Installation problem with expat, ruins SAX support

[SauliusP.]

Hi,

Today after Calibre's update I've noticed, that my DOCX Input plugin does not work anymore. Error is:


Traceback (most recent call last):
File "site.py", line 58, in main
File "site-packages/calibre/utils/ipc/worker.py", line 191, in main
File "site-packages/calibre/gui2/convert/gui_conversion.py", line 31, in gui_convert_override
File "site-packages/calibre/gui2/convert/gui_conversion.py", line 25, in gui_convert
File "site-packages/calibre/ebooks/conversion/plumber.py", line 979, in run
File "site-packages/calibre/customize/conversion.py", line 208, in __call__
File "calibre_plugins.docx_input_plugin.docx_input" , line 29, in convert
File "calibre_plugins.docx_input_plugin.docx_input" , line 58, in _parseStyles
File "xml/sax/__init__.py", line 93, in make_parser
SAXReaderNotAvailable: No parsers found


I have reinstalled the Calibre and there is expat failure during the installation too:

.............

Symlinking /opt/calibre/epub-fix to /usr/bin/epub-fix
Symlinking /opt/calibre/ebook-device to /usr/bin/ebook-device
Symlinking /opt/calibre/calibredb to /usr/bin/calibredb
Setting up bash completion...
Traceback (most recent call last):
File "site-packages/calibre/utils/config.py", line 264, in refresh
File "site-packages/calibre/utils/config.py", line 253, in raw_to_object
File "plistlib.py", line 103, in readPlistFromString
File "plistlib.py", line 78, in readPlist
File "plistlib.py", line 401, in parse
File "xml/parsers/expat.py", line 4, in <module>
ImportError: /opt/calibre/lib/python2.7/lib-dynload/pyexpat.so: undefined symbol: XML_SetHashSalt
Installing bash completion to /etc/bash_completion.d/calibre



Is there any quick workaround with this problem? And I wonder if it would affect other OSes?
I am on Ubuntu 11.10.

[kovidgoyal]

Are you on a 32-bit or a 64-bit OS? Generally speaking dont use expat, lxml is much better.

[kovidgoyal]

The fix is to update the version of libexpat on your system to one that defines XML_SetHashSalt (this is a good idea since this function is needed for a security fix in python), I'll upload a new linux build of calibre with a bundled libexpat in a couple of hours.

[SauliusP.]

64 bit. I had a quick scan-through lxml, but it would require quite a lot of rework. And it supports only startElementNS, not startElement, which is also not that convenient...

Originally Posted by kovidgoyal

Are you on a 32-bit or a 64-bit OS? Generally speaking dont use expat, lxml is much better.

[SauliusP.]

Thank you for new version. I will give lxml a try anyway, if you suggest it is better.

[kovidgoyal]

The way to use lxml is not to write a parser, but to use the builtin parsers, and then use XPath on the parsed document. This is what calibre uses in its conversion pipeline.

[kovidgoyal]

Builds with bundled libexpat are now live.