Merge branch 'kovidgoyal:master' into master

recipes/mediapart.recipe

@@ -82,13 +82,25 @@ class Mediapart(BasicNewsRecipe):
     feeds = []
 
     sections = [
-        'france', 'international', 'economie', 'culture-idees', 'politique', 'ecologie', 'fil-dactualites'
+        'france', 'international', 'economie', 'culture-et-idees', 'politique', 'ecologie', 'fil-dactualites'
     ]
 
+    mediapart_feed = 'https://www.mediapart.fr/articles/feed'
     for sec in sections:
-        a = 'https://news.google.com/rss/search?q=when:27h+allinurl:mediapart.fr%2Fjournal{}&hl=fr-FR&gl=FR&ceid=FR:fr'
-        feeds.append((sec.capitalize(), a.format('%2F' + sec + '%2F')))
-    feeds.append(('Autres', a.format('')))
+        feeds.append((sec.capitalize(), mediapart_feed))
+    feeds.append(('Autres', mediapart_feed))
+
+    def parse_feeds(self):
+        feeds = BasicNewsRecipe.parse_feeds(self)
+        for feed in feeds:
+            feed_name = feed.title.lower()
+            for article in feed.articles:
+                if feed_name != 'autres'  and feed_name not in article.url:
+                    feed.articles.remove(article)
+                if feed_name == 'autres' and any(section in article.url for section in self.sections):
+                    feed.articles.remove(article)
+
+        return feeds
 
     def get_browser(self):
         # -- Handle login
@@ -99,11 +111,7 @@ class Mediapart(BasicNewsRecipe):
             br['email'] = self.username
             br['password'] = self.password
             br.submit()
-
-        # -- Handle Google's cookies consent page
-        br.open('https://news.google.com')
-        br.select_form(action="https://consent.google.com/save")
-        br.submit()
+            
         return br
 
     def default_cover(self, cover_file):

setup/build.py

@@ -732,7 +732,10 @@ sip-file = {os.path.basename(sipf)!r}
         cwd = os.getcwd()
         try:
             os.chdir(os.path.join(src_dir, 'build'))
-            self.check_call([self.env.make] + ([] if iswindows else ['-j%d'%(os.cpu_count() or 1)]))
+            env = os.environ.copy()
+            if is_macos_universal_build:
+                env['ARCHS'] = 'x86_64 arm64'
+            self.check_call([self.env.make] + ([] if iswindows else ['-j%d'%(os.cpu_count() or 1)]), env=env)
             e = 'pyd' if iswindows else 'so'
             m = glob.glob(f'{ext.name}/{ext.name}.*{e}')
             if not m:

src/calibre/ebooks/metadata/epub.py

@@ -10,13 +10,14 @@ __copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
 import io
 import os
 import posixpath
-from contextlib import closing
+from contextlib import closing, suppress
 
 from calibre import CurrentDir
 from calibre.ebooks.metadata.opf import get_metadata as get_metadata_from_opf
 from calibre.ebooks.metadata.opf import set_metadata as set_metadata_opf
 from calibre.ebooks.metadata.opf2 import OPF
 from calibre.ptempfile import TemporaryDirectory
+from calibre.utils.imghdr import what as what_image_type
 from calibre.utils.localunzip import LocalZipFile
 from calibre.utils.xml_parse import safe_xml_fromstring
 from calibre.utils.zipfile import BadZipfile, ZipFile, safe_replace
@@ -36,7 +37,7 @@ class ContainerException(OCFException):
 
 class Container(dict):
 
-    def __init__(self, stream=None):
+    def __init__(self, stream=None, file_exists=None):
         if not stream:
             return
         container = safe_xml_fromstring(stream.read())
@@ -49,6 +50,12 @@ class Container(dict):
             mt, fp = rootfile.get('media-type'), rootfile.get('full-path')
             if not mt or not fp:
                 raise EPubException("<rootfile/> element malformed")
+
+            if file_exists and not file_exists(fp):
+                # Some Kobo epubs have multiple rootfile entries, but only one
+                # exists.  Ignore the ones that don't exist.
+                continue
+
             self[mt] = fp
 
 
@@ -95,7 +102,7 @@ class OCFReader(OCF):
 
         try:
             with closing(self.open(OCF.CONTAINER_PATH)) as f:
-                self.container = Container(f)
+                self.container = Container(f, self.exists)
         except KeyError:
             raise EPubException("missing OCF container.xml file")
         self.opf_path = self.container[OPF.MIMETYPE]
@@ -125,6 +132,14 @@ class OCFReader(OCF):
     def read_bytes(self, name):
         return self.open(name).read()
 
+    def exists(self, path):
+        try:
+            self.open(path).close()
+            return True
+        except OSError:
+            return False
+
+
 
 class OCFZipReader(OCFReader):
 
@@ -153,6 +168,13 @@ class OCFZipReader(OCFReader):
     def read_bytes(self, name):
         return self.archive.read(name)
 
+    def exists(self, path):
+        try:
+            self.archive.getinfo(path)
+            return True
+        except KeyError:
+            return False
+
 
 def get_zip_reader(stream, root=None):
     try:
@@ -192,6 +214,19 @@ def render_cover(cpage, zf, reader=None):
             cpage = os.path.join(tdir, cpage)
             if not os.path.exists(cpage):
                 return
+
+            with suppress(Exception):
+                # In the case of manga, the first spine item may be an image
+                # already, so treat it as a raster cover.
+                file_format = what_image_type(cpage)
+                if file_format == "jpeg":
+                    # Only JPEG is allowed since elsewhere we assume raster covers
+                    # are JPEG.  In principle we could convert other image formats
+                    # but this is already an out-of-spec case that happens to
+                    # arise in books from some stores.
+                    with open(cpage, "rb") as source:
+                        return source.read()
+
             return render_html_svg_workaround(cpage, default_log, root=tdir)