Segmentation fault with unoconv

[Sworddragon]

BPO	20756
Nosy	@ned-deily
Files	backtrace.txt: GDB backtrace test.odt: LibreOffice test document

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2014-06-21.19:27:09.769>
created_at = <Date 2014-02-24.02:13:53.417>
labels = ['interpreter-core', 'invalid', 'type-crash']
title = 'Segmentation fault with unoconv'
updated_at = <Date 2014-06-21.19:32:54.792>
user = 'https://bugs.python.org/Sworddragon'

bugs.python.org fields:

activity = <Date 2014-06-21.19:32:54.792>
actor = 'ned.deily'
assignee = 'none'
closed = True
closed_date = <Date 2014-06-21.19:27:09.769>
closer = 'Sworddragon'
components = ['Interpreter Core']
creation = <Date 2014-02-24.02:13:53.417>
creator = 'Sworddragon'
dependencies = []
files = ['34206', '34207']
hgrepos = []
issue_num = 20756
keywords = []
message_count = 5.0
messages = ['212052', '212054', '212055', '212056', '221192']
nosy_count = 2.0
nosy_names = ['ned.deily', 'Sworddragon']
pr_nums = []
priority = 'normal'
resolution = 'not a bug'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'crash'
url = 'https://bugs.python.org/issue20756'
versions = ['Python 3.4']

[Sworddragon]

With Python 3.4.0 RC1 on using the command "unoconv -o test.pdf test.odt" I'm getting a segmentation fault. In the attachments are the used LibreOffice document and a GDB backtrace. The used version of unoconv was 0.6-6 from Ubuntu 14.04 dev and can be currently found here: http://packages.ubuntu.com/de/source/trusty/unoconv

[ned-deily]

I don't know anything about these programs but it appears libpyuno.so is linked with a specific version of Python 3. Was it rebuilt linked against Python 3.4, instead of Python 3.3?

http://packages.ubuntu.com/en/trusty/python3-uno

[Sworddragon]

Was it rebuilt linked against Python 3.4, instead of Python 3.3?

I don't know. Is ../Python/pystate.c that throws the error not a part of Python?

[ned-deily]

The point is that you may be inadvertently mixing a shared library (libpyuno.so) linked to a Python 3.3 interpreter with a base Python 3.4 interpreter. Unless python3-uno is very careful to use the stable ABI there is no guarantee of binary level compatibility between Python versions (3.3 vs 3.4):

http://docs.python.org/dev/c-api/stable.html

[Sworddragon]

I have retested this with the correct linked version and it is working fine now so I'm closing this ticket.