Issue 112603 - sax: use rtl::ByteSequence for trivial global objects instead of uno::Sequence
Summary: sax: use rtl::ByteSequence for trivial global objects instead of uno::Sequence
Status: CLOSED FIXED
Alias: None
Product: xml
Classification: Code
Component: code (show other issues)
Version: DEV300m83
Hardware: All Linux, all
: P3 Trivial (vote)
Target Milestone: 3.4.1
Assignee: Stephan Bergmann
QA Contact: issues@xml
URL:
Keywords:
Depends on:
Blocks: 107490
  Show dependency tree
 
Reported: 2010-06-22 14:56 UTC by caolanm
Modified: 2010-09-22 09:27 UTC (History)
1 user (show)

See Also:
Issue Type: PATCH
Latest Confirmation in: ---
Developer Difficulty: ---


Attachments
like so (6.59 KB, patch)
2010-06-22 14:57 UTC, caolanm
no flags Details | Diff

Note You need to log in before you can comment on or make changes to this issue.
Description caolanm 2010-06-22 14:56:43 UTC
A global uno::Sequence object is a little problematic in that in its dtor
uno::Sequence looks up the types of its contained objects to do correct release
on them. This needs the cppu infrastructure to be working, but that's rather
problematic at shut-down time where there's some nasty life-cycle issues in cppu. 

Currently the cppu foo leaks, so this isn't a noticeable issue, but a leaking
cppu causes the threads used by e.g. unopkg to not exit correctly and outlive
process exit and eventual destruction of all used data.

For sax, the whole issue could be side-stepped by using rtl::BytesSequences and
toUnoSequence which is a simple inline cast, so no detrimental performance issue
should exist.
Comment 1 caolanm 2010-06-22 14:57:13 UTC
Created attachment 70149 [details]
like so
Comment 2 Mathias_Bauer 2010-06-22 16:13:37 UTC
Stephan, please comment
Comment 3 Stephan Bergmann 2010-06-22 16:18:08 UTC
looks good
Comment 4 caolanm 2010-09-06 09:25:33 UTC
I'll take this then as these are the only cppu::Sequences to ctor before main
and I'd like to clear them out of the way
Comment 6 caolanm 2010-09-06 11:34:46 UTC
cmc->sb: willing to verify this change ?, i.e.
http://hg.services.openoffice.org/cws/cmcfixes78/rev/883150fba4cd
Comment 7 Stephan Bergmann 2010-09-06 11:58:50 UTC
verified by code review
Comment 8 Stephan Bergmann 2010-09-22 09:27:59 UTC
.