| Summary: | Broken IBus IM with the mga5 firefox update to 38.1.0. | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Edward d'Auvergne <true.bugman> |
| Component: | RPM Packages | Assignee: | QA Team <qa-bugs> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | major | ||
| Priority: | Normal | CC: | davidwhodgins, fundawang, joequant, sysadmin-bugs, yochenhsieh |
| Version: | 5 | Keywords: | validated_update |
| Target Milestone: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | advisory MGA5-64-OK MGA5-32-OK | ||
| Source RPM: | ibus | CVE: | |
| Status comment: | |||
|
Description
Edward d'Auvergne
2015-07-07 09:38:13 CEST
Strangely, downgrading to the original firefox-31.7.0-1.mga5.x86_64.rpm version does not restore IBus support. I am typing with Hindi turned on. Installing the lib64suil* packages fixed the issue (or it could have been ibus-devel, though unlikely). However, even more strangely, uninstalling them does not revert the behaviour of not being able to use IBus!
Actually, with a lot of testing of restarting the computer with IBus set to either English or Chinese, the bug appears to be highly inconsistent. Sometimes it works in Firefox, and sometimes it does not. I am wondering if it has something to do with this obscure bug fix:
https://github.com/ibus/ibus/commit/83451e476b5ec98aa38e333038077f9e5b2b087d
This was in the IBus 1.5.10 bugfix release (https://github.com/ibus/ibus/releases). Or maybe this other 1.5.10 bugfix:
https://github.com/ibus/ibus/commit/1b668f667eaef08dfbbbdf04a0865ea086f4250a
Mga5 is using 1.5.8 released on 24 Jul 2014. Would it be possible to have 1.5.10 pushed into the testing repositories? IBus 1.5.9 and 1.5.10 are pure bugfix releases, see:
https://github.com/ibus/ibus/releases/tag/1.5.10
https://github.com/ibus/ibus/releases/tag/1.5.9
Samuel Verschelde
2015-07-07 11:30:22 CEST
Assignee:
bugsquad =>
fundawang
Samuel Verschelde
2015-07-07 11:59:07 CEST
CC:
(none) =>
yochenhsieh Hello,
Considering that 1.5.10 contains mostly bugfixes, I submit it to core/updates_testing. Please test it throughly in case of any possible regression or inclusion of new bugs.
Suggested advisory:
========================
ibus 1.5.10 was released just 2 days after we have release freeze for cauldron. It includes many bugfixes that should(might) resolve (un)reported bugs that relate to input methods since then.
References:
https://github.com/ibus/ibus/releases/tag/1.5.10
https://github.com/ibus/ibus/releases/tag/1.5.9
========================
Updated packages in core/updates_testing:
========================
ibus-1.5.10-1.mga5
ibus-gtk-1.5.10-1.mga5
ibus-gtk3-1.5.10-1.mga5
ibus-ui-gtk3-1.5.10-1.mga5
lib(64)ibus1.0_5-1.5.10-1.mga5
lib(64)ibus-gir1.0-1.5.10-1.mga5
ibus-devel-1.5.10-1.mga5
Source RPMs:
ibus-1.5.10-1.mga5.src.rpm
If it's meant as an update candidate, which it looks like, don't forget to assign to QA :) Re-assign to QA. Thanks for reminder. The version/release to be tested is ibus-1.5.10-1.mga5, as noted in comment#3. Assignee:
fundawang =>
qa-bugs Advisory committed to svn. CC:
(none) =>
davidwhodgins Testing completed Mageia 5 i586 for ibus-1.5.10-1.mga5.src.rpm: - can input Chinese in firefox-38.1.0-1.mga5 - no regression or new bugs so far. It would be helpful if someone can help test x86-64 and input in both i586 and x86-64 programs, e.g. firefox and skype. Just in case I did anything wrong. Whiteboard:
advisory =>
advisory MGA5-32-OK Joseph, Funda are you able to help with testing this one please? CC:
(none) =>
fundawang, joequant I have bad news for you... After a few days retrials, the problem seems still there but not always replicable. I cannot figure out a specific procedure to reproduce it. But this problem does only happen with firefox. A workaround is to run firefox with following command: export GTK_IM_MODULE=xim && firefox Whiteboard:
advisory MGA5-32-OK =>
advisory Extensive testing of the core/updates_testing 1.5.10 version of IBus seems to work fine for me in 32-bit Skype and 64-bit Firefox. One of the fixes was simply a longer timeout, but I cannot replicate the bug any more (possibly because my computer loads so fast that this is not triggered - it's both a fast computer and runs of a SSD). Tested for about two weeks on another new install of mga5 32bit with firefox 38.1 and ibus 1.5.10. I have no problem with this new install and everything seems work fine here. At most circumstances, ibus 1.5.10 will resolve this bug. Sometimes, but not always, the problem still appears (see comment #9) and workaround is available. Adding MGA5-32-OK as per comment 11 and MGA5-64-OK as per comment 10. If I understood your test results correctly, the update is ready to be pushed. Am I right? Whiteboard:
advisory =>
advisory MGA5-64-OK MGA5-32-OK Yes, I think this update is ready. Thanks everyone for the assistance.:) Validating, then. Keywords:
(none) =>
validated_update An update for this issue has been pushed to Mageia Updates repository. http://advisories.mageia.org/MGAA-2015-0075.html Status:
NEW =>
RESOLVED |