keys":["host1:1234","host2:5678"],"body":[["dir","abc",[]]]}]`,
+	},
+	{ // 4
+		caddyfile: `host {
+	foo "bar baz"
+}`,
+		json: `[{"keys":["host"],"body":[["foo","bar baz"]]}]`,
+	},
+	{ // 5
+		caddyfile: `host, host:80 {
+	foo "bar \"baz\""
+}`,
+		json: `[{"keys":["host","host:80"],"body":[["foo","bar \"baz\""]]}]`,
+	},
+	{ // 6
+		caddyfile: `host {
+	foo "bar
+baz"
+}`,
+		json: `[{"keys":["host"],"body":[["foo","bar\nbaz"]]}]`,
+	},
+	{ // 7
+		caddyfile: `host {
+	dir 123 4.56 true
+}`,
+		json: `[{"keys":["host"],"body":[["dir","123","4.56","true"]]}]`, // NOTE: I guess we assume numbers and booleans should be encoded as strings...?
+	},
+	{ // 8
+		caddyfile: `http://host, https://host {
+}`,
+		json: `[{"keys":["http://host","https://host"],"body":[]}]`, // hosts in JSON are always host:port format (if port is specified), for consistency
+	},
+	{ // 9
+		caddyfile: `host {
+	dir1 a b
+	dir2 c d
+}`,
+		json: `[{"keys":["host"],"body":[["dir1","a","b"],["dir2","c","d"]]}]`,
+	},
+	{ // 10
+		caddyfile: `host {
+	dir a b
+	dir c d
+}`,
+		json: `[{"keys":["host"],"body":[["dir","a","b"],["dir","c","d"]]}]`,
+	},
+	{ // 11
+		caddyfile: `host {
+	dir1 a b
+	dir2 {
+		c
+		d
+	}
+}`,
+		json: `[{"keys":["host"],"body":[["dir1","a","b"],["dir2",[["c"],["d"]]]]}]`,
+	},
+	{ // 12
+		caddyfile: `host1 {
+	dir1
+}
+
+host2 {
+	dir2
+}`,
+		json: `[{"keys":["host1"],"body":[["dir1"]]},{"keys":["host2"],"body":[["dir2"]]}]`,
+	},
+}
+
+func TestToJSON(t *testing.T) {
+	for i, test := range tests {
+		output, err := ToJSON([]byte(test.caddyfile))
+		if err != nil {
+			t.Errorf("Test %d: %v", i, err)
+		}
+		if string(output) != test.json {
+			t.Errorf("Test %d\nExpected:\n'%s'\nActual:\n'%s'", i, test.json, string(output))
+		}
+	}
+}
+
+func TestFromJSON(t *testing.T) {
+	for i, test := range tests {
+		output, err := FromJSON([]byte(test.json))
+		if err != nil {
+			t.Errorf("Test %d: %v", i, err)
+		}
+		if string(output) != test.caddyfile {
+			t.Errorf("Test %d\nExpected:\n'%s'\nActual:\n'%s'", i, test.caddyfile, string(output))
+		}
+	}
+}
+
+// TODO: Will these tests come in handy somewhere else?
+/*
+func TestStandardizeAddress(t *testing.T) {
+	// host:https should be converted to https://host
+	output, err := ToJSON([]byte(`host:https`))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if expected, actual := `[{"keys":["https://host"],"body":[]}]`, string(output); expected != actual {
+		t.Errorf("Expected:\n'%s'\nActual:\n'%s'", expected, actual)
+	}
+
+	output, err = FromJSON([]byte(`[{"keys":["https://host"],"body":[]}]`))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if expected, actual := "https://host {\n}", string(output); expected != actual {
+		t.Errorf("Expected:\n'%s'\nActual:\n'%s'", expected, actual)
+	}
+
+	// host: should be converted to just host
+	output, err = ToJSON([]byte(`host:`))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if expected, actual := `[{"keys":["host"],"body":[]}]`, string(output); expected != actual {
+		t.Errorf("Expected:\n'%s'\nActual:\n'%s'", expected, actual)
+	}
+	output, err = FromJSON([]byte(`[{"keys":["host:"],"body":[]}]`))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if expected, actual := "host {\n}", string(output); expected != actual {
+		t.Errorf("Expected:\n'%s'\nActual:\n'%s'", expected, actual)
+	}
+}
+*/

caddyfile/lexer.go

@@ -0,0 +1,122 @@
+package caddyfile
+
+import (
+	"bufio"
+	"io"
+	"unicode"
+)
+
+type (
+	// lexer is a utility which can get values, token by
+	// token, from a Reader. A token is a word, and tokens
+	// are separated by whitespace. A word can be enclosed
+	// in quotes if it contains whitespace.
+	lexer struct {
+		reader *bufio.Reader
+		token  Token
+		line   int
+	}
+
+	// Token represents a single parsable unit.
+	Token struct {
+		File string
+		Line int
+		Text string
+	}
+)
+
+// load prepares the lexer to scan an input for tokens.
+func (l *lexer) load(input io.Reader) error {
+	l.reader = bufio.NewReader(input)
+	l.line = 1
+	return nil
+}
+
+// next loads the next token into the lexer.
+// A token is delimited by whitespace, unless
+// the token starts with a quotes character (")
+// in which case the token goes until the closing
+// quotes (the enclosing quotes are not included).
+// Inside quoted strings, quotes may be escaped
+// with a preceding \ character. No other chars
+// may be escaped. The rest of the line is skipped
+// if a "#" character is read in. Returns true if
+// a token was loaded; false otherwise.
+func (l *lexer) next() bool {
+	var val []rune
+	var comment, quoted, escaped bool
+
+	makeToken := func() bool {
+		l.token.Text = string(val)
+		return true
+	}
+
+	for {
+		ch, _, err := l.reader.ReadRune()
+		if err != nil {
+			if len(val) > 0 {
+				return makeToken()
+			}
+			if err == io.EOF {
+				return false
+			}
+			panic(err)
+		}
+
+		if quoted {
+			if !escaped {
+				if ch == '\\' {
+					escaped = true
+					continue
+				} else if ch == '"' {
+					quoted = false
+					return makeToken()
+				}
+			}
+			if ch == '\n' {
+				l.line++
+			}
+			if escaped {
+				// only escape quotes
+				if ch != '"' {
+					val = append(val, '\\')
+				}
+			}
+			val = append(val, ch)
+			escaped = false
+			continue
+		}
+
+		if unicode.IsSpace(ch) {
+			if ch == '\r' {
+				continue
+			}
+			if ch == '\n' {
+				l.line++
+				comment = false
+			}
+			if len(val) > 0 {
+				return makeToken()
+			}
+			continue
+		}
+
+		if ch == '#' {
+			comment = true
+		}
+
+		if comment {
+			continue
+		}
+
+		if len(val) == 0 {
+			l.token = Token{Line: l.line}
+			if ch == '"' {
+				quoted = true
+				continue
+			}
+		}
+
+		val = append(val, ch)
+	}
+}

caddyfile/lexer_test.go

@@ -0,0 +1,165 @@
+package caddyfile
+
+import (
+	"strings"
+	"testing"
+)
+
+type lexerTestCase struct {
+	input    string
+	expected []Token
+}
+
+func TestLexer(t *testing.T) {
+	testCases := []lexerTestCase{
+		{
+			input: `host:123`,
+			expected: []Token{
+				{Line: 1, Text: "host:123"},
+			},
+		},
+		{
+			input: `host:123
+
+					directive`,
+			expected: []Token{
+				{Line: 1, Text: "host:123"},
+				{Line: 3, Text: "directive"},
+			},
+		},
+		{
+			input: `host:123 {
+						directive
+					}`,
+			expected: []Token{
+				{Line: 1, Text: "host:123"},
+				{Line: 1, Text: "{"},
+				{Line: 2, Text: "directive"},
+				{Line: 3, Text: "}"},
+			},
+		},
+		{
+			input: `host:123 { directive }`,
+			expected: []Token{
+				{Line: 1, Text: "host:123"},
+				{Line: 1, Text: "{"},
+				{Line: 1, Text: "directive"},
+				{Line: 1, Text: "}"},
+			},
+		},
+		{
+			input: `host:123 {
+						#comment
+						directive
+						# comment
+						foobar # another comment
+					}`,
+			expected: []Token{
+				{Line: 1, Text: "host:123"},
+				{Line: 1, Text: "{"},
+				{Line: 3, Text: "directive"},
+				{Line: 5, Text: "foobar"},
+				{Line: 6, Text: "}"},
+			},
+		},
+		{
+			input: `a "quoted value" b
+					foobar`,
+			expected: []Token{
+				{Line: 1, Text: "a"},
+				{Line: 1, Text: "quoted value"},
+				{Line: 1, Text: "b"},
+				{Line: 2, Text: "foobar"},
+			},
+		},
+		{
+			input: `A "quoted \"value\" inside" B`,
+			expected: []Token{
+				{Line: 1, Text: "A"},
+				{Line: 1, Text: `quoted "value" inside`},
+				{Line: 1, Text: "B"},
+			},
+		},
+		{
+			input: `"don't\escape"`,
+			expected: []Token{
+				{Line: 1, Text: `don't\escape`},
+			},
+		},
+		{
+			input: `"don't\\escape"`,
+			expected: []Token{
+				{Line: 1, Text: `don't\\escape`},
+			},
+		},
+		{
+			input: `A "quoted value with line
+					break inside" {
+						foobar
+					}`,
+			expected: []Token{
+				{Line: 1, Text: "A"},
+				{Line: 1, Text: "quoted value with line\n\t\t\t\t\tbreak inside"},
+				{Line: 2, Text: "{"},
+				{Line: 3, Text: "foobar"},
+				{Line: 4, Text: "}"},
+			},
+		},
+		{
+			input: `"C:\php\php-cgi.exe"`,
+			expected: []Token{
+				{Line: 1, Text: `C:\php\php-cgi.exe`},
+			},
+		},
+		{
+			input: `empty "" string`,
+			expected: []Token{
+				{Line: 1, Text: `empty`},
+				{Line: 1, Text: ``},
+				{Line: 1, Text: `string`},
+			},
+		},
+		{
+			input: "skip those\r\nCR characters",
+			expected: []Token{
+				{Line: 1, Text: "skip"},
+				{Line: 1, Text: "those"},
+				{Line: 2, Text: "CR"},
+				{Line: 2, Text: "characters"},
+			},
+		},
+	}
+
+	for i, testCase := range testCases {
+		actual := tokenize(testCase.input)
+		lexerCompare(t, i, testCase.expected, actual)
+	}
+}
+
+func tokenize(input string) (tokens []Token) {
+	l := lexer{}
+	l.load(strings.NewReader(input))
+	for l.next() {
+		tokens = append(tokens, l.token)
+	}
+	return
+}
+
+func lexerCompare(t *testing.T, n int, expected, actual []Token) {
+	if len(expected) != len(actual) {
+		t.Errorf("Test case %d: expected %d token(s) but got %d", n, len(expected), len(actual))
+	}
+
+	for i := 0; i < len(actual) && i < len(expected); i++ {
+		if actual[i].Line != expected[i].Line {
+			t.Errorf("Test case %d token %d ('%s'): expected line %d but was line %d",
+				n, i, expected[i].Text, expected[i].Line, actual[i].Line)
+			break
+		}
+		if actual[i].Text != expected[i].Text {
+			t.Errorf("Test case %d token %d: expected text '%s' but was '%s'",
+				n, i, expected[i].Text, actual[i].Text)
+			break
+		}
+	}
+}

caddyfile/parse.go

@@ -0,0 +1,367 @@
+package caddyfile
+
+import (
+	"io"
+	"os"
+	"path/filepath"
+	"strings"
+)
+
+// Parse parses the input just enough to group tokens, in
+// order, by server block. No further parsing is performed.
+// Server blocks are returned in the order in which they appear.
+// Directives that do not appear in validDirectives will cause
+// an error. If you do not want to check for valid directives,
+// pass in nil instead.
+func Parse(filename string, input io.Reader, validDirectives []string) ([]ServerBlock, error) {
+	p := parser{Dispenser: NewDispenser(filename, input), validDirectives: validDirectives}
+	blocks, err := p.parseAll()
+	return blocks, err
+}
+
+// allTokens lexes the entire input, but does not parse it.
+// It returns all the tokens from the input, unstructured
+// and in order.
+func allTokens(input io.Reader) (tokens []Token) {
+	l := new(lexer)
+	l.load(input)
+	for l.next() {
+		tokens = append(tokens, l.token)
+	}
+	return
+}
+
+type parser struct {
+	Dispenser
+	block           ServerBlock // current server block being parsed
+	validDirectives []string    // a directive must be valid or it's an error
+	eof             bool        // if we encounter a valid EOF in a hard place
+}
+
+func (p *parser) parseAll() ([]ServerBlock, error) {
+	var blocks []ServerBlock
+
+	for p.Next() {
+		err := p.parseOne()
+		if err != nil {
+			return blocks, err
+		}
+		if len(p.block.Keys) > 0 {
+			blocks = append(blocks, p.block)
+		}
+	}
+
+	return blocks, nil
+}
+
+func (p *parser) parseOne() error {
+	p.block = ServerBlock{Tokens: make(map[string][]Token)}
+
+	err := p.begin()
+	if err != nil {
+		return err
+	}
+
+	return nil
+}
+
+func (p *parser) begin() error {
+	if len(p.tokens) == 0 {
+		return nil
+	}
+
+	err := p.addresses()
+	if err != nil {
+		return err
+	}
+
+	if p.eof {
+		// this happens if the Caddyfile consists of only
+		// a line of addresses and nothing else
+		return nil
+	}
+
+	err = p.blockContents()
+	if err != nil {
+		return err
+	}
+
+	return nil
+}
+
+func (p *parser) addresses() error {
+	var expectingAnother bool
+
+	for {
+		tkn := replaceEnvVars(p.Val())
+
+		// special case: import directive replaces tokens during parse-time
+		if tkn == "import" && p.isNewLine() {
+			err := p.doImport()
+			if err != nil {
+				return err
+			}
+			continue
+		}
+
+		// Open brace definitely indicates end of addresses
+		if tkn == "{" {
+			if expectingAnother {
+				return p.Errf("Expected another address but had '%s' - check for extra comma", tkn)
+			}
+			break
+		}
+
+		if tkn != "" { // empty token possible if user typed ""
+			// Trailing comma indicates another address will follow, which
+			// may possibly be on the next line
+			if tkn[len(tkn)-1] == ',' {
+				tkn = tkn[:len(tkn)-1]
+				expectingAnother = true
+			} else {
+				expectingAnother = false // but we may still see another one on this line
+			}
+
+			p.block.Keys = append(p.block.Keys, tkn)
+		}
+
+		// Advance token and possibly break out of loop or return error
+		hasNext := p.Next()
+		if expectingAnother && !hasNext {
+			return p.EOFErr()
+		}
+		if !hasNext {
+			p.eof = true
+			break // EOF
+		}
+		if !expectingAnother && p.isNewLine() {
+			break
+		}
+	}
+
+	return nil
+}
+
+func (p *parser) blockContents() error {
+	errOpenCurlyBrace := p.openCurlyBrace()
+	if errOpenCurlyBrace != nil {
+		// single-server configs don't need curly braces
+		p.cursor--
+	}
+
+	err := p.directives()
+	if err != nil {
+		return err
+	}
+
+	// Only look for close curly brace if there was an opening
+	if errOpenCurlyBrace == nil {
+		err = p.closeCurlyBrace()
+		if err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+// directives parses through all the lines for directives
+// and it expects the next token to be the first
+// directive. It goes until EOF or closing curly brace
+// which ends the server block.
+func (p *parser) directives() error {
+	for p.Next() {
+		// end of server block
+		if p.Val() == "}" {
+			break
+		}
+
+		// special case: import directive replaces tokens during parse-time
+		if p.Val() == "import" {
+			err := p.doImport()
+			if err != nil {
+				return err
+			}
+			p.cursor-- // cursor is advanced when we continue, so roll back one more
+			continue
+		}
+
+		// normal case: parse a directive on this line
+		if err := p.directive(); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+// doImport swaps out the import directive and its argument
+// (a total of 2 tokens) with the tokens in the specified file
+// or globbing pattern. When the function returns, the cursor
+// is on the token before where the import directive was. In
+// other words, call Next() to access the first token that was
+// imported.
+func (p *parser) doImport() error {
+	// syntax check
+	if !p.NextArg() {
+		return p.ArgErr()
+	}
+	importPattern := p.Val()
+	if p.NextArg() {
+		return p.Err("Import takes only one argument (glob pattern or file)")
+	}
+
+	// do glob
+	matches, err := filepath.Glob(importPattern)
+	if err != nil {
+		return p.Errf("Failed to use import pattern %s: %v", importPattern, err)
+	}
+	if len(matches) == 0 {
+		return p.Errf("No files matching import pattern %s", importPattern)
+	}
+
+	// splice out the import directive and its argument (2 tokens total)
+	tokensBefore := p.tokens[:p.cursor-1]
+	tokensAfter := p.tokens[p.cursor+1:]
+
+	// collect all the imported tokens
+	var importedTokens []Token
+	for _, importFile := range matches {
+		newTokens, err := p.doSingleImport(importFile)
+		if err != nil {
+			return err
+		}
+		importedTokens = append(importedTokens, newTokens...)
+	}
+
+	// splice the imported tokens in the place of the import statement
+	// and rewind cursor so Next() will land on first imported token
+	p.tokens = append(tokensBefore, append(importedTokens, tokensAfter...)...)
+	p.cursor--
+
+	return nil
+}
+
+// doSingleImport lexes the individual file at importFile and returns
+// its tokens or an error, if any.
+func (p *parser) doSingleImport(importFile string) ([]Token, error) {
+	file, err := os.Open(importFile)
+	if err != nil {
+		return nil, p.Errf("Could not import %s: %v", importFile, err)
+	}
+	defer file.Close()
+	importedTokens := allTokens(file)
+
+	// Tack the filename onto these tokens so errors show the imported file's name
+	filename := filepath.Base(importFile)
+	for i := 0; i < len(importedTokens); i++ {
+		importedTokens[i].File = filename
+	}
+
+	return importedTokens, nil
+}
+
+// directive collects tokens until the directive's scope
+// closes (either end of line or end of curly brace block).
+// It expects the currently-loaded token to be a directive
+// (or } that ends a server block). The collected tokens
+// are loaded into the current server block for later use
+// by directive setup functions.
+func (p *parser) directive() error {
+	dir := p.Val()
+	nesting := 0
+
+	// TODO: More helpful error message ("did you mean..." or "maybe you need to install its server type")
+	if !p.validDirective(dir) {
+		return p.Errf("Unknown directive '%s'", dir)
+	}
+
+	// The directive itself is appended as a relevant token
+	p.block.Tokens[dir] = append(p.block.Tokens[dir], p.tokens[p.cursor])
+
+	for p.Next() {
+		if p.Val() == "{" {
+			nesting++
+		} else if p.isNewLine() && nesting == 0 {
+			p.cursor-- // read too far
+			break
+		} else if p.Val() == "}" && nesting > 0 {
+			nesting--
+		} else if p.Val() == "}" && nesting == 0 {
+			return p.Err("Unexpected '}' because no matching opening brace")
+		}
+		p.tokens[p.cursor].Text = replaceEnvVars(p.tokens[p.cursor].Text)
+		p.block.Tokens[dir] = append(p.block.Tokens[dir], p.tokens[p.cursor])
+	}
+
+	if nesting > 0 {
+		return p.EOFErr()
+	}
+	return nil
+}
+
+// openCurlyBrace expects the current token to be an
+// opening curly brace. This acts like an assertion
+// because it returns an error if the token is not
+// a opening curly brace. It does NOT advance the token.
+func (p *parser) openCurlyBrace() error {
+	if p.Val() != "{" {
+		return p.SyntaxErr("{")
+	}
+	return nil
+}
+
+// closeCurlyBrace expects the current token to be
+// a closing curly brace. This acts like an assertion
+// because it returns an error if the token is not
+// a closing curly brace. It does NOT advance the token.
+func (p *parser) closeCurlyBrace() error {
+	if p.Val() != "}" {
+		return p.SyntaxErr("}")
+	}
+	return nil
+}
+
+// validDirective returns true if dir is in p.validDirectives.
+func (p *parser) validDirective(dir string) bool {
+	if p.validDirectives == nil {
+		return true
+	}
+	for _, d := range p.validDirectives {
+		if d == dir {
+			return true
+		}
+	}
+	return false
+}
+
+// replaceEnvVars replaces environment variables that appear in the token
+// and understands both the $UNIX and %WINDOWS% syntaxes.
+func replaceEnvVars(s string) string {
+	s = replaceEnvReferences(s, "{%", "%}")
+	s = replaceEnvReferences(s, "{$", "}")
+	return s
+}
+
+// replaceEnvReferences performs the actual replacement of env variables
+// in s, given the placeholder start and placeholder end strings.
+func replaceEnvReferences(s, refStart, refEnd string) string {
+	index := strings.Index(s, refStart)
+	for index != -1 {
+		endIndex := strings.Index(s, refEnd)
+		if endIndex != -1 {
+			ref := s[index : endIndex+len(refEnd)]
+			s = strings.Replace(s, ref, os.Getenv(ref[len(refStart):len(ref)-len(refEnd)]), -1)
+		} else {
+			return s
+		}
+		index = strings.Index(s, refStart)
+	}
+	return s
+}
+
+// ServerBlock associates any number of keys (usually addresses
+// of some sort) with tokens (grouped by directive name).
+type ServerBlock struct {
+	Keys   []string
+	Tokens map[string][]Token
+}

caddyfile/parse_test.go

@@ -0,0 +1,399 @@
+package caddyfile
+
+import (
+	"os"
+	"strings"
+	"testing"
+)
+
+func TestAllTokens(t *testing.T) {
+	input := strings.NewReader("a b c\nd e")
+	expected := []string{"a", "b", "c", "d", "e"}
+	tokens := allTokens(input)
+
+	if len(tokens) != len(expected) {
+		t.Fatalf("Expected %d tokens, got %d", len(expected), len(tokens))
+	}
+
+	for i, val := range expected {
+		if tokens[i].Text != val {
+			t.Errorf("Token %d should be '%s' but was '%s'", i, val, tokens[i].Text)
+		}
+	}
+}
+
+func TestParseOneAndImport(t *testing.T) {
+	testParseOne := func(input string) (ServerBlock, error) {
+		p := testParser(input)
+		p.Next() // parseOne doesn't call Next() to start, so we must
+		err := p.parseOne()
+		return p.block, err
+	}
+
+	for i, test := range []struct {
+		input     string
+		shouldErr bool
+		keys      []string
+		tokens    map[string]int // map of directive name to number of tokens expected
+	}{
+		{`localhost`, false, []string{
+			"localhost",
+		}, map[string]int{}},
+
+		{`localhost
+		  dir1`, false, []string{
+			"localhost",
+		}, map[string]int{
+			"dir1": 1,
+		}},
+
+		{`localhost:1234
+		  dir1 foo bar`, false, []string{
+			"localhost:1234",
+		}, map[string]int{
+			"dir1": 3,
+		}},
+
+		{`localhost {
+		    dir1
+		  }`, false, []string{
+			"localhost",
+		}, map[string]int{
+			"dir1": 1,
+		}},
+
+		{`localhost:1234 {
+		    dir1 foo bar
+		    dir2
+		  }`, false, []string{
+			"localhost:1234",
+		}, map[string]int{
+			"dir1": 3,
+			"dir2": 1,
+		}},
+
+		{`http://localhost https://localhost
+		  dir1 foo bar`, false, []string{
+			"http://localhost",
+			"https://localhost",
+		}, map[string]int{
+			"dir1": 3,
+		}},
+
+		{`http://localhost https://localhost {
+		    dir1 foo bar
+		  }`, false, []string{
+			"http://localhost",
+			"https://localhost",
+		}, map[string]int{
+			"dir1": 3,
+		}},
+
+		{`http://localhost, https://localhost {
+		    dir1 foo bar
+		  }`, false, []string{
+			"http://localhost",
+			"https://localhost",
+		}, map[string]int{
+			"dir1": 3,
+		}},
+
+		{`http://localhost, {
+		  }`, true, []string{
+			"http://localhost",
+		}, map[string]int{}},
+
+		{`host1:80, http://host2.com
+		  dir1 foo bar
+		  dir2 baz`, false, []string{
+			"host1:80",
+			"http://host2.com",
+		}, map[string]int{
+			"dir1": 3,
+			"dir2": 2,
+		}},
+
+		{`http://host1.com,
+		  http://host2.com,
+		  https://host3.com`, false, []string{
+			"http://host1.com",
+			"http://host2.com",
+			"https://host3.com",
+		}, map[string]int{}},
+
+		{`http://host1.com:1234, https://host2.com
+		  dir1 foo {
+		    bar baz
+		  }
+		  dir2`, false, []string{
+			"http://host1.com:1234",
+			"https://host2.com",
+		}, map[string]int{
+			"dir1": 6,
+			"dir2": 1,
+		}},
+
+		{`127.0.0.1
+		  dir1 {
+		    bar baz
+		  }
+		  dir2 {
+		    foo bar
+		  }`, false, []string{
+			"127.0.0.1",
+		}, map[string]int{
+			"dir1": 5,
+			"dir2": 5,
+		}},
+
+		{`localhost
+		  dir1 {
+		    foo`, true, []string{
+			"localhost",
+		}, map[string]int{
+			"dir1": 3,
+		}},
+
+		{`localhost
+		  dir1 {
+		  }`, false, []string{
+			"localhost",
+		}, map[string]int{
+			"dir1": 3,
+		}},
+
+		{`localhost
+		  dir1 {
+		  } }`, true, []string{
+			"localhost",
+		}, map[string]int{
+			"dir1": 3,
+		}},
+
+		{`localhost
+		  dir1 {
+		    nested {
+		      foo
+		    }
+		  }
+		  dir2 foo bar`, false, []string{
+			"localhost",
+		}, map[string]int{
+			"dir1": 7,
+			"dir2": 3,
+		}},
+
+		{``, false, []string{}, map[string]int{}},
+
+		{`localhost
+		  dir1 arg1
+		  import testdata/import_test1.txt`, false, []string{
+			"localhost",
+		}, map[string]int{
+			"dir1": 2,
+			"dir2": 3,
+			"dir3": 1,
+		}},
+
+		{`import testdata/import_test2.txt`, false, []string{
+			"host1",
+		}, map[string]int{
+			"dir1": 1,
+			"dir2": 2,
+		}},
+
+		{`import testdata/import_test1.txt testdata/import_test2.txt`, true, []string{}, map[string]int{}},
+
+		{`import testdata/not_found.txt`, true, []string{}, map[string]int{}},
+
+		{`""`, false, []string{}, map[string]int{}},
+
+		{``, false, []string{}, map[string]int{}},
+	} {
+		result, err := testParseOne(test.input)
+
+		if test.shouldErr && err == nil {
+			t.Errorf("Test %d: Expected an error, but didn't get one", i)
+		}
+		if !test.shouldErr && err != nil {
+			t.Errorf("Test %d: Expected no error, but got: %v", i, err)
+		}
+
+		if len(result.Keys) != len(test.keys) {
+			t.Errorf("Test %d: Expected %d keys, got %d",
+				i, len(test.keys), len(result.Keys))
+			continue
+		}
+		for j, addr := range result.Keys {
+			if addr != test.keys[j] {
+				t.Errorf("Test %d, key %d: Expected '%s', but was '%s'",
+					i, j, test.keys[j], addr)
+			}
+		}
+
+		if len(result.Tokens) != len(test.tokens) {
+			t.Errorf("Test %d: Expected %d directives, had %d",
+				i, len(test.tokens), len(result.Tokens))
+			continue
+		}
+		for directive, tokens := range result.Tokens {
+			if len(tokens) != test.tokens[directive] {
+				t.Errorf("Test %d, directive '%s': Expected %d tokens, counted %d",
+					i, directive, test.tokens[directive], len(tokens))
+				continue
+			}
+		}
+	}
+}
+
+func TestParseAll(t *testing.T) {
+	for i, test := range []struct {
+		input     string
+		shouldErr bool
+		keys      [][]string // keys per server block, in order
+	}{
+		{`localhost`, false, [][]string{
+			{"localhost"},
+		}},
+
+		{`localhost:1234`, false, [][]string{
+			{"localhost:1234"},
+		}},
+
+		{`localhost:1234 {
+		  }
+		  localhost:2015 {
+		  }`, false, [][]string{
+			{"localhost:1234"},
+			{"localhost:2015"},
+		}},
+
+		{`localhost:1234, http://host2`, false, [][]string{
+			{"localhost:1234", "http://host2"},
+		}},
+
+		{`localhost:1234, http://host2,`, true, [][]string{}},
+
+		{`http://host1.com, http://host2.com {
+		  }
+		  https://host3.com, https://host4.com {
+		  }`, false, [][]string{
+			{"http://host1.com", "http://host2.com"},
+			{"https://host3.com", "https://host4.com"},
+		}},
+
+		{`import testdata/import_glob*.txt`, false, [][]string{
+			{"glob0.host0"},
+			{"glob0.host1"},
+			{"glob1.host0"},
+			{"glob2.host0"},
+		}},
+	} {
+		p := testParser(test.input)
+		blocks, err := p.parseAll()
+
+		if test.shouldErr && err == nil {
+			t.Errorf("Test %d: Expected an error, but didn't get one", i)
+		}
+		if !test.shouldErr && err != nil {
+			t.Errorf("Test %d: Expected no error, but got: %v", i, err)
+		}
+
+		if len(blocks) != len(test.keys) {
+			t.Errorf("Test %d: Expected %d server blocks, got %d",
+				i, len(test.keys), len(blocks))
+			continue
+		}
+		for j, block := range blocks {
+			if len(block.Keys) != len(test.keys[j]) {
+				t.Errorf("Test %d: Expected %d keys in block %d, got %d",
+					i, len(test.keys[j]), j, len(block.Keys))
+				continue
+			}
+			for k, addr := range block.Keys {
+				if addr != test.keys[j][k] {
+					t.Errorf("Test %d, block %d, key %d: Expected '%s', but got '%s'",
+						i, j, k, test.keys[j][k], addr)
+				}
+			}
+		}
+	}
+}
+
+func TestEnvironmentReplacement(t *testing.T) {
+	os.Setenv("PORT", "8080")
+	os.Setenv("ADDRESS", "servername.com")
+	os.Setenv("FOOBAR", "foobar")
+
+	// basic test; unix-style env vars
+	p := testParser(`{$ADDRESS}`)
+	blocks, _ := p.parseAll()
+	if actual, expected := blocks[0].Keys[0], "servername.com"; expected != actual {
+		t.Errorf("Expected key to be '%s' but was '%s'", expected, actual)
+	}
+
+	// multiple vars per token
+	p = testParser(`{$ADDRESS}:{$PORT}`)
+	blocks, _ = p.parseAll()
+	if actual, expected := blocks[0].Keys[0], "servername.com:8080"; expected != actual {
+		t.Errorf("Expected key to be '%s' but was '%s'", expected, actual)
+	}
+
+	// windows-style var and unix style in same token
+	p = testParser(`{%ADDRESS%}:{$PORT}`)
+	blocks, _ = p.parseAll()
+	if actual, expected := blocks[0].Keys[0], "servername.com:8080"; expected != actual {
+		t.Errorf("Expected key to be '%s' but was '%s'", expected, actual)
+	}
+
+	// reverse order
+	p = testParser(`{$ADDRESS}:{%PORT%}`)
+	blocks, _ = p.parseAll()
+	if actual, expected := blocks[0].Keys[0], "servername.com:8080"; expected != actual {
+		t.Errorf("Expected key to be '%s' but was '%s'", expected, actual)
+	}
+
+	// env var in server block body as argument
+	p = testParser(":{%PORT%}\ndir1 {$FOOBAR}")
+	blocks, _ = p.parseAll()
+	if actual, expected := blocks[0].Keys[0], ":8080"; expected != actual {
+		t.Errorf("Expected key to be '%s' but was '%s'", expected, actual)
+	}
+	if actual, expected := blocks[0].Tokens["dir1"][1].Text, "foobar"; expected != actual {
+		t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual)
+	}
+
+	// combined windows env vars in argument
+	p = testParser(":{%PORT%}\ndir1 {%ADDRESS%}/{%FOOBAR%}")
+	blocks, _ = p.parseAll()
+	if actual, expected := blocks[0].Tokens["dir1"][1].Text, "servername.com/foobar"; expected != actual {
+		t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual)
+	}
+
+	// malformed env var (windows)
+	p = testParser(":1234\ndir1 {%ADDRESS}")
+	blocks, _ = p.parseAll()
+	if actual, expected := blocks[0].Tokens["dir1"][1].Text, "{%ADDRESS}"; expected != actual {
+		t.Errorf("Expected host to be '%s' but was '%s'", expected, actual)
+	}
+
+	// malformed (non-existent) env var (unix)
+	p = testParser(`:{$PORT$}`)
+	blocks, _ = p.parseAll()
+	if actual, expected := blocks[0].Keys[0], ":"; expected != actual {
+		t.Errorf("Expected key to be '%s' but was '%s'", expected, actual)
+	}
+
+	// in quoted field
+	p = testParser(":1234\ndir1 \"Test {$FOOBAR} test\"")
+	blocks, _ = p.parseAll()
+	if actual, expected := blocks[0].Tokens["dir1"][1].Text, "Test foobar test"; expected != actual {
+		t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual)
+	}
+}
+
+func testParser(input string) parser {
+	buf := strings.NewReader(input)
+	p := parser{Dispenser: NewDispenser("Test", buf)}
+	return p
+}

caddyhttp/basicauth/basicauth.go

@@ -0,0 +1,148 @@
+// Package basicauth implements HTTP Basic Authentication.
+package basicauth
+
+import (
+	"bufio"
+	"crypto/subtle"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"path/filepath"
+	"strings"
+	"sync"
+
+	"github.com/jimstudt/http-authentication/basic"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+// BasicAuth is middleware to protect resources with a username and password.
+// Note that HTTP Basic Authentication is not secure by itself and should
+// not be used to protect important assets without HTTPS. Even then, the
+// security of HTTP Basic Auth is disputed. Use discretion when deciding
+// what to protect with BasicAuth.
+type BasicAuth struct {
+	Next     httpserver.Handler
+	SiteRoot string
+	Rules    []Rule
+}
+
+// ServeHTTP implements the httpserver.Handler interface.
+func (a BasicAuth) ServeHTTP(w http.ResponseWriter, r *http.Request) (int, error) {
+
+	var hasAuth bool
+	var isAuthenticated bool
+
+	for _, rule := range a.Rules {
+		for _, res := range rule.Resources {
+			if !httpserver.Path(r.URL.Path).Matches(res) {
+				continue
+			}
+
+			// Path matches; parse auth header
+			username, password, ok := r.BasicAuth()
+			hasAuth = true
+
+			// Check credentials
+			if !ok ||
+				username != rule.Username ||
+				!rule.Password(password) {
+				//subtle.ConstantTimeCompare([]byte(password), []byte(rule.Password)) != 1 {
+				continue
+			}
+
+			// Flag set only on successful authentication
+			isAuthenticated = true
+		}
+	}
+
+	if hasAuth {
+		if !isAuthenticated {
+			w.Header().Set("WWW-Authenticate", "Basic")
+			return http.StatusUnauthorized, nil
+		}
+		// "It's an older code, sir, but it checks out. I was about to clear them."
+		return a.Next.ServeHTTP(w, r)
+	}
+
+	// Pass-thru when no paths match
+	return a.Next.ServeHTTP(w, r)
+}
+
+// Rule represents a BasicAuth rule. A username and password
+// combination protect the associated resources, which are
+// file or directory paths.
+type Rule struct {
+	Username  string
+	Password  func(string) bool
+	Resources []string
+}
+
+// PasswordMatcher determines whether a password matches a rule.
+type PasswordMatcher func(pw string) bool
+
+var (
+	htpasswords   map[string]map[string]PasswordMatcher
+	htpasswordsMu sync.Mutex
+)
+
+// GetHtpasswdMatcher matches password rules.
+func GetHtpasswdMatcher(filename, username, siteRoot string) (PasswordMatcher, error) {
+	filename = filepath.Join(siteRoot, filename)
+	htpasswordsMu.Lock()
+	if htpasswords == nil {
+		htpasswords = make(map[string]map[string]PasswordMatcher)
+	}
+	pm := htpasswords[filename]
+	if pm == nil {
+		fh, err := os.Open(filename)
+		if err != nil {
+			return nil, fmt.Errorf("open %q: %v", filename, err)
+		}
+		defer fh.Close()
+		pm = make(map[string]PasswordMatcher)
+		if err = parseHtpasswd(pm, fh); err != nil {
+			return nil, fmt.Errorf("parsing htpasswd %q: %v", fh.Name(), err)
+		}
+		htpasswords[filename] = pm
+	}
+	htpasswordsMu.Unlock()
+	if pm[username] == nil {
+		return nil, fmt.Errorf("username %q not found in %q", username, filename)
+	}
+	return pm[username], nil
+}
+
+func parseHtpasswd(pm map[string]PasswordMatcher, r io.Reader) error {
+	scanner := bufio.NewScanner(r)
+	for scanner.Scan() {
+		line := strings.TrimSpace(scanner.Text())
+		if line == "" || strings.IndexByte(line, '#') == 0 {
+			continue
+		}
+		i := strings.IndexByte(line, ':')
+		if i <= 0 {
+			return fmt.Errorf("malformed line, no color: %q", line)
+		}
+		user, encoded := line[:i], line[i+1:]
+		for _, p := range basic.DefaultSystems {
+			matcher, err := p(encoded)
+			if err != nil {
+				return err
+			}
+			if matcher != nil {
+				pm[user] = matcher.MatchesPassword
+				break
+			}
+		}
+	}
+	return scanner.Err()
+}
+
+// PlainMatcher returns a PasswordMatcher that does a constant-time
+// byte-wise comparison.
+func PlainMatcher(passw string) PasswordMatcher {
+	return func(pw string) bool {
+		return subtle.ConstantTimeCompare([]byte(pw), []byte(passw)) == 1
+	}
+}

caddyhttp/basicauth/basicauth_test.go

@@ -0,0 +1,146 @@
+package basicauth
+
+import (
+	"encoding/base64"
+	"fmt"
+	"io/ioutil"
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"path/filepath"
+	"testing"
+
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func TestBasicAuth(t *testing.T) {
+	rw := BasicAuth{
+		Next: httpserver.HandlerFunc(contentHandler),
+		Rules: []Rule{
+			{Username: "test", Password: PlainMatcher("ttest"), Resources: []string{"/testing"}},
+		},
+	}
+
+	tests := []struct {
+		from   string
+		result int
+		cred   string
+	}{
+		{"/testing", http.StatusUnauthorized, "ttest:test"},
+		{"/testing", http.StatusOK, "test:ttest"},
+		{"/testing", http.StatusUnauthorized, ""},
+	}
+
+	for i, test := range tests {
+
+		req, err := http.NewRequest("GET", test.from, nil)
+		if err != nil {
+			t.Fatalf("Test %d: Could not create HTTP request %v", i, err)
+		}
+		auth := "Basic " + base64.StdEncoding.EncodeToString([]byte(test.cred))
+		req.Header.Set("Authorization", auth)
+
+		rec := httptest.NewRecorder()
+		result, err := rw.ServeHTTP(rec, req)
+		if err != nil {
+			t.Fatalf("Test %d: Could not ServeHTTP %v", i, err)
+		}
+		if result != test.result {
+			t.Errorf("Test %d: Expected Header '%d' but was '%d'",
+				i, test.result, result)
+		}
+		if result == http.StatusUnauthorized {
+			headers := rec.Header()
+			if val, ok := headers["Www-Authenticate"]; ok {
+				if val[0] != "Basic" {
+					t.Errorf("Test %d, Www-Authenticate should be %s provided %s", i, "Basic", val[0])
+				}
+			} else {
+				t.Errorf("Test %d, should provide a header Www-Authenticate", i)
+			}
+		}
+
+	}
+
+}
+
+func TestMultipleOverlappingRules(t *testing.T) {
+	rw := BasicAuth{
+		Next: httpserver.HandlerFunc(contentHandler),
+		Rules: []Rule{
+			{Username: "t", Password: PlainMatcher("p1"), Resources: []string{"/t"}},
+			{Username: "t1", Password: PlainMatcher("p2"), Resources: []string{"/t/t"}},
+		},
+	}
+
+	tests := []struct {
+		from   string
+		result int
+		cred   string
+	}{
+		{"/t", http.StatusOK, "t:p1"},
+		{"/t/t", http.StatusOK, "t:p1"},
+		{"/t/t", http.StatusOK, "t1:p2"},
+		{"/a", http.StatusOK, "t1:p2"},
+		{"/t/t", http.StatusUnauthorized, "t1:p3"},
+		{"/t", http.StatusUnauthorized, "t1:p2"},
+	}
+
+	for i, test := range tests {
+
+		req, err := http.NewRequest("GET", test.from, nil)
+		if err != nil {
+			t.Fatalf("Test %d: Could not create HTTP request %v", i, err)
+		}
+		auth := "Basic " + base64.StdEncoding.EncodeToString([]byte(test.cred))
+		req.Header.Set("Authorization", auth)
+
+		rec := httptest.NewRecorder()
+		result, err := rw.ServeHTTP(rec, req)
+		if err != nil {
+			t.Fatalf("Test %d: Could not ServeHTTP %v", i, err)
+		}
+		if result != test.result {
+			t.Errorf("Test %d: Expected Header '%d' but was '%d'",
+				i, test.result, result)
+		}
+
+	}
+
+}
+
+func contentHandler(w http.ResponseWriter, r *http.Request) (int, error) {
+	fmt.Fprintf(w, r.URL.String())
+	return http.StatusOK, nil
+}
+
+func TestHtpasswd(t *testing.T) {
+	htpasswdPasswd := "IedFOuGmTpT8"
+	htpasswdFile := `sha1:{SHA}dcAUljwz99qFjYR0YLTXx0RqLww=
+md5:$apr1$l42y8rex$pOA2VJ0x/0TwaFeAF9nX61`
+
+	htfh, err := ioutil.TempFile("", "basicauth-")
+	if err != nil {
+		t.Skipf("Error creating temp file (%v), will skip htpassword test")
+		return
+	}
+	defer os.Remove(htfh.Name())
+	if _, err = htfh.Write([]byte(htpasswdFile)); err != nil {
+		t.Fatalf("write htpasswd file %q: %v", htfh.Name(), err)
+	}
+	htfh.Close()
+
+	for i, username := range []string{"sha1", "md5"} {
+		rule := Rule{Username: username, Resources: []string{"/testing"}}
+
+		siteRoot := filepath.Dir(htfh.Name())
+		filename := filepath.Base(htfh.Name())
+		if rule.Password, err = GetHtpasswdMatcher(filename, rule.Username, siteRoot); err != nil {
+			t.Fatalf("GetHtpasswdMatcher(%q, %q): %v", htfh.Name(), rule.Username, err)
+		}
+		t.Logf("%d. username=%q", i, rule.Username)
+		if !rule.Password(htpasswdPasswd) || rule.Password(htpasswdPasswd+"!") {
+			t.Errorf("%d (%s) password does not match.", i, rule.Username)
+		}
+	}
+}

caddyhttp/basicauth/setup.go

@@ -0,0 +1,82 @@
+package basicauth
+
+import (
+	"strings"
+
+	"github.com/mholt/caddy"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func init() {
+	caddy.RegisterPlugin("basicauth", caddy.Plugin{
+		ServerType: "http",
+		Action:     setup,
+	})
+}
+
+// setup configures a new BasicAuth middleware instance.
+func setup(c *caddy.Controller) error {
+	cfg := httpserver.GetConfig(c)
+	root := cfg.Root
+
+	rules, err := basicAuthParse(c)
+	if err != nil {
+		return err
+	}
+
+	basic := BasicAuth{Rules: rules}
+
+	cfg.AddMiddleware(func(next httpserver.Handler) httpserver.Handler {
+		basic.Next = next
+		basic.SiteRoot = root
+		return basic
+	})
+
+	return nil
+}
+
+func basicAuthParse(c *caddy.Controller) ([]Rule, error) {
+	var rules []Rule
+	cfg := httpserver.GetConfig(c)
+
+	var err error
+	for c.Next() {
+		var rule Rule
+
+		args := c.RemainingArgs()
+
+		switch len(args) {
+		case 2:
+			rule.Username = args[0]
+			if rule.Password, err = passwordMatcher(rule.Username, args[1], cfg.Root); err != nil {
+				return rules, c.Errf("Get password matcher from %s: %v", c.Val(), err)
+			}
+
+			for c.NextBlock() {
+				rule.Resources = append(rule.Resources, c.Val())
+				if c.NextArg() {
+					return rules, c.Errf("Expecting only one resource per line (extra '%s')", c.Val())
+				}
+			}
+		case 3:
+			rule.Resources = append(rule.Resources, args[0])
+			rule.Username = args[1]
+			if rule.Password, err = passwordMatcher(rule.Username, args[2], cfg.Root); err != nil {
+				return rules, c.Errf("Get password matcher from %s: %v", c.Val(), err)
+			}
+		default:
+			return rules, c.ArgErr()
+		}
+
+		rules = append(rules, rule)
+	}
+
+	return rules, nil
+}
+
+func passwordMatcher(username, passw, siteRoot string) (PasswordMatcher, error) {
+	if !strings.HasPrefix(passw, "htpasswd=") {
+		return PlainMatcher(passw), nil
+	}
+	return GetHtpasswdMatcher(passw[9:], username, siteRoot)
+}

caddyhttp/basicauth/setup_test.go

@@ -0,0 +1,132 @@
+package basicauth
+
+import (
+	"fmt"
+	"io/ioutil"
+	"os"
+	"strings"
+	"testing"
+
+	"github.com/mholt/caddy"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func TestSetup(t *testing.T) {
+	c := caddy.NewTestController("http", `basicauth user pwd`)
+	err := setup(c)
+	if err != nil {
+		t.Errorf("Expected no errors, but got: %v", err)
+	}
+	mids := httpserver.GetConfig(c).Middleware()
+	if len(mids) == 0 {
+		t.Fatal("Expected middleware, got 0 instead")
+	}
+
+	handler := mids[0](httpserver.EmptyNext)
+	myHandler, ok := handler.(BasicAuth)
+	if !ok {
+		t.Fatalf("Expected handler to be type BasicAuth, got: %#v", handler)
+	}
+
+	if !httpserver.SameNext(myHandler.Next, httpserver.EmptyNext) {
+		t.Error("'Next' field of handler was not set properly")
+	}
+}
+
+func TestBasicAuthParse(t *testing.T) {
+	htpasswdPasswd := "IedFOuGmTpT8"
+	htpasswdFile := `sha1:{SHA}dcAUljwz99qFjYR0YLTXx0RqLww=
+md5:$apr1$l42y8rex$pOA2VJ0x/0TwaFeAF9nX61`
+
+	var skipHtpassword bool
+	htfh, err := ioutil.TempFile(".", "basicauth-")
+	if err != nil {
+		t.Logf("Error creating temp file (%v), will skip htpassword test", err)
+		skipHtpassword = true
+	} else {
+		if _, err = htfh.Write([]byte(htpasswdFile)); err != nil {
+			t.Fatalf("write htpasswd file %q: %v", htfh.Name(), err)
+		}
+		htfh.Close()
+		defer os.Remove(htfh.Name())
+	}
+
+	tests := []struct {
+		input     string
+		shouldErr bool
+		password  string
+		expected  []Rule
+	}{
+		{`basicauth user pwd`, false, "pwd", []Rule{
+			{Username: "user"},
+		}},
+		{`basicauth user pwd {
+		}`, false, "pwd", []Rule{
+			{Username: "user"},
+		}},
+		{`basicauth user pwd {
+			/resource1
+			/resource2
+		}`, false, "pwd", []Rule{
+			{Username: "user", Resources: []string{"/resource1", "/resource2"}},
+		}},
+		{`basicauth /resource user pwd`, false, "pwd", []Rule{
+			{Username: "user", Resources: []string{"/resource"}},
+		}},
+		{`basicauth /res1 user1 pwd1
+		  basicauth /res2 user2 pwd2`, false, "pwd", []Rule{
+			{Username: "user1", Resources: []string{"/res1"}},
+			{Username: "user2", Resources: []string{"/res2"}},
+		}},
+		{`basicauth user`, true, "", []Rule{}},
+		{`basicauth`, true, "", []Rule{}},
+		{`basicauth /resource user pwd asdf`, true, "", []Rule{}},
+
+		{`basicauth sha1 htpasswd=` + htfh.Name(), false, htpasswdPasswd, []Rule{
+			{Username: "sha1"},
+		}},
+	}
+
+	for i, test := range tests {
+		actual, err := basicAuthParse(caddy.NewTestController("http", test.input))
+
+		if err == nil && test.shouldErr {
+			t.Errorf("Test %d didn't error, but it should have", i)
+		} else if err != nil && !test.shouldErr {
+			t.Errorf("Test %d errored, but it shouldn't have; got '%v'", i, err)
+		}
+
+		if len(actual) != len(test.expected) {
+			t.Fatalf("Test %d expected %d rules, but got %d",
+				i, len(test.expected), len(actual))
+		}
+
+		for j, expectedRule := range test.expected {
+			actualRule := actual[j]
+
+			if actualRule.Username != expectedRule.Username {
+				t.Errorf("Test %d, rule %d: Expected username '%s', got '%s'",
+					i, j, expectedRule.Username, actualRule.Username)
+			}
+
+			if strings.Contains(test.input, "htpasswd=") && skipHtpassword {
+				continue
+			}
+			pwd := test.password
+			if len(actual) > 1 {
+				pwd = fmt.Sprintf("%s%d", pwd, j+1)
+			}
+			if !actualRule.Password(pwd) || actualRule.Password(test.password+"!") {
+				t.Errorf("Test %d, rule %d: Expected password '%v', got '%v'",
+					i, j, test.password, actualRule.Password(""))
+			}
+
+			expectedRes := fmt.Sprintf("%v", expectedRule.Resources)
+			actualRes := fmt.Sprintf("%v", actualRule.Resources)
+			if actualRes != expectedRes {
+				t.Errorf("Test %d, rule %d: Expected resource list %s, but got %s",
+					i, j, expectedRes, actualRes)
+			}
+		}
+	}
+}

caddyhttp/bind/bind.go

@@ -0,0 +1,24 @@
+package bind
+
+import (
+	"github.com/mholt/caddy"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func init() {
+	caddy.RegisterPlugin("bind", caddy.Plugin{
+		ServerType: "http",
+		Action:     setupBind,
+	})
+}
+
+func setupBind(c *caddy.Controller) error {
+	config := httpserver.GetConfig(c)
+	for c.Next() {
+		if !c.Args(&config.ListenHost) {
+			return c.ArgErr()
+		}
+		config.TLS.ListenHost = config.ListenHost // necessary for ACME challenges, see issue #309
+	}
+	return nil
+}

caddyhttp/bind/bind_test.go

@@ -0,0 +1,24 @@
+package bind
+
+import (
+	"testing"
+
+	"github.com/mholt/caddy"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func TestSetupBind(t *testing.T) {
+	c := caddy.NewTestController("http", `bind 1.2.3.4`)
+	err := setupBind(c)
+	if err != nil {
+		t.Fatalf("Expected no errors, but got: %v", err)
+	}
+
+	cfg := httpserver.GetConfig(c)
+	if got, want := cfg.ListenHost, "1.2.3.4"; got != want {
+		t.Errorf("Expected the config's ListenHost to be %s, was %s", want, got)
+	}
+	if got, want := cfg.TLS.ListenHost, "1.2.3.4"; got != want {
+		t.Errorf("Expected the TLS config's ListenHost to be %s, was %s", want, got)
+	}
+}

caddyhttp/browse/browse.go

@@ -0,0 +1,432 @@
+// Package browse provides middleware for listing files in a directory
+// when directory path is requested instead of a specific file.
+package browse
+
+import (
+	"bytes"
+	"encoding/json"
+	"net/http"
+	"net/url"
+	"os"
+	"path"
+	"sort"
+	"strconv"
+	"strings"
+	"text/template"
+	"time"
+
+	"github.com/dustin/go-humanize"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+	"github.com/mholt/caddy/caddyhttp/staticfiles"
+)
+
+// Browse is an http.Handler that can show a file listing when
+// directories in the given paths are specified.
+type Browse struct {
+	Next          httpserver.Handler
+	Configs       []Config
+	IgnoreIndexes bool
+}
+
+// Config is a configuration for browsing in a particular path.
+type Config struct {
+	PathScope string
+	Root      http.FileSystem
+	Variables interface{}
+	Template  *template.Template
+}
+
+// A Listing is the context used to fill out a template.
+type Listing struct {
+	// The name of the directory (the last element of the path)
+	Name string
+
+	// The full path of the request
+	Path string
+
+	// Whether the parent directory is browsable
+	CanGoUp bool
+
+	// The items (files and folders) in the path
+	Items []FileInfo
+
+	// The number of directories in the listing
+	NumDirs int
+
+	// The number of files (items that aren't directories) in the listing
+	NumFiles int
+
+	// Which sorting order is used
+	Sort string
+
+	// And which order
+	Order string
+
+	// If ≠0 then Items have been limited to that many elements
+	ItemsLimitedTo int
+
+	// Optional custom variables for use in browse templates
+	User interface{}
+
+	httpserver.Context
+}
+
+// BreadcrumbMap returns l.Path where every element is a map
+// of URLs and path segment names.
+func (l Listing) BreadcrumbMap() map[string]string {
+	result := map[string]string{}
+
+	if len(l.Path) == 0 {
+		return result
+	}
+
+	// skip trailing slash
+	lpath := l.Path
+	if lpath[len(lpath)-1] == '/' {
+		lpath = lpath[:len(lpath)-1]
+	}
+
+	parts := strings.Split(lpath, "/")
+	for i, part := range parts {
+		if i == 0 && part == "" {
+			// Leading slash (root)
+			result["/"] = "/"
+			continue
+		}
+		result[strings.Join(parts[:i+1], "/")] = part
+	}
+
+	return result
+}
+
+// FileInfo is the info about a particular file or directory
+type FileInfo struct {
+	IsDir   bool
+	Name    string
+	Size    int64
+	URL     string
+	ModTime time.Time
+	Mode    os.FileMode
+}
+
+// HumanSize returns the size of the file as a human-readable string
+// in IEC format (i.e. power of 2 or base 1024).
+func (fi FileInfo) HumanSize() string {
+	return humanize.IBytes(uint64(fi.Size))
+}
+
+// HumanModTime returns the modified time of the file as a human-readable string.
+func (fi FileInfo) HumanModTime(format string) string {
+	return fi.ModTime.Format(format)
+}
+
+// Implement sorting for Listing
+type byName Listing
+type bySize Listing
+type byTime Listing
+
+// By Name
+func (l byName) Len() int      { return len(l.Items) }
+func (l byName) Swap(i, j int) { l.Items[i], l.Items[j] = l.Items[j], l.Items[i] }
+
+// Treat upper and lower case equally
+func (l byName) Less(i, j int) bool {
+	return strings.ToLower(l.Items[i].Name) < strings.ToLower(l.Items[j].Name)
+}
+
+// By Size
+func (l bySize) Len() int      { return len(l.Items) }
+func (l bySize) Swap(i, j int) { l.Items[i], l.Items[j] = l.Items[j], l.Items[i] }
+
+const directoryOffset = -1 << 31 // = math.MinInt32
+func (l bySize) Less(i, j int) bool {
+	iSize, jSize := l.Items[i].Size, l.Items[j].Size
+	if l.Items[i].IsDir {
+		iSize = directoryOffset + iSize
+	}
+	if l.Items[j].IsDir {
+		jSize = directoryOffset + jSize
+	}
+	return iSize < jSize
+}
+
+// By Time
+func (l byTime) Len() int           { return len(l.Items) }
+func (l byTime) Swap(i, j int)      { l.Items[i], l.Items[j] = l.Items[j], l.Items[i] }
+func (l byTime) Less(i, j int) bool { return l.Items[i].ModTime.Before(l.Items[j].ModTime) }
+
+// Add sorting method to "Listing"
+// it will apply what's in ".Sort" and ".Order"
+func (l Listing) applySort() {
+	// Check '.Order' to know how to sort
+	if l.Order == "desc" {
+		switch l.Sort {
+		case "name":
+			sort.Sort(sort.Reverse(byName(l)))
+		case "size":
+			sort.Sort(sort.Reverse(bySize(l)))
+		case "time":
+			sort.Sort(sort.Reverse(byTime(l)))
+		default:
+			// If not one of the above, do nothing
+			return
+		}
+	} else { // If we had more Orderings we could add them here
+		switch l.Sort {
+		case "name":
+			sort.Sort(byName(l))
+		case "size":
+			sort.Sort(bySize(l))
+		case "time":
+			sort.Sort(byTime(l))
+		default:
+			// If not one of the above, do nothing
+			return
+		}
+	}
+}
+
+func directoryListing(files []os.FileInfo, canGoUp bool, urlPath string) (Listing, bool) {
+	var (
+		fileinfos           []FileInfo
+		dirCount, fileCount int
+		hasIndexFile        bool
+	)
+
+	for _, f := range files {
+		name := f.Name()
+
+		for _, indexName := range staticfiles.IndexPages {
+			if name == indexName {
+				hasIndexFile = true
+				break
+			}
+		}
+
+		if f.IsDir() {
+			name += "/"
+			dirCount++
+		} else {
+			fileCount++
+		}
+
+		url := url.URL{Path: "./" + name} // prepend with "./" to fix paths with ':' in the name
+
+		fileinfos = append(fileinfos, FileInfo{
+			IsDir:   f.IsDir(),
+			Name:    f.Name(),
+			Size:    f.Size(),
+			URL:     url.String(),
+			ModTime: f.ModTime().UTC(),
+			Mode:    f.Mode(),
+		})
+	}
+
+	return Listing{
+		Name:     path.Base(urlPath),
+		Path:     urlPath,
+		CanGoUp:  canGoUp,
+		Items:    fileinfos,
+		NumDirs:  dirCount,
+		NumFiles: fileCount,
+	}, hasIndexFile
+}
+
+// ServeHTTP determines if the request is for this plugin, and if all prerequisites are met.
+// If so, control is handed over to ServeListing.
+func (b Browse) ServeHTTP(w http.ResponseWriter, r *http.Request) (int, error) {
+	var bc *Config
+	// See if there's a browse configuration to match the path
+	for i := range b.Configs {
+		if httpserver.Path(r.URL.Path).Matches(b.Configs[i].PathScope) {
+			bc = &b.Configs[i]
+			goto inScope
+		}
+	}
+	return b.Next.ServeHTTP(w, r)
+inScope:
+
+	// Browse works on existing directories; delegate everything else
+	requestedFilepath, err := bc.Root.Open(r.URL.Path)
+	if err != nil {
+		switch {
+		case os.IsPermission(err):
+			return http.StatusForbidden, err
+		case os.IsExist(err):
+			return http.StatusNotFound, err
+		default:
+			return b.Next.ServeHTTP(w, r)
+		}
+	}
+	defer requestedFilepath.Close()
+
+	info, err := requestedFilepath.Stat()
+	if err != nil {
+		switch {
+		case os.IsPermission(err):
+			return http.StatusForbidden, err
+		case os.IsExist(err):
+			return http.StatusGone, err
+		default:
+			return b.Next.ServeHTTP(w, r)
+		}
+	}
+	if !info.IsDir() {
+		return b.Next.ServeHTTP(w, r)
+	}
+
+	// Do not reply to anything else because it might be nonsensical
+	switch r.Method {
+	case http.MethodGet, http.MethodHead:
+		// proceed, noop
+	case "PROPFIND", http.MethodOptions:
+		return http.StatusNotImplemented, nil
+	default:
+		return b.Next.ServeHTTP(w, r)
+	}
+
+	// Browsing navigation gets messed up if browsing a directory
+	// that doesn't end in "/" (which it should, anyway)
+	if !strings.HasSuffix(r.URL.Path, "/") {
+		http.Redirect(w, r, r.URL.Path+"/", http.StatusTemporaryRedirect)
+		return 0, nil
+	}
+
+	return b.ServeListing(w, r, requestedFilepath, bc)
+}
+
+func (b Browse) loadDirectoryContents(requestedFilepath http.File, urlPath string) (*Listing, bool, error) {
+	files, err := requestedFilepath.Readdir(-1)
+	if err != nil {
+		return nil, false, err
+	}
+
+	// Determine if user can browse up another folder
+	var canGoUp bool
+	curPathDir := path.Dir(strings.TrimSuffix(urlPath, "/"))
+	for _, other := range b.Configs {
+		if strings.HasPrefix(curPathDir, other.PathScope) {
+			canGoUp = true
+			break
+		}
+	}
+
+	// Assemble listing of directory contents
+	listing, hasIndex := directoryListing(files, canGoUp, urlPath)
+
+	return &listing, hasIndex, nil
+}
+
+// handleSortOrder gets and stores for a Listing the 'sort' and 'order',
+// and reads 'limit' if given. The latter is 0 if not given.
+//
+// This sets Cookies.
+func (b Browse) handleSortOrder(w http.ResponseWriter, r *http.Request, scope string) (sort string, order string, limit int, err error) {
+	sort, order, limitQuery := r.URL.Query().Get("sort"), r.URL.Query().Get("order"), r.URL.Query().Get("limit")
+
+	// If the query 'sort' or 'order' is empty, use defaults or any values previously saved in Cookies
+	switch sort {
+	case "":
+		sort = "name"
+		if sortCookie, sortErr := r.Cookie("sort"); sortErr == nil {
+			sort = sortCookie.Value
+		}
+	case "name", "size", "type":
+		http.SetCookie(w, &http.Cookie{Name: "sort", Value: sort, Path: scope, Secure: r.TLS != nil})
+	}
+
+	switch order {
+	case "":
+		order = "asc"
+		if orderCookie, orderErr := r.Cookie("order"); orderErr == nil {
+			order = orderCookie.Value
+		}
+	case "asc", "desc":
+		http.SetCookie(w, &http.Cookie{Name: "order", Value: order, Path: scope, Secure: r.TLS != nil})
+	}
+
+	if limitQuery != "" {
+		limit, err = strconv.Atoi(limitQuery)
+		if err != nil { // if the 'limit' query can't be interpreted as a number, return err
+			return
+		}
+	}
+
+	return
+}
+
+// ServeListing returns a formatted view of 'requestedFilepath' contents'.
+func (b Browse) ServeListing(w http.ResponseWriter, r *http.Request, requestedFilepath http.File, bc *Config) (int, error) {
+	listing, containsIndex, err := b.loadDirectoryContents(requestedFilepath, r.URL.Path)
+	if err != nil {
+		switch {
+		case os.IsPermission(err):
+			return http.StatusForbidden, err
+		case os.IsExist(err):
+			return http.StatusGone, err
+		default:
+			return http.StatusInternalServerError, err
+		}
+	}
+	if containsIndex && !b.IgnoreIndexes { // directory isn't browsable
+		return b.Next.ServeHTTP(w, r)
+	}
+	listing.Context = httpserver.Context{
+		Root: bc.Root,
+		Req:  r,
+		URL:  r.URL,
+	}
+	listing.User = bc.Variables
+
+	// Copy the query values into the Listing struct
+	var limit int
+	listing.Sort, listing.Order, limit, err = b.handleSortOrder(w, r, bc.PathScope)
+	if err != nil {
+		return http.StatusBadRequest, err
+	}
+
+	listing.applySort()
+
+	if limit > 0 && limit <= len(listing.Items) {
+		listing.Items = listing.Items[:limit]
+		listing.ItemsLimitedTo = limit
+	}
+
+	var buf *bytes.Buffer
+	acceptHeader := strings.ToLower(strings.Join(r.Header["Accept"], ","))
+	switch {
+	case strings.Contains(acceptHeader, "application/json"):
+		if buf, err = b.formatAsJSON(listing, bc); err != nil {
+			return http.StatusInternalServerError, err
+		}
+		w.Header().Set("Content-Type", "application/json; charset=utf-8")
+
+	default: // There's no 'application/json' in the 'Accept' header; browse normally
+		if buf, err = b.formatAsHTML(listing, bc); err != nil {
+			return http.StatusInternalServerError, err
+		}
+		w.Header().Set("Content-Type", "text/html; charset=utf-8")
+
+	}
+
+	buf.WriteTo(w)
+
+	return http.StatusOK, nil
+}
+
+func (b Browse) formatAsJSON(listing *Listing, bc *Config) (*bytes.Buffer, error) {
+	marsh, err := json.Marshal(listing.Items)
+	if err != nil {
+		return nil, err
+	}
+
+	buf := new(bytes.Buffer)
+	_, err = buf.Write(marsh)
+	return buf, err
+}
+
+func (b Browse) formatAsHTML(listing *Listing, bc *Config) (*bytes.Buffer, error) {
+	buf := new(bytes.Buffer)
+	err := bc.Template.Execute(buf, listing)
+	return buf, err
+}

caddyhttp/browse/browse_test.go

@@ -0,0 +1,355 @@
+package browse
+
+import (
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"net/url"
+	"os"
+	"path/filepath"
+	"sort"
+	"testing"
+	"text/template"
+	"time"
+
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func TestSort(t *testing.T) {
+	// making up []fileInfo with bogus values;
+	// to be used to make up our "listing"
+	fileInfos := []FileInfo{
+		{
+			Name:    "fizz",
+			Size:    4,
+			ModTime: time.Now().AddDate(-1, 1, 0),
+		},
+		{
+			Name:    "buzz",
+			Size:    2,
+			ModTime: time.Now().AddDate(0, -3, 3),
+		},
+		{
+			Name:    "bazz",
+			Size:    1,
+			ModTime: time.Now().AddDate(0, -2, -23),
+		},
+		{
+			Name:    "jazz",
+			Size:    3,
+			ModTime: time.Now(),
+		},
+	}
+	listing := Listing{
+		Name:    "foobar",
+		Path:    "/fizz/buzz",
+		CanGoUp: false,
+		Items:   fileInfos,
+	}
+
+	// sort by name
+	listing.Sort = "name"
+	listing.applySort()
+	if !sort.IsSorted(byName(listing)) {
+		t.Errorf("The listing isn't name sorted: %v", listing.Items)
+	}
+
+	// sort by size
+	listing.Sort = "size"
+	listing.applySort()
+	if !sort.IsSorted(bySize(listing)) {
+		t.Errorf("The listing isn't size sorted: %v", listing.Items)
+	}
+
+	// sort by Time
+	listing.Sort = "time"
+	listing.applySort()
+	if !sort.IsSorted(byTime(listing)) {
+		t.Errorf("The listing isn't time sorted: %v", listing.Items)
+	}
+
+	// reverse by name
+	listing.Sort = "name"
+	listing.Order = "desc"
+	listing.applySort()
+	if !isReversed(byName(listing)) {
+		t.Errorf("The listing isn't reversed by name: %v", listing.Items)
+	}
+
+	// reverse by size
+	listing.Sort = "size"
+	listing.Order = "desc"
+	listing.applySort()
+	if !isReversed(bySize(listing)) {
+		t.Errorf("The listing isn't reversed by size: %v", listing.Items)
+	}
+
+	// reverse by time
+	listing.Sort = "time"
+	listing.Order = "desc"
+	listing.applySort()
+	if !isReversed(byTime(listing)) {
+		t.Errorf("The listing isn't reversed by time: %v", listing.Items)
+	}
+}
+
+func TestBrowseHTTPMethods(t *testing.T) {
+	tmpl, err := template.ParseFiles("testdata/photos.tpl")
+	if err != nil {
+		t.Fatalf("An error occured while parsing the template: %v", err)
+	}
+
+	b := Browse{
+		Next: httpserver.HandlerFunc(func(w http.ResponseWriter, r *http.Request) (int, error) {
+			return http.StatusTeapot, nil // not t.Fatalf, or we will not see what other methods yield
+		}),
+		Configs: []Config{
+			{
+				PathScope: "/photos",
+				Root:      http.Dir("./testdata"),
+				Template:  tmpl,
+			},
+		},
+	}
+
+	rec := httptest.NewRecorder()
+	for method, expected := range map[string]int{
+		http.MethodGet:     http.StatusOK,
+		http.MethodHead:    http.StatusOK,
+		http.MethodOptions: http.StatusNotImplemented,
+		"PROPFIND":         http.StatusNotImplemented,
+	} {
+		req, err := http.NewRequest(method, "/photos/", nil)
+		if err != nil {
+			t.Fatalf("Test: Could not create HTTP request: %v", err)
+		}
+
+		code, _ := b.ServeHTTP(rec, req)
+		if code != expected {
+			t.Errorf("Wrong status with HTTP Method %s: expected %d, got %d", method, expected, code)
+		}
+	}
+}
+
+func TestBrowseTemplate(t *testing.T) {
+	tmpl, err := template.ParseFiles("testdata/photos.tpl")
+	if err != nil {
+		t.Fatalf("An error occured while parsing the template: %v", err)
+	}
+
+	b := Browse{
+		Next: httpserver.HandlerFunc(func(w http.ResponseWriter, r *http.Request) (int, error) {
+			t.Fatalf("Next shouldn't be called")
+			return 0, nil
+		}),
+		Configs: []Config{
+			{
+				PathScope: "/photos",
+				Root:      http.Dir("./testdata"),
+				Template:  tmpl,
+			},
+		},
+	}
+
+	req, err := http.NewRequest("GET", "/photos/", nil)
+	if err != nil {
+		t.Fatalf("Test: Could not create HTTP request: %v", err)
+	}
+
+	rec := httptest.NewRecorder()
+
+	code, _ := b.ServeHTTP(rec, req)
+	if code != http.StatusOK {
+		t.Fatalf("Wrong status, expected %d, got %d", http.StatusOK, code)
+	}
+
+	respBody := rec.Body.String()
+	expectedBody := `<!DOCTYPE html>
+<html>
+<head>
+<title>Template</title>
+</head>
+<body>
+<h1>Header</h1>
+
+<h1>/photos/</h1>
+
+<a href="./test.html">test.html</a><br>
+
+<a href="./test2.html">test2.html</a><br>
+
+<a href="./test3.html">test3.html</a><br>
+
+</body>
+</html>
+`
+
+	if respBody != expectedBody {
+		t.Fatalf("Expected body: %v got: %v", expectedBody, respBody)
+	}
+
+}
+
+func TestBrowseJson(t *testing.T) {
+	b := Browse{
+		Next: httpserver.HandlerFunc(func(w http.ResponseWriter, r *http.Request) (int, error) {
+			t.Fatalf("Next shouldn't be called")
+			return 0, nil
+		}),
+		Configs: []Config{
+			{
+				PathScope: "/photos/",
+				Root:      http.Dir("./testdata"),
+			},
+		},
+	}
+
+	//Getting the listing from the ./testdata/photos, the listing returned will be used to validate test results
+	testDataPath := filepath.Join("./testdata", "photos")
+	file, err := os.Open(testDataPath)
+	if err != nil {
+		if os.IsPermission(err) {
+			t.Fatalf("Os Permission Error")
+		}
+	}
+	defer file.Close()
+
+	files, err := file.Readdir(-1)
+	if err != nil {
+		t.Fatalf("Unable to Read Contents of the directory")
+	}
+	var fileinfos []FileInfo
+
+	for i, f := range files {
+		name := f.Name()
+
+		// Tests fail in CI environment because all file mod times are the same for
+		// some reason, making the sorting unpredictable. To hack around this,
+		// we ensure here that each file has a different mod time.
+		chTime := f.ModTime().UTC().Add(-(time.Duration(i) * time.Second))
+		if err := os.Chtimes(filepath.Join(testDataPath, name), chTime, chTime); err != nil {
+			t.Fatal(err)
+		}
+
+		if f.IsDir() {
+			name += "/"
+		}
+
+		url := url.URL{Path: "./" + name}
+
+		fileinfos = append(fileinfos, FileInfo{
+			IsDir:   f.IsDir(),
+			Name:    f.Name(),
+			Size:    f.Size(),
+			URL:     url.String(),
+			ModTime: chTime,
+			Mode:    f.Mode(),
+		})
+	}
+	listing := Listing{Items: fileinfos} // this listing will be used for validation inside the tests
+
+	tests := []struct {
+		QueryURL       string
+		SortBy         string
+		OrderBy        string
+		Limit          int
+		shouldErr      bool
+		expectedResult []FileInfo
+	}{
+		//test case 1: testing for default sort and  order and without the limit parameter, default sort is by name and the default order is ascending
+		//without the limit query entire listing will be produced
+		{"/", "", "", -1, false, listing.Items},
+		//test case 2: limit is set to 1, orderBy and sortBy is default
+		{"/?limit=1", "", "", 1, false, listing.Items[:1]},
+		//test case 3 : if the listing request is bigger than total size of listing then it should return everything
+		{"/?limit=100000000", "", "", 100000000, false, listing.Items},
+		//test case 4 : testing for negative limit
+		{"/?limit=-1", "", "", -1, false, listing.Items},
+		//test case 5 : testing with limit set to -1 and order set to descending
+		{"/?limit=-1&order=desc", "", "desc", -1, false, listing.Items},
+		//test case 6 : testing with limit set to 2 and order set to descending
+		{"/?limit=2&order=desc", "", "desc", 2, false, listing.Items},
+		//test case 7 : testing with limit set to 3 and order set to descending
+		{"/?limit=3&order=desc", "", "desc", 3, false, listing.Items},
+		//test case 8 : testing with limit set to 3 and order set to ascending
+		{"/?limit=3&order=asc", "", "asc", 3, false, listing.Items},
+		//test case 9 : testing with limit set to 1111111 and order set to ascending
+		{"/?limit=1111111&order=asc", "", "asc", 1111111, false, listing.Items},
+		//test case 10 : testing with limit set to default and order set to ascending and sorting by size
+		{"/?order=asc&sort=size", "size", "asc", -1, false, listing.Items},
+		//test case 11 : testing with limit set to default and order set to ascending and sorting by last modified
+		{"/?order=asc&sort=time", "time", "asc", -1, false, listing.Items},
+		//test case 12 : testing with limit set to 1 and order set to ascending and sorting by last modified
+		{"/?order=asc&sort=time&limit=1", "time", "asc", 1, false, listing.Items},
+		//test case 13 : testing with limit set to -100 and order set to ascending and sorting by last modified
+		{"/?order=asc&sort=time&limit=-100", "time", "asc", -100, false, listing.Items},
+		//test case 14 : testing with limit set to -100 and order set to ascending and sorting by size
+		{"/?order=asc&sort=size&limit=-100", "size", "asc", -100, false, listing.Items},
+	}
+
+	for i, test := range tests {
+		var marsh []byte
+		req, err := http.NewRequest("GET", "/photos"+test.QueryURL, nil)
+
+		if err == nil && test.shouldErr {
+			t.Errorf("Test %d didn't error, but it should have", i)
+		} else if err != nil && !test.shouldErr {
+			t.Errorf("Test %d errored, but it shouldn't have; got '%v'", i, err)
+		}
+
+		req.Header.Set("Accept", "application/json")
+		rec := httptest.NewRecorder()
+
+		code, err := b.ServeHTTP(rec, req)
+
+		if code != http.StatusOK {
+			t.Fatalf("In test %d: Wrong status, expected %d, got %d", i, http.StatusOK, code)
+		}
+		if rec.HeaderMap.Get("Content-Type") != "application/json; charset=utf-8" {
+			t.Fatalf("Expected Content type to be application/json; charset=utf-8, but got %s ", rec.HeaderMap.Get("Content-Type"))
+		}
+
+		actualJSONResponse := rec.Body.String()
+		copyOflisting := listing
+		if test.SortBy == "" {
+			copyOflisting.Sort = "name"
+		} else {
+			copyOflisting.Sort = test.SortBy
+		}
+		if test.OrderBy == "" {
+			copyOflisting.Order = "asc"
+		} else {
+			copyOflisting.Order = test.OrderBy
+		}
+
+		copyOflisting.applySort()
+
+		limit := test.Limit
+		if limit <= len(copyOflisting.Items) && limit > 0 {
+			marsh, err = json.Marshal(copyOflisting.Items[:limit])
+		} else { // if the 'limit' query is empty, or has the wrong value, list everything
+			marsh, err = json.Marshal(copyOflisting.Items)
+		}
+
+		if err != nil {
+			t.Fatalf("Unable to Marshal the listing ")
+		}
+		expectedJSON := string(marsh)
+
+		if actualJSONResponse != expectedJSON {
+			t.Errorf("JSON response doesn't match the expected for test number %d with sort=%s, order=%s\nExpected response %s\nActual response = %s\n",
+				i+1, test.SortBy, test.OrderBy, expectedJSON, actualJSONResponse)
+		}
+	}
+}
+
+// "sort" package has "IsSorted" function, but no "IsReversed";
+func isReversed(data sort.Interface) bool {
+	n := data.Len()
+	for i := n - 1; i > 0; i-- {
+		if !data.Less(i, i-1) {
+			return false
+		}
+	}
+	return true
+}

caddyhttp/browse/setup.go

@@ -0,0 +1,427 @@
+package browse
+
+import (
+	"fmt"
+	"io/ioutil"
+	"net/http"
+	"text/template"
+
+	"github.com/mholt/caddy"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func init() {
+	caddy.RegisterPlugin("browse", caddy.Plugin{
+		ServerType: "http",
+		Action:     setup,
+	})
+}
+
+// setup configures a new Browse middleware instance.
+func setup(c *caddy.Controller) error {
+	configs, err := browseParse(c)
+	if err != nil {
+		return err
+	}
+
+	b := Browse{
+		Configs:       configs,
+		IgnoreIndexes: false,
+	}
+
+	httpserver.GetConfig(c).AddMiddleware(func(next httpserver.Handler) httpserver.Handler {
+		b.Next = next
+		return b
+	})
+
+	return nil
+}
+
+func browseParse(c *caddy.Controller) ([]Config, error) {
+	var configs []Config
+
+	cfg := httpserver.GetConfig(c)
+
+	appendCfg := func(bc Config) error {
+		for _, c := range configs {
+			if c.PathScope == bc.PathScope {
+				return fmt.Errorf("duplicate browsing config for %s", c.PathScope)
+			}
+		}
+		configs = append(configs, bc)
+		return nil
+	}
+
+	for c.Next() {
+		var bc Config
+
+		// First argument is directory to allow browsing; default is site root
+		if c.NextArg() {
+			bc.PathScope = c.Val()
+		} else {
+			bc.PathScope = "/"
+		}
+		bc.Root = http.Dir(cfg.Root)
+		theRoot, err := bc.Root.Open("/") // catch a missing path early
+		if err != nil {
+			return configs, err
+		}
+		defer theRoot.Close()
+		_, err = theRoot.Readdir(-1)
+		if err != nil {
+			return configs, err
+		}
+
+		// Second argument would be the template file to use
+		var tplText string
+		if c.NextArg() {
+			tplBytes, err := ioutil.ReadFile(c.Val())
+			if err != nil {
+				return configs, err
+			}
+			tplText = string(tplBytes)
+		} else {
+			tplText = defaultTemplate
+		}
+
+		// Build the template
+		tpl, err := template.New("listing").Parse(tplText)
+		if err != nil {
+			return configs, err
+		}
+		bc.Template = tpl
+
+		// Save configuration
+		err = appendCfg(bc)
+		if err != nil {
+			return configs, err
+		}
+	}
+
+	return configs, nil
+}
+
+// The default template to use when serving up directory listings
+const defaultTemplate = `<!DOCTYPE html>
+<html>
+	<head>
+		<title>{{.Name}}</title>
+		<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<style>
+* { padding: 0; margin: 0; }
+
+body {
+	font-family: sans-serif;
+	text-rendering: optimizespeed;
+}
+
+a {
+	color: #006ed3;
+	text-decoration: none;
+}
+
+a:hover,
+h1 a:hover {
+	color: #319cff;
+}
+
+header,
+#summary {
+	padding-left: 5%;
+	padding-right: 5%;
+}
+
+th:first-child,
+td:first-child {
+	padding-left: 5%;
+}
+
+th:last-child,
+td:last-child {
+	padding-right: 5%;
+}
+
+header {
+	padding-top: 25px;
+	padding-bottom: 15px;
+	background-color: #f2f2f2;
+}
+
+h1 {
+	font-size: 20px;
+	font-weight: normal;
+	white-space: nowrap;
+	overflow-x: hidden;
+	text-overflow: ellipsis;
+}
+
+h1 a {
+	color: inherit;
+}
+
+h1 a:hover {
+	text-decoration: underline;
+}
+
+main {
+	display: block;
+}
+
+.meta {
+	font-size: 12px;
+	font-family: Verdana, sans-serif;
+	border-bottom: 1px solid #9C9C9C;
+	padding-top: 15px;
+	padding-bottom: 15px;
+}
+
+.meta-item {
+	margin-right: 1em;
+}
+
+table {
+	width: 100%;
+	border-collapse: collapse;
+}
+
+tr {
+	border-bottom: 1px dashed #dadada;
+}
+
+tbody tr:hover {
+	background-color: #ffffec;
+}
+
+th,
+td {
+	text-align: left;
+	padding: 10px 0;
+}
+
+th {
+	padding-top: 15px;
+	padding-bottom: 15px;
+	font-size: 16px;
+	white-space: nowrap;
+}
+
+th a {
+	color: black;
+}
+
+th svg {
+	vertical-align: middle;
+}
+
+td {
+	font-size: 14px;
+}
+
+td:first-child {
+	width: 50%;
+}
+
+th:last-child,
+td:last-child {
+	text-align: right;
+}
+
+td:first-child svg {
+	position: absolute;
+}
+
+td .name,
+td .goup {
+	margin-left: 1.75em;
+	word-break: break-all;
+	overflow-wrap: break-word;
+	white-space: pre-wrap;
+}
+
+footer {
+	padding: 40px 20px;
+	font-size: 12px;
+	text-align: center;
+}
+
+@media (max-width: 600px) {
+	.hideable {
+		display: none;
+	}
+
+	td:first-child {
+		width: auto;
+	}
+
+	th:nth-child(2),
+	td:nth-child(2) {
+		padding-right: 5%;
+		text-align: right;
+	}
+}
+</style>
+	</head>
+	<body>
+		<svg version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" height="0" width="0" style="position: absolute;">
+			<defs>
+				<!-- Folder -->
+				<linearGradient id="f" y2="640" gradientUnits="userSpaceOnUse" x2="244.84" gradientTransform="matrix(.97319 0 0 1.0135 -.50695 -13.679)" y1="415.75" x1="244.84">
+					<stop stop-color="#b3ddfd" offset="0"/>
+					<stop stop-color="#69c" offset="1"/>
+				</linearGradient>
+				<linearGradient id="e" y2="571.06" gradientUnits="userSpaceOnUse" x2="238.03" gradientTransform="translate(0,2)" y1="346.05" x1="236.26">
+					<stop stop-color="#ace" offset="0"/>
+					<stop stop-color="#369" offset="1"/>
+				</linearGradient>
+				<g id="folder" transform="translate(-266.06 -193.36)">
+					<g transform="matrix(.066019 0 0 .066019 264.2 170.93)">
+						<g transform="matrix(1.4738 0 0 1.4738 -52.053 -166.93)">
+							<path fill="#69c" d="m98.424 343.78c-11.08 0-20 8.92-20 20v48.5 33.719 105.06c0 11.08 8.92 20 20 20h279.22c11.08 0 20-8.92 20-20v-138.78c0-11.08-8.92-20-20-20h-117.12c-7.5478-1.1844-9.7958-6.8483-10.375-11.312v-5.625-11.562c0-11.08-8.92-20-20-20h-131.72z"/>
+							<rect rx="12.885" ry="12.199" height="227.28" width="366.69" y="409.69" x="54.428" fill="#369"/>
+							<path fill="url(#e)" d="m98.424 345.78c-11.08 0-20 8.92-20 20v48.5 33.719 105.06c0 11.08 8.92 20 20 20h279.22c11.08 0 20-8.92 20-20v-138.78c0-11.08-8.92-20-20-20h-117.12c-7.5478-1.1844-9.7958-6.8483-10.375-11.312v-5.625-11.562c0-11.08-8.92-20-20-20h-131.72z"/>
+							<rect rx="12.885" ry="12.199" height="227.28" width="366.69" y="407.69" x="54.428" fill="url(#f)"/>
+						</g>
+					</g>
+				</g>
+
+				<!-- File -->
+				<linearGradient id="a">
+					<stop stop-color="#cbcbcb" offset="0"/>
+					<stop stop-color="#f0f0f0" offset=".34923"/>
+					<stop stop-color="#e2e2e2" offset="1"/>
+				</linearGradient>
+				<linearGradient id="d" y2="686.15" xlink:href="#a" gradientUnits="userSpaceOnUse" y1="207.83" gradientTransform="matrix(.28346 0 0 .31053 -608.52 485.11)" x2="380.1" x1="749.25"/>
+				<linearGradient id="c" y2="287.74" xlink:href="#a" gradientUnits="userSpaceOnUse" y1="169.44" gradientTransform="matrix(.28342 0 0 .31057 -608.52 485.11)" x2="622.33" x1="741.64"/>
+				<linearGradient id="b" y2="418.54" gradientUnits="userSpaceOnUse" y1="236.13" gradientTransform="matrix(.29343 0 0 .29999 -608.52 485.11)" x2="330.88" x1="687.96">
+					<stop stop-color="#fff" offset="0"/>
+					<stop stop-color="#fff" stop-opacity="0" offset="1"/>
+				</linearGradient>
+				<g id="file" transform="translate(-278.15 -216.59)">
+					<g fill-rule="evenodd" transform="matrix(.19775 0 0 .19775 381.05 112.68)">
+						<path d="m-520.17 525.5v36.739 36.739 36.739 36.739h33.528 33.528 33.528 33.528v-36.739-36.739-36.739l-33.528-36.739h-33.528-33.528-33.528z" stroke-opacity=".36478" stroke-width=".42649" fill="#fff"/>
+						<g>
+							<path d="m-520.11 525.68v36.739 36.739 36.739 36.739h33.528 33.528 33.528 33.528v-36.739-36.739-36.739l-33.528-36.739h-33.528-33.528-33.528z" stroke-opacity=".36478" stroke="#000" stroke-width=".42649" fill="url(#d)"/>
+							<path d="m-386 562.42c-10.108-2.9925-23.206-2.5682-33.101-0.86253 1.7084-10.962 1.922-24.701-0.4271-35.877l33.528 36.739z" stroke-width=".95407pt" fill="url(#c)"/>
+							<path d="m-519.13 537-0.60402 134.7h131.68l0.0755-33.296c-2.9446 1.1325-32.692-40.998-70.141-39.186-37.483 1.8137-27.785-56.777-61.006-62.214z" stroke-width="1pt" fill="url(#b)"/>
+						</g>
+					</g>
+				</g>
+
+				<!-- Up arrow -->
+				<g id="up-arrow" transform="translate(-279.22 -208.12)">
+					<path transform="matrix(.22413 0 0 .12089 335.67 164.35)" stroke-width="0" d="m-194.17 412.01h-28.827-28.827l14.414-24.965 14.414-24.965 14.414 24.965z"/>
+				</g>
+
+				<!-- Down arrow -->
+				<g id="down-arrow" transform="translate(-279.22 -208.12)">
+					<path transform="matrix(.22413 0 0 -.12089 335.67 257.93)" stroke-width="0" d="m-194.17 412.01h-28.827-28.827l14.414-24.965 14.414-24.965 14.414 24.965z"/>
+				</g>
+			</defs>
+		</svg>
+
+		<header>
+			<h1>
+				{{range $url, $name := .BreadcrumbMap}}<a href="{{$url}}">{{$name}}</a>{{if ne $url "/"}}/{{end}}{{end}}
+			</h1>
+		</header>
+		<main>
+			<div class="meta">
+				<div id="summary">
+					<span class="meta-item"><b>{{.NumDirs}}</b> director{{if eq 1 .NumDirs}}y{{else}}ies{{end}}</span>
+					<span class="meta-item"><b>{{.NumFiles}}</b> file{{if ne 1 .NumFiles}}s{{end}}</span>
+					{{- if ne 0 .ItemsLimitedTo}}
+					<span class="meta-item">(of which only <b>{{.ItemsLimitedTo}}</b> are displayed)</span>
+					{{- end}}
+				</div>
+			</div>
+			<div class="listing">
+				<table aria-describedby="summary">
+					<thead>
+					<tr>
+						<th>
+							{{- if and (eq .Sort "name") (ne .Order "desc")}}
+							<a href="?sort=name&order=desc{{if ne 0 .ItemsLimitedTo}}&limit={{.ItemsLimitedTo}}{{end}}">Name <svg width="1em" height=".4em" version="1.1" viewBox="0 0 12.922194 6.0358899"><use xlink:href="#up-arrow"></use></svg></a>
+							{{- else if and (eq .Sort "name") (ne .Order "asc")}}
+							<a href="?sort=name&order=asc{{if ne 0 .ItemsLimitedTo}}&limit={{.ItemsLimitedTo}}{{end}}">Name <svg width="1em" height=".4em" version="1.1" viewBox="0 0 12.922194 6.0358899"><use xlink:href="#down-arrow"></use></svg></a>
+							{{- else}}
+							<a href="?sort=name&order=asc{{if ne 0 .ItemsLimitedTo}}&limit={{.ItemsLimitedTo}}{{end}}">Name</a>
+							{{- end}}
+						</th>
+						<th>
+							{{- if and (eq .Sort "size") (ne .Order "desc")}}
+							<a href="?sort=size&order=desc{{if ne 0 .ItemsLimitedTo}}&limit={{.ItemsLimitedTo}}{{end}}">Size <svg width="1em" height=".4em" version="1.1" viewBox="0 0 12.922194 6.0358899"><use xlink:href="#up-arrow"></use></svg></a>
+							{{- else if and (eq .Sort "size") (ne .Order "asc")}}
+							<a href="?sort=size&order=asc{{if ne 0 .ItemsLimitedTo}}&limit={{.ItemsLimitedTo}}{{end}}">Size <svg width="1em" height=".4em" version="1.1" viewBox="0 0 12.922194 6.0358899"><use xlink:href="#down-arrow"></use></svg></a>
+							{{- else}}
+							<a href="?sort=size&order=asc{{if ne 0 .ItemsLimitedTo}}&limit={{.ItemsLimitedTo}}{{end}}">Size</a>
+							{{- end}}
+						</th>
+						<th class="hideable">
+							{{- if and (eq .Sort "time") (ne .Order "desc")}}
+							<a href="?sort=time&order=desc{{if ne 0 .ItemsLimitedTo}}&limit={{.ItemsLimitedTo}}{{end}}">Modified <svg width="1em" height=".4em" version="1.1" viewBox="0 0 12.922194 6.0358899"><use xlink:href="#up-arrow"></use></svg></a>
+							{{- else if and (eq .Sort "time") (ne .Order "asc")}}
+							<a href="?sort=time&order=asc{{if ne 0 .ItemsLimitedTo}}&limit={{.ItemsLimitedTo}}{{end}}">Modified <svg width="1em" height=".4em" version="1.1" viewBox="0 0 12.922194 6.0358899"><use xlink:href="#down-arrow"></use></svg></a>
+							{{- else}}
+							<a href="?sort=time&order=asc{{if ne 0 .ItemsLimitedTo}}&limit={{.ItemsLimitedTo}}{{end}}">Modified</a>
+							{{- end}}
+						</th>
+					</tr>
+					</thead>
+					<tbody>
+					{{- if .CanGoUp}}
+					<tr>
+						<td>
+							<a href="..">
+								<span class="goup">Go up</span>
+							</a>
+						</td>
+						<td>&mdash;</td>
+						<td class="hideable">&mdash;</td>
+					</tr>
+					{{- end}}
+					{{- range .Items}}
+					<tr>
+						<td>
+							<a href="{{.URL}}">
+								{{- if .IsDir}}
+								<svg width="1.5em" height="1em" version="1.1" viewBox="0 0 35.678803 28.527945"><use xlink:href="#folder"></use></svg>
+								{{- else}}
+								<svg width="1.5em" height="1em" version="1.1" viewBox="0 0 26.604381 29.144726"><use xlink:href="#file"></use></svg>
+								{{- end}}
+								<span class="name">{{.Name}}</span>
+							</a>
+						</td>
+						{{- if .IsDir}}
+						<td data-order="-1">&mdash;</td>
+						{{- else}}
+						<td data-order="{{.Size}}">{{.HumanSize}}</td>
+						{{- end}}
+						<td class="hideable"><time datetime="{{.HumanModTime "2006-01-02T15:04:05Z"}}">{{.HumanModTime "01/02/2006 03:04:05 PM -07:00"}}</time></td>
+					</tr>
+					{{- end}}
+					</tbody>
+				</table>
+			</div>
+		</main>
+		<footer>
+			Served with <a rel="noopener noreferrer" href="https://caddyserver.com">Caddy</a>.
+		</footer>
+		<script type="text/javascript">
+			function localizeDatetime(e, index, ar) {
+				if (e.textContent === undefined) {
+					return;
+				}
+				var d = new Date(e.getAttribute('datetime'));
+				if (isNaN(d)) {
+					d = new Date(e.textContent);
+					if (isNaN(d)) {
+						return;
+					}
+				}
+				e.textContent = d.toLocaleString();
+			}
+			var timeList = Array.prototype.slice.call(document.getElementsByTagName("time"));
+			timeList.forEach(localizeDatetime);
+		</script>
+	</body>
+</html>`

caddyhttp/browse/setup_test.go

@@ -0,0 +1,69 @@
+package browse
+
+import (
+	"io/ioutil"
+	"os"
+	"path/filepath"
+	"strconv"
+	"testing"
+	"time"
+
+	"github.com/mholt/caddy"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func TestSetup(t *testing.T) {
+	tempDirPath := os.TempDir()
+	_, err := os.Stat(tempDirPath)
+	if err != nil {
+		t.Fatalf("BeforeTest: Failed to find an existing directory for testing! Error was: %v", err)
+	}
+	nonExistantDirPath := filepath.Join(tempDirPath, strconv.Itoa(int(time.Now().UnixNano())))
+
+	tempTemplate, err := ioutil.TempFile(".", "tempTemplate")
+	if err != nil {
+		t.Fatalf("BeforeTest: Failed to create a temporary file in the working directory! Error was: %v", err)
+	}
+	defer os.Remove(tempTemplate.Name())
+
+	tempTemplatePath := filepath.Join(".", tempTemplate.Name())
+
+	for i, test := range []struct {
+		input             string
+		expectedPathScope []string
+		shouldErr         bool
+	}{
+		// test case #0 tests handling of multiple pathscopes
+		{"browse " + tempDirPath + "\n browse .", []string{tempDirPath, "."}, false},
+
+		// test case #1 tests instantiation of Config with default values
+		{"browse /", []string{"/"}, false},
+
+		// test case #2 tests detectaction of custom template
+		{"browse . " + tempTemplatePath, []string{"."}, false},
+
+		// test case #3 tests detection of non-existent template
+		{"browse . " + nonExistantDirPath, nil, true},
+
+		// test case #4 tests detection of duplicate pathscopes
+		{"browse " + tempDirPath + "\n browse " + tempDirPath, nil, true},
+	} {
+
+		c := caddy.NewTestController("http", test.input)
+		err := setup(c)
+		if err != nil && !test.shouldErr {
+			t.Errorf("Test case #%d recieved an error of %v", i, err)
+		}
+		if test.expectedPathScope == nil {
+			continue
+		}
+		mids := httpserver.GetConfig(c).Middleware()
+		mid := mids[len(mids)-1]
+		recievedConfigs := mid(nil).(Browse).Configs
+		for j, config := range recievedConfigs {
+			if config.PathScope != test.expectedPathScope[j] {
+				t.Errorf("Test case #%d expected a pathscope of %v, but got %v", i, test.expectedPathScope, config.PathScope)
+			}
+		}
+	}
+}

caddyhttp/browse/testdata/header.html

@@ -0,0 +1 @@
+<h1>Header</h1>

caddyhttp/browse/testdata/photos.tpl

@@ -0,0 +1,13 @@
+<!DOCTYPE html>
+<html>
+<head>
+<title>Template</title>
+</head>
+<body>
+{{.Include "header.html"}}
+<h1>{{.Path}}</h1>
+{{range .Items}}
+<a href="{{.URL}}">{{.Name}}</a><br>
+{{end}}
+</body>
+</html>

caddyhttp/browse/testdata/photos/test.html

@@ -0,0 +1,8 @@
+<!DOCTYPE html>
+<html>
+<head>
+<title>Test</title>
+</head>
+<body>
+</body>
+</html>

caddyhttp/browse/testdata/photos/test2.html

@@ -0,0 +1,8 @@
+<!DOCTYPE html>
+<html>
+<head>
+<title>Test 2</title>
+</head>
+<body>
+</body>
+</html>

caddyhttp/browse/testdata/photos/test3.html

@@ -0,0 +1,3 @@
+<!DOCTYPE html>
+<html>
+</html>

caddyhttp/caddyhttp.go

@@ -0,0 +1,29 @@
+package caddyhttp
+
+import (
+	// plug in the server
+	_ "github.com/mholt/caddy/caddyhttp/httpserver"
+
+	// plug in the standard directives
+	_ "github.com/mholt/caddy/caddyhttp/basicauth"
+	_ "github.com/mholt/caddy/caddyhttp/bind"
+	_ "github.com/mholt/caddy/caddyhttp/browse"
+	_ "github.com/mholt/caddy/caddyhttp/errors"
+	_ "github.com/mholt/caddy/caddyhttp/expvar"
+	_ "github.com/mholt/caddy/caddyhttp/extensions"
+	_ "github.com/mholt/caddy/caddyhttp/fastcgi"
+	_ "github.com/mholt/caddy/caddyhttp/gzip"
+	_ "github.com/mholt/caddy/caddyhttp/header"
+	_ "github.com/mholt/caddy/caddyhttp/internalsrv"
+	_ "github.com/mholt/caddy/caddyhttp/log"
+	_ "github.com/mholt/caddy/caddyhttp/markdown"
+	_ "github.com/mholt/caddy/caddyhttp/mime"
+	_ "github.com/mholt/caddy/caddyhttp/pprof"
+	_ "github.com/mholt/caddy/caddyhttp/proxy"
+	_ "github.com/mholt/caddy/caddyhttp/redirect"
+	_ "github.com/mholt/caddy/caddyhttp/rewrite"
+	_ "github.com/mholt/caddy/caddyhttp/root"
+	_ "github.com/mholt/caddy/caddyhttp/templates"
+	_ "github.com/mholt/caddy/caddyhttp/websocket"
+	_ "github.com/mholt/caddy/startupshutdown"
+)

caddyhttp/caddyhttp_test.go

@@ -0,0 +1,19 @@
+package caddyhttp
+
+import (
+	"strings"
+	"testing"
+
+	"github.com/mholt/caddy"
+)
+
+// TODO: this test could be improved; the purpose is to
+// ensure that the standard plugins are in fact plugged in
+// and registered properly; this is a quick/naive way to do it.
+func TestStandardPlugins(t *testing.T) {
+	numStandardPlugins := 25 // importing caddyhttp plugs in this many plugins
+	s := caddy.DescribePlugins()
+	if got, want := strings.Count(s, "\n"), numStandardPlugins+5; got != want {
+		t.Errorf("Expected all standard plugins to be plugged in, got:\n%s", s)
+	}
+}

caddyhttp/errors/errors.go

@@ -0,0 +1,140 @@
+// Package errors implements an HTTP error handling middleware.
+package errors
+
+import (
+	"fmt"
+	"io"
+	"log"
+	"net/http"
+	"os"
+	"runtime"
+	"strings"
+	"time"
+
+	"github.com/mholt/caddy"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func init() {
+	caddy.RegisterPlugin("errors", caddy.Plugin{
+		ServerType: "http",
+		Action:     setup,
+	})
+}
+
+// ErrorHandler handles HTTP errors (and errors from other middleware).
+type ErrorHandler struct {
+	Next       httpserver.Handler
+	ErrorPages map[int]string // map of status code to filename
+	LogFile    string
+	Log        *log.Logger
+	LogRoller  *httpserver.LogRoller
+	Debug      bool // if true, errors are written out to client rather than to a log
+}
+
+func (h ErrorHandler) ServeHTTP(w http.ResponseWriter, r *http.Request) (int, error) {
+	defer h.recovery(w, r)
+
+	status, err := h.Next.ServeHTTP(w, r)
+
+	if err != nil {
+		errMsg := fmt.Sprintf("%s [ERROR %d %s] %v", time.Now().Format(timeFormat), status, r.URL.Path, err)
+		if h.Debug {
+			// Write error to response instead of to log
+			w.Header().Set("Content-Type", "text/plain; charset=utf-8")
+			w.WriteHeader(status)
+			fmt.Fprintln(w, errMsg)
+			return 0, err // returning 0 signals that a response has been written
+		}
+		h.Log.Println(errMsg)
+	}
+
+	if status >= 400 {
+		h.errorPage(w, r, status)
+		return 0, err
+	}
+
+	return status, err
+}
+
+// errorPage serves a static error page to w according to the status
+// code. If there is an error serving the error page, a plaintext error
+// message is written instead, and the extra error is logged.
+func (h ErrorHandler) errorPage(w http.ResponseWriter, r *http.Request, code int) {
+	// See if an error page for this status code was specified
+	if pagePath, ok := h.ErrorPages[code]; ok {
+		// Try to open it
+		errorPage, err := os.Open(pagePath)
+		if err != nil {
+			// An additional error handling an error... <insert grumpy cat here>
+			h.Log.Printf("%s [NOTICE %d %s] could not load error page: %v",
+				time.Now().Format(timeFormat), code, r.URL.String(), err)
+			httpserver.DefaultErrorFunc(w, r, code)
+			return
+		}
+		defer errorPage.Close()
+
+		// Copy the page body into the response
+		w.Header().Set("Content-Type", "text/html; charset=utf-8")
+		w.WriteHeader(code)
+		_, err = io.Copy(w, errorPage)
+
+		if err != nil {
+			// Epic fail... sigh.
+			h.Log.Printf("%s [NOTICE %d %s] could not respond with %s: %v",
+				time.Now().Format(timeFormat), code, r.URL.String(), pagePath, err)
+			httpserver.DefaultErrorFunc(w, r, code)
+		}
+
+		return
+	}
+
+	// Default error response
+	httpserver.DefaultErrorFunc(w, r, code)
+}
+
+func (h ErrorHandler) recovery(w http.ResponseWriter, r *http.Request) {
+	rec := recover()
+	if rec == nil {
+		return
+	}
+
+	// Obtain source of panic
+	// From: https://gist.github.com/swdunlop/9629168
+	var name, file string // function name, file name
+	var line int
+	var pc [16]uintptr
+	n := runtime.Callers(3, pc[:])
+	for _, pc := range pc[:n] {
+		fn := runtime.FuncForPC(pc)
+		if fn == nil {
+			continue
+		}
+		file, line = fn.FileLine(pc)
+		name = fn.Name()
+		if !strings.HasPrefix(name, "runtime.") {
+			break
+		}
+	}
+
+	// Trim file path
+	delim := "/caddy/"
+	pkgPathPos := strings.Index(file, delim)
+	if pkgPathPos > -1 && len(file) > pkgPathPos+len(delim) {
+		file = file[pkgPathPos+len(delim):]
+	}
+
+	panicMsg := fmt.Sprintf("%s [PANIC %s] %s:%d - %v", time.Now().Format(timeFormat), r.URL.String(), file, line, rec)
+	if h.Debug {
+		// Write error and stack trace to the response rather than to a log
+		var stackBuf [4096]byte
+		stack := stackBuf[:runtime.Stack(stackBuf[:], false)]
+		httpserver.WriteTextResponse(w, http.StatusInternalServerError, fmt.Sprintf("%s\n\n%s", panicMsg, stack))
+	} else {
+		// Currently we don't use the function name, since file:line is more conventional
+		h.Log.Printf(panicMsg)
+		h.errorPage(w, r, http.StatusInternalServerError)
+	}
+}
+
+const timeFormat = "02/Jan/2006:15:04:05 -0700"

caddyhttp/errors/errors_test.go

@@ -0,0 +1,168 @@
+package errors
+
+import (
+	"bytes"
+	"errors"
+	"fmt"
+	"log"
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"path/filepath"
+	"strconv"
+	"strings"
+	"testing"
+
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func TestErrors(t *testing.T) {
+	// create a temporary page
+	path := filepath.Join(os.TempDir(), "errors_test.html")
+	f, err := os.Create(path)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer os.Remove(path)
+
+	const content = "This is a error page"
+	_, err = f.WriteString(content)
+	if err != nil {
+		t.Fatal(err)
+	}
+	f.Close()
+
+	buf := bytes.Buffer{}
+	em := ErrorHandler{
+		ErrorPages: map[int]string{
+			http.StatusNotFound:  path,
+			http.StatusForbidden: "not_exist_file",
+		},
+		Log: log.New(&buf, "", 0),
+	}
+	_, notExistErr := os.Open("not_exist_file")
+
+	testErr := errors.New("test error")
+	tests := []struct {
+		next         httpserver.Handler
+		expectedCode int
+		expectedBody string
+		expectedLog  string
+		expectedErr  error
+	}{
+		{
+			next:         genErrorHandler(http.StatusOK, nil, "normal"),
+			expectedCode: http.StatusOK,
+			expectedBody: "normal",
+			expectedLog:  "",
+			expectedErr:  nil,
+		},
+		{
+			next:         genErrorHandler(http.StatusMovedPermanently, testErr, ""),
+			expectedCode: http.StatusMovedPermanently,
+			expectedBody: "",
+			expectedLog:  fmt.Sprintf("[ERROR %d %s] %v\n", http.StatusMovedPermanently, "/", testErr),
+			expectedErr:  testErr,
+		},
+		{
+			next:         genErrorHandler(http.StatusBadRequest, nil, ""),
+			expectedCode: 0,
+			expectedBody: fmt.Sprintf("%d %s\n", http.StatusBadRequest,
+				http.StatusText(http.StatusBadRequest)),
+			expectedLog: "",
+			expectedErr: nil,
+		},
+		{
+			next:         genErrorHandler(http.StatusNotFound, nil, ""),
+			expectedCode: 0,
+			expectedBody: content,
+			expectedLog:  "",
+			expectedErr:  nil,
+		},
+		{
+			next:         genErrorHandler(http.StatusForbidden, nil, ""),
+			expectedCode: 0,
+			expectedBody: fmt.Sprintf("%d %s\n", http.StatusForbidden,
+				http.StatusText(http.StatusForbidden)),
+			expectedLog: fmt.Sprintf("[NOTICE %d /] could not load error page: %v\n",
+				http.StatusForbidden, notExistErr),
+			expectedErr: nil,
+		},
+	}
+
+	req, err := http.NewRequest("GET", "/", nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	for i, test := range tests {
+		em.Next = test.next
+		buf.Reset()
+		rec := httptest.NewRecorder()
+		code, err := em.ServeHTTP(rec, req)
+
+		if err != test.expectedErr {
+			t.Errorf("Test %d: Expected error %v, but got %v",
+				i, test.expectedErr, err)
+		}
+		if code != test.expectedCode {
+			t.Errorf("Test %d: Expected status code %d, but got %d",
+				i, test.expectedCode, code)
+		}
+		if body := rec.Body.String(); body != test.expectedBody {
+			t.Errorf("Test %d: Expected body %q, but got %q",
+				i, test.expectedBody, body)
+		}
+		if log := buf.String(); !strings.Contains(log, test.expectedLog) {
+			t.Errorf("Test %d: Expected log %q, but got %q",
+				i, test.expectedLog, log)
+		}
+	}
+}
+
+func TestVisibleErrorWithPanic(t *testing.T) {
+	const panicMsg = "I'm a panic"
+	eh := ErrorHandler{
+		ErrorPages: make(map[int]string),
+		Debug:      true,
+		Next: httpserver.HandlerFunc(func(w http.ResponseWriter, r *http.Request) (int, error) {
+			panic(panicMsg)
+		}),
+	}
+
+	req, err := http.NewRequest("GET", "/", nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	rec := httptest.NewRecorder()
+
+	code, err := eh.ServeHTTP(rec, req)
+
+	if code != 0 {
+		t.Errorf("Expected error handler to return 0 (it should write to response), got status %d", code)
+	}
+	if err != nil {
+		t.Errorf("Expected error handler to return nil error (it should panic!), but got '%v'", err)
+	}
+
+	body := rec.Body.String()
+
+	if !strings.Contains(body, "[PANIC /] caddyhttp/errors/errors_test.go") {
+		t.Errorf("Expected response body to contain error log line, but it didn't:\n%s", body)
+	}
+	if !strings.Contains(body, panicMsg) {
+		t.Errorf("Expected response body to contain panic message, but it didn't:\n%s", body)
+	}
+	if len(body) < 500 {
+		t.Errorf("Expected response body to contain stack trace, but it was too short: len=%d", len(body))
+	}
+}
+
+func genErrorHandler(status int, err error, body string) httpserver.Handler {
+	return httpserver.HandlerFunc(func(w http.ResponseWriter, r *http.Request) (int, error) {
+		if len(body) > 0 {
+			w.Header().Set("Content-Length", strconv.Itoa(len(body)))
+			fmt.Fprint(w, body)
+		}
+		return status, err
+	})
+}

caddyhttp/errors/setup.go

@@ -0,0 +1,152 @@
+package errors
+
+import (
+	"io"
+	"log"
+	"os"
+	"path/filepath"
+	"strconv"
+
+	"github.com/hashicorp/go-syslog"
+	"github.com/mholt/caddy"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+// setup configures a new errors middleware instance.
+func setup(c *caddy.Controller) error {
+	handler, err := errorsParse(c)
+	if err != nil {
+		return err
+	}
+
+	// Open the log file for writing when the server starts
+	c.OnStartup(func() error {
+		var err error
+		var writer io.Writer
+
+		switch handler.LogFile {
+		case "visible":
+			handler.Debug = true
+		case "stdout":
+			writer = os.Stdout
+		case "stderr":
+			writer = os.Stderr
+		case "syslog":
+			writer, err = gsyslog.NewLogger(gsyslog.LOG_ERR, "LOCAL0", "caddy")
+			if err != nil {
+				return err
+			}
+		default:
+			if handler.LogFile == "" {
+				writer = os.Stderr // default
+				break
+			}
+
+			var file *os.File
+			file, err = os.OpenFile(handler.LogFile, os.O_RDWR|os.O_CREATE|os.O_APPEND, 0644)
+			if err != nil {
+				return err
+			}
+			if handler.LogRoller != nil {
+				file.Close()
+
+				handler.LogRoller.Filename = handler.LogFile
+
+				writer = handler.LogRoller.GetLogWriter()
+			} else {
+				writer = file
+			}
+		}
+
+		handler.Log = log.New(writer, "", 0)
+		return nil
+	})
+
+	httpserver.GetConfig(c).AddMiddleware(func(next httpserver.Handler) httpserver.Handler {
+		handler.Next = next
+		return handler
+	})
+
+	return nil
+}
+
+func errorsParse(c *caddy.Controller) (*ErrorHandler, error) {
+	// Very important that we make a pointer because the startup
+	// function that opens the log file must have access to the
+	// same instance of the handler, not a copy.
+	handler := &ErrorHandler{ErrorPages: make(map[int]string)}
+
+	cfg := httpserver.GetConfig(c)
+
+	optionalBlock := func() (bool, error) {
+		var hadBlock bool
+
+		for c.NextBlock() {
+			hadBlock = true
+
+			what := c.Val()
+			if !c.NextArg() {
+				return hadBlock, c.ArgErr()
+			}
+			where := c.Val()
+
+			if what == "log" {
+				if where == "visible" {
+					handler.Debug = true
+				} else {
+					handler.LogFile = where
+					if c.NextArg() {
+						if c.Val() == "{" {
+							c.IncrNest()
+							logRoller, err := httpserver.ParseRoller(c)
+							if err != nil {
+								return hadBlock, err
+							}
+							handler.LogRoller = logRoller
+						}
+					}
+				}
+			} else {
+				// Error page; ensure it exists
+				where = filepath.Join(cfg.Root, where)
+				f, err := os.Open(where)
+				if err != nil {
+					log.Printf("[WARNING] Unable to open error page '%s': %v", where, err)
+				}
+				f.Close()
+
+				whatInt, err := strconv.Atoi(what)
+				if err != nil {
+					return hadBlock, c.Err("Expecting a numeric status code, got '" + what + "'")
+				}
+				handler.ErrorPages[whatInt] = where
+			}
+		}
+		return hadBlock, nil
+	}
+
+	for c.Next() {
+		// weird hack to avoid having the handler values overwritten.
+		if c.Val() == "}" {
+			continue
+		}
+		// Configuration may be in a block
+		hadBlock, err := optionalBlock()
+		if err != nil {
+			return handler, err
+		}
+
+		// Otherwise, the only argument would be an error log file name or 'visible'
+		if !hadBlock {
+			if c.NextArg() {
+				if c.Val() == "visible" {
+					handler.Debug = true
+				} else {
+					handler.LogFile = c.Val()
+				}
+			}
+		}
+	}
+
+	return handler, nil
+}

caddyhttp/errors/setup_test.go

@@ -0,0 +1,155 @@
+package errors
+
+import (
+	"testing"
+
+	"github.com/mholt/caddy"
+	"github.com/mholt/caddy/caddyhttp/httpserver"
+)
+
+func TestSetup(t *testing.T) {
+	c := caddy.NewTestController("http", `errors`)
+	err := setup(c)
+	if err != nil {
+		t.Errorf("Expected no errors, got: %v", err)
+	}
+	mids := httpserver.GetConfig(c).Middleware()
+	if len(mids) == 0 {
+		t.Fatal("Expected middlewares, was nil instead")
+	}
+
+	handler := mids[0](httpserver.EmptyNext)
+	myHandler, ok := handler.(*ErrorHandler)
+	if !ok {
+		t.Fatalf("Expected handler to be type ErrorHandler, got: %#v", handler)
+	}
+
+	if myHandler.LogFile != "" {
+		t.Errorf("Expected '%s' as the default LogFile", "")
+	}
+	if myHandler.LogRoller != nil {
+		t.Errorf("Expected LogRoller to be nil, got: %v", *myHandler.LogRoller)
+	}
+	if !httpserver.SameNext(myHandler.Next, httpserver.EmptyNext) {
+		t.Error("'Next' field of handler was not set properly")
+	}
+
+	// Test Startup function -- TODO
+	// if len(c.Startup) == 0 {
+	// 	t.Fatal("Expected 1 startup function, had 0")
+	// }
+	// c.Startup[0]()
+	// if myHandler.Log == nil {
+	// 	t.Error("Expected Log to be non-nil after startup because Debug is not enabled")
+	// }
+}
+
+func TestErrorsParse(t *testing.T) {
+	tests := []struct {
+		inputErrorsRules     string
+		shouldErr            bool
+		expectedErrorHandler ErrorHandler
+	}{
+		{`errors`, false, ErrorHandler{
+			LogFile: "",
+		}},
+		{`errors errors.txt`, false, ErrorHandler{
+			LogFile: "errors.txt",
+		}},
+		{`errors visible`, false, ErrorHandler{
+			LogFile: "",
+			Debug:   true,
+		}},
+		{`errors { log visible }`, false, ErrorHandler{
+			LogFile: "",
+			Debug:   true,
+		}},
+		{`errors { log errors.txt
+        404 404.html
+        500 500.html
+}`, false, ErrorHandler{
+			LogFile: "errors.txt",
+			ErrorPages: map[int]string{
+				404: "404.html",
+				500: "500.html",
+			},
+		}},
+		{`errors { log errors.txt { size 2 age 10 keep 3 } }`, false, ErrorHandler{
+			LogFile: "errors.txt",
+			LogRoller: &httpserver.LogRoller{
+				MaxSize:    2,
+				MaxAge:     10,
+				MaxBackups: 3,
+				LocalTime:  true,
+			},
+		}},
+		{`errors { log errors.txt {
+            size 3
+            age 11
+            keep 5
+        }
+        404 404.html
+        503 503.html
+}`, false, ErrorHandler{
+			LogFile: "errors.txt",
+			ErrorPages: map[int]string{
+				404: "404.html",
+				503: "503.html",
+			},
+			LogRoller: &httpserver.LogRoller{
+				MaxSize:    3,
+				MaxAge:     11,
+				MaxBackups: 5,
+				LocalTime:  true,
+			},
+		}},
+	}
+	for i, test := range tests {
+		actualErrorsRule, err := errorsParse(caddy.NewTestController("http", test.inputErrorsRules))
+
+		if err == nil && test.shouldErr {
+			t.Errorf("Test %d didn't error, but it should have", i)
+		} else if err != nil && !test.shouldErr {
+			t.Errorf("Test %d errored, but it shouldn't have; got '%v'", i, err)
+		}
+		if actualErrorsRule.LogFile != test.expectedErrorHandler.LogFile {
+			t.Errorf("Test %d expected LogFile to be %s, but got %s",
+				i, test.expectedErrorHandler.LogFile, actualErrorsRule.LogFile)
+		}
+		if actualErrorsRule.Debug != test.expectedErrorHandler.Debug {
+			t.Errorf("Test %d expected Debug to be %v, but got %v",
+				i, test.expectedErrorHandler.Debug, actualErrorsRule.Debug)
+		}
+		if actualErrorsRule.LogRoller != nil && test.expectedErrorHandler.LogRoller == nil || actualErrorsRule.LogRoller == nil && test.expectedErrorHandler.LogRoller != nil {
+			t.Fatalf("Test %d expected LogRoller to be %v, but got %v",
+				i, test.expectedErrorHandler.LogRoller, actualErrorsRule.LogRoller)
+		}
+		if len(actualErrorsRule.ErrorPages) != len(test.expectedErrorHandler.ErrorPages) {
+			t.Fatalf("Test %d expected %d no of Error pages, but got %d ",
+				i, len(test.expectedErrorHandler.ErrorPages), len(actualErrorsRule.ErrorPages))
+		}
+		if actualErrorsRule.LogRoller != nil && test.expectedErrorHandler.LogRoller != nil {
+			if actualErrorsRule.LogRoller.Filename != test.expectedErrorHandler.LogRoller.Filename {
+				t.Fatalf("Test %d expected LogRoller Filename to be %s, but got %s",
+					i, test.expectedErrorHandler.LogRoller.Filename, actualErrorsRule.LogRoller.Filename)
+			}
+			if actualErrorsRule.LogRoller.MaxAge != test.expectedErrorHandler.LogRoller.MaxAge {
+				t.Fatalf("Test %d expected LogRoller MaxAge to be %d, but got %d",
+					i, test.expectedErrorHandler.LogRoller.MaxAge, actualErrorsRule.LogRoller.MaxAge)
+			}
+			if actualErrorsRule.LogRoller.MaxBackups != test.expectedErrorHandler.LogRoller.MaxBackups {
+				t.Fatalf("Test %d expected LogRoller MaxBackups to be %d, but got %d",
+					i, test.expectedErrorHandler.LogRoller.MaxBackups, actualErrorsRule.LogRoller.MaxBackups)
+			}
+			if actualErrorsRule.LogRoller.MaxSize != test.expectedErrorHandler.LogRoller.MaxSize {
+				t.Fatalf("Test %d expected LogRoller MaxSize to be %d, but got %d",
+					i, test.expectedErrorHandler.LogRoller.MaxSize, actualErrorsRule.LogRoller.MaxSize)
+			}
+			if actualErrorsRule.LogRoller.LocalTime != test.expectedErrorHandler.LogRoller.LocalTime {
+				t.Fatalf("Test %d expected LogRoller LocalTime to be %t, but got %t",
+					i, test.expectedErrorHandler.LogRoller.LocalTime, actualErrorsRule.LogRoller.LocalTime)
+			}
+		}
+	}
+
+}