Windows: Edit book: Fix a crash when using the check book tool with a book that contains malformed markup. Fixes #2045133 [Editor closes immediately if epub-check finds parsing-failures in too much text-files](https://bugs.launchpad.net/calibre/+bug/2045133)

Bloody lxml doesnt initialize libxml2 correctly.

bypy/sources.json

@@ -321,17 +321,17 @@
     {
         "name": "libxml2",
         "unix": {
-            "filename": "libxml2-2.11.5.tar.xz",
-            "hash": "sha256:3727b078c360ec69fa869de14bd6f75d7ee8d36987b071e6928d4720a28df3a6",
-            "urls": ["https://download.gnome.org/sources/libxml2/2.11/{filename}"]
+            "filename": "libxml2-2.12.1.tar.xz",
+            "hash": "sha256:8982b9ccdf7f456e30d8f7012d50858c6623e495333b6191def455c7e95427eb",
+            "urls": ["https://download.gnome.org/sources/libxml2/2.12/{filename}"]
         }
     },
 
     {
         "name": "libxslt",
         "unix": {
-            "filename": "libxslt-1.1.38.tar.xz",
-            "hash": "sha256:1f32450425819a09acaff2ab7a5a7f8a2ec7956e505d7beeb45e843d0e1ecab1",
+            "filename": "libxslt-1.1.39.tar.xz",
+            "hash": "sha256:2a20ad621148339b0759c4d4e96719362dee64c9a096dbba625ba053846349f0",
             "urls": ["https://download.gnome.org/sources/libxslt/1.1/{filename}"]
         }
     },

bypy/windows/site.py

@@ -60,6 +60,16 @@ def set_quit():
     builtins.exit = _sitebuiltins.Quitter('exit', eof)
 
 
+def workaround_lxml_bug():
+    # Without calling xmlInitParser() import lxml causes a segfault
+    import ctypes
+    x = ctypes.WinDLL('libxml2.dll')
+    x.xmlInitParser()
+    workaround_lxml_bug.libxml2 = x
+    from lxml import etree
+    del etree
+
+
 def main():
     sys.meta_path.insert(0, PydImporter())
     os.add_dll_directory(os.path.abspath(os.path.join(sys.app_dir, 'app', 'bin')))
@@ -75,6 +85,8 @@ def main():
     set_helper()
     set_quit()
 
+    workaround_lxml_bug()
+
     return run_entry_point()
 
 

src/calibre/ebooks/oeb/polish/check/parsing.py

@@ -292,7 +292,7 @@ def check_xml_parsing(name, mt, raw):
     except XMLSyntaxError as err:
         try:
             line, col = err.position
-        except:
+        except Exception:
             line = col = None
         return errors + [errcls(error_message(err), name, line, col)]
     except Exception as err: