python crash in lxml with "Element("parent").find("bar")"

[Alexis Murzeau]

Hi,

When doing this on a Debian sid machine:
# apt get install locales-all
$ LC_ALL=zh_CN python3 -c 'from lxml.etree import Element; Element("parent").find("bar")'

Python crashes with a segmentation fault.
Without "LC_ALL=zh_CN", it works fine.

It may be related to this issue in Cython:
[BUG] Possible segfault in exception handling code #4722
https://github.com/cython/cython/issues/4722

This crashes too (using the same env values as in the above bug report):
LC_ALL="" LANG="en_US.ISO8859-15" python3 -c 'from lxml.etree import Element; Element("parent").find("bar")'

This bug is encountered when doing reproducible builds tests using reprotest.

gdb backtrace: see in attachment.

This is with these packages:
 - python3.10 3.10.4-3
 - python3-lxml 4.8.0-1
 - locales-all 2.33-7

Python : sys.version_info(major=3, minor=10, micro=4, releaselevel='final', serial=0)
lxml.etree : (4, 8, 0, 0)
libxml used : (2, 9, 13)
libxml compiled : (2, 9, 12)
libxslt used : (1, 1, 34)
libxslt compiled : (1, 1, 34)

[bastimeyer (bastimeyer)]

The cython issue at
https://github.com/cython/cython/issues/4722
has been fixed, and recent lxml builds seem to be working fine with the updated cython version.

lxml 4.9.0-2 doesn't seem to have been pushed to pypi though:
https://github.com/lxml/lxml/releases/tag/lxml-4.9.0-2
https://pypi.org/project/lxml/#history

Could you please fix that? Since this release is just a couple of hours old, I'm not sure if this is already being looked into.

Thanks!

Related issue:
https://github.com/streamlink/streamlink/issues/4562

[scoder (scoder)]

This has been fixed in lxml 4.9.0.