Commit Graph

27165 Commits

Author SHA1 Message Date
Kovid Goyal 1f2aa8a55b Allow merging of icu branch into trunk by falling back to the old icu module if the old binary plugin is detected. 2014-03-08 22:18:29 +05:30
Kovid Goyal b76cc3e9ab Speed up searching a little by using a dedicated function for testing if a string contains a substring using primary collation (replaces using primary_find() 2014-03-08 22:08:31 +05:30
Kovid Goyal 4eaee89487 Fix ICU find returning incorrect position and length parameters when non-BMP characters are present on wide python builds 2014-03-08 21:46:46 +05:30
Kovid Goyal 27327e811b Clearer error message when compiling on python >= 3.3 2014-03-08 21:19:05 +05:30
Kovid Goyal f078cd7168 Performance improvements and code cleanup for the ICU module 2014-03-08 21:13:26 +05:30
Kovid Goyal b8e414f18b Revert a part of the previous commit that was left in by mistake and also add a test for handling of positions when the haystack contains non-BMP chars 2014-03-08 21:12:38 +05:30
Kovid Goyal bd88666bb0 Command line interface to filesystem matcher
Also fix a couple of bugs in the matcher algorithms
2014-03-08 20:55:20 +05:30
Kovid Goyal 09be666ea0 When reading metadata from filenames, do not apply the fallback regexp to read metadata if the user specified regexp puts the entire filename into the title. The fallback is only used if the user specified expression does not match the filename at all. 2014-03-08 18:11:08 +05:30
Kovid Goyal 35c837b839 Move matcher module into the utils package 2014-03-08 11:24:15 +05:30
Kovid Goyal b672f4ed11 Subsequence matcher: Use primary collation 2014-03-08 10:59:19 +05:30
Kovid Goyal 6e9afc0398 ... 2014-03-08 10:56:31 +05:30
Kovid Goyal 37ac52ad32 Allow capsule based access to the ICU collator 2014-03-08 09:28:02 +05:30
Kovid Goyal 5816c4aaeb ... 2014-03-08 08:27:27 +05:30
Kovid Goyal a33265cf46 Content server: Fix (maybe) an error on some windows computers with a non-standard default encoding
See http://www.mobileread.com/forums/showthread.php?t=235366
2014-03-08 08:08:46 +05:30
Kovid Goyal 606ee69582 ... 2014-03-08 07:28:46 +05:30
Kovid Goyal 290462909f Linux binary install script: Fix error on linux systems where the system python has an encoding of None set on stdout. Assume encoding is utf-8 in this case. 2014-03-07 21:51:00 +05:30
Kovid Goyal 0dc884efc5 Some minor tweaks to ensure correctness
Also fix compilation of the subsequence matcher on OS X
v1.27.0
2014-03-07 09:56:07 +05:30
Kovid Goyal 90868e0262 pep8 2014-03-07 09:06:52 +05:30
Kovid Goyal d410d11e55 pep8 2014-03-07 09:05:58 +05:30
Kovid Goyal 6f30914a52 version 1.27 2014-03-07 09:03:37 +05:30
Kovid Goyal 286312b63a Merge branch 'master' of https://github.com/cbhaley/calibre 2014-03-07 08:23:38 +05:30
Charles Haley 72c26969c8 Undo the update CC version number. It should be updated *after* CC is generally released. 2014-03-07 03:14:21 +01:00
Kovid Goyal 8b7956f996 An option to multiple add books from inside a ZIP or RAR file. Right click the Add Books button and choose 'Add multiple books from archive'. 2014-03-06 22:07:18 +05:30
Kovid Goyal bdb2d0a5d9 MOBI Input: Fix empty <a> tags not being handled correctly in rare cases when the markup is bad enough to be parsed only using the HTML 5 algorithm. Fixes #1288430 [Private bug](https://bugs.launchpad.net/calibre/+bug/1288430) 2014-03-06 21:47:43 +05:30
Kovid Goyal 20dd82d501 PEP8 2014-03-06 21:35:51 +05:30
Kovid Goyal 05ef329823 test for character_name() 2014-03-06 19:29:30 +05:30
Kovid Goyal efafb69fcd Fix icu_character_name not working for non-BMP chars on windows 2014-03-06 19:05:57 +05:30
Kovid Goyal 4b954101b7 Use macro instead of sizeof as that is recommended by the python developers 2014-03-06 17:47:48 +05:30
Kovid Goyal 68a37c565d Use new roundtrip code for ICU get_character_name 2014-03-06 17:35:18 +05:30
Kovid Goyal a8d8f92e9f Edit book: Fix column number of cursor one less than it should be.
Also fix non-BMP characters before cursor not being reported correctly.
2014-03-06 17:30:35 +05:30
Kovid Goyal 4c9e1733ec Return NULL when decoding from UTF-32 fails 2014-03-06 17:22:41 +05:30
Kovid Goyal 82b6335a3c Use the faster py->icu string conversion in the matcher 2014-03-06 16:58:46 +05:30
Kovid Goyal df6a06c8b7 Faster roundtripping of strings between ICU and python 2014-03-06 16:41:27 +05:30
Kovid Goyal e05c4e669b Tiny performance improvement 2014-03-06 14:36:17 +05:30
Kovid Goyal 0930fef90a API for normalization using ICU 2014-03-06 14:27:15 +05:30
Kovid Goyal 63cba4c884 Insert special char: Allow searching for non BMP characters, by using the ICU database of names rather than python's outdated one. 2014-03-06 13:10:48 +05:30
Kovid Goyal b36c6211b0 Merge branch 'master' of https://github.com/cbhaley/calibre 2014-03-06 12:45:29 +05:30
Kovid Goyal fe1b7fe739 ... 2014-03-06 12:44:22 +05:30
Kovid Goyal 717eba06b9 ... 2014-03-06 12:42:49 +05:30
Charles Haley c048add7cf Update CC version number in preparation for general release next week. 2014-03-06 08:11:42 +01:00
Kovid Goyal 935af249a4 Fix warnings when building matcher in Visual Studio 2014-03-06 11:54:45 +05:30
Kovid Goyal 22ca2bcac4 Edit Book: Add support for showing the names of non-BMP characters on narrow python builds (windows and OS X) 2014-03-06 11:45:11 +05:30
Kovid Goyal 3901051e2e Refactor subsequence matcher to not use recursion and support multithreading. Also add a python implementation for easy experimentation with the algorithm. 2014-03-06 10:11:27 +05:30
Kovid Goyal c634b7c946 ... 2014-03-06 09:31:25 +05:30
Kovid Goyal fea4b9cbac Update American Thinker
Fixes #1288371 [American Thinker news](https://bugs.launchpad.net/calibre/+bug/1288371)
2014-03-05 23:54:31 +05:30
Kovid Goyal f19f998b25 Change docs for in process servers to discourage their use 2014-03-05 22:57:20 +05:30
Kovid Goyal fc7874a993 Update HNA and Heise 2014-03-05 22:15:05 +05:30
Kovid Goyal 4398dda15f Use a 2D array for memoization, gives clearer code 2014-03-05 08:44:02 +05:30
Kovid Goyal ed45f1a6aa ... 2014-03-05 08:13:18 +05:30
Kovid Goyal 58c084e68b Do not show system tray notification popups while analyzing books on device, as they can be slow on systems that display multiple messages serially instead of replacing the last message 2014-03-05 08:07:18 +05:30