From Fedora Project Wiki
(Created page with 'previous meeting | next meeting == Agenda == Time:[http://www.timeanddate.com/worldclock/fixe...')
 
m (internal link cleaning)
 
(4 intermediate revisions by one other user not shown)
Line 1: Line 1:
[[I18N/InputMethods/Meetings/2009-08-19|previous meeting]] | [[I18N/InputMethods/Meetings/2009-09-02|next meeting]]
[[I18N/InputMethods/Meetings/2009-08-19|previous meeting]] | [[I18N/InputMethods/Meetings/2009-09-09|next meeting]]


== Agenda ==
== Agenda ==
Line 9: Line 9:


== Summary==
== Summary==
* @input-methods waiting on anaconda for top package page
* ibus persists after logout: https://bugzilla.redhat.com/show_bug.cgi?id=518970
* ibus-qt and compose key
* ibus-table build failures
* gob2 x86_64 issue
* dlopen/dlclose of im-scim.so causes segfault - https://bugzilla.redhat.com/show_bug.cgi?id=517001


== Log ==
== Log ==
{|
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | IM meeting soon
|-
| colspan="3" |  who's in for the IM meeting?* pravins in
|-
| colspan="3" |  Kaio in
|-
| colspan="3" |  paragan in
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | wonder if phuang_, tagoh3 and fujiwarat are around
|-
| colspan="3" |  dyek2 dychen
|-
| colspan="3" |  and dychen and Kaio
|-
| style="background-color: #42427e" | fujiwarat
| style="color: #42427e" colspan="2" | hi
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | yeah
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | ah hi Kaio
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | hi
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | cool
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | hi
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | quick update on the @input-methods:
|-
| colspan="3" |  so far clumens kind of pushed back saying most people won't understand what "Input Methods" are anyway... (for anaconda top package listing)
|-
| colspan="3" |  anyway we followed up - let's see how it goes
|-
| colspan="3" |  otherwise we may just leave it to post-f12 even
|-
| colspan="3" |  I would rather avoid all the "how do i install IM?" and "why am I using <xim>?" type faq's but anyway
|-
| colspan="3" |  (if we turn off @input-methods)
|-
| style="background-color: #488888" | pravins
| style="color: #488888" colspan="2" | juhp: yeah, that looks nice
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | .bug 518970
|-
| style="background-color: #8c4a4a" | zodbot
| style="color: #8c4a4a" colspan="2" | juhp: Bug 518970 ibus-daemon persists after desktop session - https://bugzilla.redhat.com/show_bug.cgi?id=518970
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | phuang_: any comment?
|-
| colspan="3" |  hmm guess ibus-daemon can run in console too?
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | even with --xim?
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | hmm
|-
| colspan="3" |  dunno :)
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | I did not meet this problem. I will test it later
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | ok please
|-
| colspan="3" |  I can reproduce 100%
|-
| colspan="3" |  anyone else seen it?
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | was about to do it - no lucks on rawhide installation yet, again ;o)
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | ah
|-
| colspan="3" |  anaconda?
|-
| colspan="3" |  I have been getting funny Live spin errors too this week :/
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | dunno. it freezes on creating fs on virtio.
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | I am running an up-to-date pre-alpha live install here
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | guess better trying 12alpha
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | yeah
|-
| colspan="3" |  that might install ;)
|-
| colspan="3" |  .bug 517603
|-
| style="background-color: #8c4a4a" | zodbot
| style="color: #8c4a4a" colspan="2" | juhp: Bug 517603 anaconda failed in get_table - https://bugzilla.redhat.com/show_bug.cgi?id=517603
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | anyway
|-
| colspan="3" |  phuang_: ibus-qt works well for me - well done!
|-
| colspan="3" |  is it hard to get underline in preedit?
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | juhp, Just tested ibus in rawhide, this problem does not happen
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | juhp: Hi
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | phuang_: recent install?
|-
| colspan="3" |  phuang_: you logged in twice?
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | juhp, yeah
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | hmm
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | let me try it again
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | juhp: ibus can run in console. :-)
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | phuang_: still happens for me anyway
|-
| colspan="3" |  hmm perhaps I should try on a fresh install... if only...
|-
| colspan="3" |  phuang_: when is your tree do you know?
|-
| colspan="3" |  dyek2 dychen
|-
| colspan="3" |  dychen: right
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | juhp, I am not sure. But I often execute 'yum update'
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | well maybe let's wait to see if others can reproduce
|-
| colspan="3" |  phuang_: ok
|-
| colspan="3" |  Kaio: how's the tables packaging?
|-
| colspan="3" |  dyek2 dychen
|-
| colspan="3" |  dychen: anything new on im from you?
|-
| style="background-color: #4d4d93" | Kaio
| style="color: #4d4d93" colspan="2" | juhp☺ mock failed, suspect yum repo is broken which caused that
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | [[I18N/InputMethods/Bugs]]
|-
| colspan="3" |  Kaio: koji scratch too?
|-
| style="background-color: #4d4d93" | Kaio
| style="color: #4d4d93" colspan="2" | juhp☺ also, improving ibus-table on my own tree, no reply from acevery about merging yet
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | ok
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | juhp: Processing  a couple of bugs of ibus-chewing.
|-
| style="background-color: #4d4d93" | Kaio
| style="color: #4d4d93" colspan="2" | juhp☺ I dont know how to do koji scratch w/o commiting .spec file and tag
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | BTW, I implemented compose key support in ibus-qt, anybody use compose keys. please help to test it
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | Kaio: koji build --scratch dist-f12 my.src.rpm
|-
| colspan="3" |  (of course preferred to do local test builds when possible)
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | juhp: bug 519108
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | .bug 519108
|-
| style="background-color: #8c4a4a" | zodbot
| style="color: #8c4a4a" colspan="2" | juhp: Bug 519108 [gob2] class and enum names convert incorrectly in mock / koji. - https://bugzilla.redhat.com/show_bug.cgi?id=519108
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | ah that?
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | which you have no problem building locally, but the mock fedora-devel-x86_64 fail.
|-
| colspan="3" |  No problem build in rawhide itself.
|-
| style="background-color: #4d4d93" | Kaio
| style="color: #4d4d93" colspan="2" | juhp☺ thx try it now
|-
| colspan="3" |  dyek2 dychen
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | Kaio: Not sure whether these are related to your bug.
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | dychen: sure
|-
| colspan="3" |  dyek2 dychen
|-
| colspan="3" |  dychen: have you reproduced on another instance?
|-
| colspan="3" |  looks a bit suspicious at first hand
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | juhp: The src rpm I attached are a minimal package that trigger the bug.
|-
| colspan="3" |  dyek2 dychen
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | dychen: so it happens in koji?
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | juhp: Yes.
|-
| colspan="3" |  dyek2 dychen
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | dychen: ok - don't see any attachment btw
|-
| colspan="3" |  dyek2 dychen
|-
| colspan="3" |  dychen: and please add a link to the koji build....
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | juhp: How long will log files of a koji scratch build last?
|-
| colspan="3" |  dyek2 dychen
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | dychen: good question
|-
| colspan="3" |  a week + ?
|-
| colspan="3" |  dyek2 dychen
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | Not quite sure how active  is gob2 maintainer.
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | dychen: true
|-
| colspan="3" |  dyek2 dychen
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | juhp: Anyway I've attach the attachment.
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | dychen: I think it is good to point to a explicit build failure
|-
| colspan="3" |  .bug 517001
|-
| style="background-color: #8c4a4a" | zodbot
| style="color: #8c4a4a" colspan="2" | juhp: Bug 517001 dlopen/dlclose of im-scim.so causes segfault - https://bugzilla.redhat.com/show_bug.cgi?id=517001
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | particularly https://bugzilla.redhat.com/show_bug.cgi?id=517001#c2
|-
| colspan="3" |  hmm
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | I think scim library need do many clean up before unload. :(
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | hmm
|-
| colspan="3" |  I see
|-
| colspan="3" |  so probably we should push it upstream
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | But why the old version gtk2 does not have this problem?
|-
| colspan="3" |  I think the gtk2 in f11 does not dlclose the im-scim.so
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | ah
|-
| colspan="3" |  I see
|-
| colspan="3" |  maybe
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | hmm, is it?
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | maybe it does?
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | at least I don't see any problems with this sample code on F-11 though
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | ok
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | or, stdc++ will do something in dlopen
|-
| colspan="3" |  I mean the new stdc++
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | tagoh3: so are you in agreement?
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | in rawhide
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | aren't
|-
| colspan="3" |  hmm
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | I'm wondering if this issue is im-scim.so specific or not.
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | ah
|-
| colspan="3" |  dunno
|-
| colspan="3" |  guess not many people use it anyway
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | just opening and closing doesn't process any code in the module though.
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | I will suggest Tasaka-san to push it upstream
|-
| style="background-color: #818144" | phuang_
| style="color: #818144" colspan="2" | tagoh3, maybe has some global c++ objects will be created in dlopen
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | unless im-scim.so doesn't do any wrong on constructors and destructors. but it's funny anyway it didn't happen on F-11 but on rawhide only.
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | http://sourceforge.net/project/stats/detail.php?group_id=108454&ugn=scim&type=svn&mode=12months
|-
| colspan="3" |  okay
|-
| colspan="3" |  so seems unrelated to scim -> ibus imsettings issue then
|-
| colspan="3" |  thought it might have been related perhaps
|-
| colspan="3" |  but anyway that is scim-bridge
|-
| colspan="3" |  so shrug
|-
| colspan="3" |  if you guys want to add some technical comments might help :)
|-
| colspan="3" |  dyek2 dychen
|-
| colspan="3" |  dychen: bug 518901 looks interesting
|-
| colspan="3" |  .bug 518901
|-
| style="background-color: #8c4a4a" | zodbot
| style="color: #8c4a4a" colspan="2" | juhp: Bug 518901 ibus-chewing would not work with locale zh_TW.Big5 - https://bugzilla.redhat.com/show_bug.cgi?id=518901
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | anyway 1 hour is up so let's stop in a moment
|-
| colspan="3" |  anything else to discuss?
|-
| style="background-color: #854685" | tagoh3
| style="color: #854685" colspan="2" | nope
|-
| style="background-color: #4d4d93" | Kaio
| style="color: #4d4d93" colspan="2" | no
|-
| style="background-color: #4b904b" | dychen
| style="color: #4b904b" colspan="2" | nope
|-
| style="background-color: #407a40" | juhp
| style="color: #407a40" colspan="2" | thanks for the meeting then!
|}
Generated by irclog2html.py 2.7 by [mailto:marius@pov.lt Marius Gedminas] - find it at [http://mg.pov.lt/irclog2html mg.pov.lt]!

Latest revision as of 21:06, 17 September 2016

previous meeting | next meeting

Agenda

Time:2009-08-26 0500 UTC

Place: #fedora-i18n at Freenode

Summary

Log

juhp IM meeting soon
who's in for the IM meeting?* pravins in
Kaio in
paragan in
juhp wonder if phuang_, tagoh3 and fujiwarat are around
dyek2 dychen
and dychen and Kaio
fujiwarat hi
phuang_ yeah
juhp ah hi Kaio
phuang_ hi
juhp cool
tagoh3 hi
juhp quick update on the @input-methods:
so far clumens kind of pushed back saying most people won't understand what "Input Methods" are anyway... (for anaconda top package listing)
anyway we followed up - let's see how it goes
otherwise we may just leave it to post-f12 even
I would rather avoid all the "how do i install IM?" and "why am I using <xim>?" type faq's but anyway
(if we turn off @input-methods)
pravins juhp: yeah, that looks nice
juhp .bug 518970
zodbot juhp: Bug 518970 ibus-daemon persists after desktop session - https://bugzilla.redhat.com/show_bug.cgi?id=518970
juhp phuang_: any comment?
hmm guess ibus-daemon can run in console too?
tagoh3 even with --xim?
juhp hmm
dunno :)
phuang_ I did not meet this problem. I will test it later
juhp ok please
I can reproduce 100%
anyone else seen it?
tagoh3 was about to do it - no lucks on rawhide installation yet, again ;o)
juhp ah
anaconda?
I have been getting funny Live spin errors too this week :/
tagoh3 dunno. it freezes on creating fs on virtio.
juhp I am running an up-to-date pre-alpha live install here
tagoh3 guess better trying 12alpha
juhp yeah
that might install ;)
.bug 517603
zodbot juhp: Bug 517603 anaconda failed in get_table - https://bugzilla.redhat.com/show_bug.cgi?id=517603
juhp anyway
phuang_: ibus-qt works well for me - well done!
is it hard to get underline in preedit?
phuang_ juhp, Just tested ibus in rawhide, this problem does not happen
dychen juhp: Hi
juhp phuang_: recent install?
phuang_: you logged in twice?
phuang_ juhp, yeah
juhp hmm
phuang_ let me try it again
dychen juhp: ibus can run in console. :-)
juhp phuang_: still happens for me anyway
hmm perhaps I should try on a fresh install... if only...
phuang_: when is your tree do you know?
dyek2 dychen
dychen: right
phuang_ juhp, I am not sure. But I often execute 'yum update'
juhp well maybe let's wait to see if others can reproduce
phuang_: ok
Kaio: how's the tables packaging?
dyek2 dychen
dychen: anything new on im from you?
Kaio juhp☺ mock failed, suspect yum repo is broken which caused that
juhp I18N/InputMethods/Bugs
Kaio: koji scratch too?
Kaio juhp☺ also, improving ibus-table on my own tree, no reply from acevery about merging yet
juhp ok
dychen juhp: Processing a couple of bugs of ibus-chewing.
Kaio juhp☺ I dont know how to do koji scratch w/o commiting .spec file and tag
phuang_ BTW, I implemented compose key support in ibus-qt, anybody use compose keys. please help to test it
juhp Kaio: koji build --scratch dist-f12 my.src.rpm
(of course preferred to do local test builds when possible)
dychen juhp: bug 519108
juhp .bug 519108
zodbot juhp: Bug 519108 [gob2] class and enum names convert incorrectly in mock / koji. - https://bugzilla.redhat.com/show_bug.cgi?id=519108
juhp ah that?
dychen which you have no problem building locally, but the mock fedora-devel-x86_64 fail.
No problem build in rawhide itself.
Kaio juhp☺ thx try it now
dyek2 dychen
dychen Kaio: Not sure whether these are related to your bug.
juhp dychen: sure
dyek2 dychen
dychen: have you reproduced on another instance?
looks a bit suspicious at first hand
dychen juhp: The src rpm I attached are a minimal package that trigger the bug.
dyek2 dychen
juhp dychen: so it happens in koji?
dychen juhp: Yes.
dyek2 dychen
juhp dychen: ok - don't see any attachment btw
dyek2 dychen
dychen: and please add a link to the koji build....
dychen juhp: How long will log files of a koji scratch build last?
dyek2 dychen
juhp dychen: good question
a week + ?
dyek2 dychen
dychen Not quite sure how active is gob2 maintainer.
juhp dychen: true
dyek2 dychen
dychen juhp: Anyway I've attach the attachment.
juhp dychen: I think it is good to point to a explicit build failure
.bug 517001
zodbot juhp: Bug 517001 dlopen/dlclose of im-scim.so causes segfault - https://bugzilla.redhat.com/show_bug.cgi?id=517001
juhp particularly https://bugzilla.redhat.com/show_bug.cgi?id=517001#c2
hmm
phuang_ I think scim library need do many clean up before unload. :(
juhp hmm
I see
so probably we should push it upstream
phuang_ But why the old version gtk2 does not have this problem?
I think the gtk2 in f11 does not dlclose the im-scim.so
juhp ah
I see
maybe
tagoh3 hmm, is it?
juhp maybe it does?
tagoh3 at least I don't see any problems with this sample code on F-11 though
juhp ok
phuang_ or, stdc++ will do something in dlopen
I mean the new stdc++
juhp tagoh3: so are you in agreement?
phuang_ in rawhide
juhp aren't
hmm
tagoh3 I'm wondering if this issue is im-scim.so specific or not.
juhp ah
dunno
guess not many people use it anyway
tagoh3 just opening and closing doesn't process any code in the module though.
juhp I will suggest Tasaka-san to push it upstream
phuang_ tagoh3, maybe has some global c++ objects will be created in dlopen
tagoh3 unless im-scim.so doesn't do any wrong on constructors and destructors. but it's funny anyway it didn't happen on F-11 but on rawhide only.
juhp http://sourceforge.net/project/stats/detail.php?group_id=108454&ugn=scim&type=svn&mode=12months
okay
so seems unrelated to scim -> ibus imsettings issue then
thought it might have been related perhaps
but anyway that is scim-bridge
so shrug
if you guys want to add some technical comments might help :)
dyek2 dychen
dychen: bug 518901 looks interesting
.bug 518901
zodbot juhp: Bug 518901 ibus-chewing would not work with locale zh_TW.Big5 - https://bugzilla.redhat.com/show_bug.cgi?id=518901
juhp anyway 1 hour is up so let's stop in a moment
anything else to discuss?
tagoh3 nope
Kaio no
dychen nope
juhp thanks for the meeting then!

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!

Retrieved from "https://fedoraproject.org/w/index.php?title=I18N/InputMethods/Meetings/2009-08-26&oldid=471538"