From Fedora Project Wiki
I18N Project Meeting 2009-01-20 Minutes
- 2009-01-20 0500 UTC on #fedora-meeting at Freenode
Agenda
Please add agenda items or propose to fedora-i18n-list ahead of the meeting.
- Features/IBus -- JensPetersen
- Yum langpacks extension -- JensPetersen
- Extending input-methods with emoticons and other unicode chars -- PravinSatpute
- default vfat fs encoding in F11 kernel (should the utf8 patch be removed? and be sure it's fixed in [hal or devicekit rhbz #470592]) -- User:Alsadi
- Fonts package renaming
- Language tagging in toolkits of input text for correct rendering and spelling-checking -- nim-nim
- Next meeting
Log
15:00:01 <juhp> https://fedoraproject.org/wiki/I18N/Meetings
15:00:11 <juhp> okay I make time now
15:00:33 <juhp> so welcome to the first fedora-i18n meeting in a while
15:00:51 <pravins> hi all
15:00:52 <juhp> hopefully will be become more regular this year :)
15:01:08 <sflanigan> g'day
15:01:20 <phuang> hi
15:01:24 <juhp> hi guys
15:01:27 <paragn> hi
15:01:34 <juhp> who else is participating in the meeting?
15:01:40 <jamesni> hi
15:01:45 <tagoh3> hi
15:02:06 <Kaio> hi
15:02:10 <juhp> we have an agenda for the meeting https://fedoraproject.org/wiki/I18N/Meetings#Agenda
15:02:24 <juhp> let's get started
15:02:39 <juhp> * IBus feature
15:02:50 <dychen_> hi
15:02:58 <juhp> first topic is ibus feature for f11
15:03:19 <juhp> phuang is pushing ahead with ibus for f11
15:03:58 --> yshao (n=yshao@60.247.97.98) has joined #fedora-meeting
15:03:59 <juhp> yesterday we switched comps to ibus so that tomorrow's rawhide should default now to ibus rather than scim
15:04:29 <phuang> juhp, thanks
15:04:52 <juhp> https://fedoraproject.org/wiki/Features/IBus has been moved to ready for fesco - so will be good to get feedback from them on the plan
15:05:08 <juhp> :)
15:05:25 <-- bpepple has quit ("Ex-Chat")
15:06:17 <juhp> there is still quite a lot of work to do to complete the feature work in time for beta
15:06:40 * asgeirf attempts to sneak in a slightly delayed 'hi!'
15:06:42 --> rahul_b (n=rbhalera@123.236.183.101) has joined #fedora-meeting
15:07:19 <juhp> phuang: it is very important that the feature is completed in good time for the beta if it is going to be in f11:)
15:08:00 <juhp> I think dychen_ and Kaio will be doing some work on the engines
15:08:20 <phuang> juhp, Yeah. I will do my best to achieve it
15:08:30 <juhp> phuang: btw I was curious how easy do you think it is to write an engine in a language other than C or python?
15:08:35 <juhp> phuang: thanks
15:08:37 <dychen_> Yes, current I am workin on the building problem between libChewing and ibus-python.
15:08:45 <Kaio> juhp☺ mainly ibus-table, and ibus-anthy
15:08:45 <juhp> dychen_: great
15:08:57 <juhp> Kaio: sounds good
15:09:32 <juhp> if you have comments on the plan now or on the talk page would be a good time/place to do that too
15:09:53 <dychen_> juhp, as soon as my libUnihan-iBus interface ready, they can use SQL table as engine. :-)
15:09:53 --> sdziallas (n=sebastia@p57A2C2C6.dip.t-dialin.net) has joined #fedora-meeting
15:09:57 <phuang> juhp, It is not very hard. I could provide some libraries for other languages to make it easy.
15:10:14 <juhp> phuang: ok - np was just thinking about that
15:10:28 <juhp> in fact there are already a few comments there
15:10:33 <juhp> https://fedoraproject.org/wiki/Talk:Features/IBus
15:10:55 <phuang> juhp, Currently, developers can only use dbus wrappers directly in other languages.
15:11:04 <juhp> first one I was discussing with nim-nim and it is big and outside the scope of f11 so I moved it to another agenda item
15:11:44 <juhp> phuang: ok - so for c/python there is a nicer api, right?
15:12:22 <juhp> second item I answered on fedora-devel-list earlier
15:12:47 <juhp> phuang: any comment on third question - about remote X support?
15:13:00 <juhp> I imagine it would be similar situation to scim
15:13:03 <phuang> yeah. DBus is transparency in c & python
15:13:09 <juhp> ok
15:13:44 <juhp> any other discussion on this important topic before we move on?
15:14:25 <juhp> so aim is to have C implemention in rawhide before Beta so that it can be tested enough for final f11
15:14:30 <-- rahulb has quit (Read error: 110 (Connection timed out))
15:14:34 <phuang> Remote X session could have a remote ibus daemon, or use locale ibus via XIM
15:14:50 <juhp> okay
15:15:27 <juhp> phuang: can you follow? and I can take care of the other two
15:15:38 <pravins> juhp: when we suppose to develop scim-sinhala etc for ibus?
15:16:04 <juhp> pravins: good question - how important do you think it is?
15:16:28 <pravins> phuang: what you think?
15:16:32 <juhp> pravins: at least we have maps for Sinhala in m17n like Indic langs but of course in long term would like to see more IMEs
15:16:54 <juhp> currently it is not default for Sinhala and not installed by default in distro anyway fwiw
15:17:04 <juhp> the default, I mean
15:17:08 <pravins> aha
15:17:18 <phuang> juhp, yeah. We have the map in m17n.
15:17:19 <juhp> it was earlier I think
15:17:21 <tagoh3> juhp: is there any languages not supported by ibus right now?
15:17:37 <tagoh3> which was supported by scim
15:18:01 <juhp> tagoh3: I can think of any off hand particularly
15:18:29 <juhp> of course there are more IMEs available at this time still for scim
15:19:05 <juhp> then again m17n handles nearly all langs
15:19:23 <juhp> but I don't think we lose any default IMEs by moving to ibus
15:19:39 <juhp> for langs that use IM
15:20:02 <juhp> shall we move on?
15:20:11 <phuang> ok
15:20:31 <juhp> # Yum langpacks extension
15:21:02 <juhp> okay asgeirf pointed out to me the recent discussion on fedora-devel-list about comps
15:21:27 <juhp> and actually notting posted a prototype yum plugin that is close to this idea
15:22:16 <juhp> so probably I will see if that can be used to solve the long-standing issue with getting conditional packages like langpacks installed
15:22:18 --- knurd_afk is now known as knurd
15:23:41 <juhp> I was even pondering a feature for it, but glad to see that other people also see the importance of this and getting rid of the language support conditional nonsense eventually hopefully
15:24:13 <juhp> hope this can be improved for f11 anyway
15:24:36 <juhp> # Extending input-methods with emoticons and other unicode chars -- PravinSatpute
15:25:00 <juhp> I might call this submodes (or minor modes in the sense of emacs) for input methods
15:25:17 <pravins> :) -> ☺ U+263A
15:25:22 <pravins> :( -> ☹ U+2639
15:25:27 <juhp> yeah
15:25:59 <juhp> I would like to have a nice unicode input method in m17n or ibus more probably
15:26:17 <juhp> but mixing of IMEs is very interesting
15:26:20 <-- sdziallas has quit ("Ex-Chat")
15:26:28 <juhp> a bit like combining xkb maps I guess
15:26:45 <juhp> phuang et al: any thoughts on that
15:27:25 <phuang> juhp, It is good
15:27:30 <juhp> we probably won't do it for f11 say unless it is easy but interesting idea for the future
15:27:49 <phuang> juhp, yeah.
15:28:05 <pravins> yeah
15:28:07 <asgeirf> it would certainly get IMs more exposure :)
15:28:08 <juhp> it would certainly do something that other OSes can't
15:28:16 <juhp> asgeirf: right that too
15:28:20 <phuang> pravins, currently, ibus gtk im module, could do this. But qt im moudule & xim can not
15:28:37 <juhp> in fact I think we need to think about how to deal with xkb - a lot of people have talked about this
15:28:50 <juhp> the current dichotomy is quite ugly
15:29:25 <juhp> to us anyway - there should just be keyboard layouts and everything else should be an input method ;)
15:30:15 <phuang> juhp, I think we could let ibus to manage xkb. But I still need do more investigation on xkb.
15:30:20 <juhp> phuang: how does ibus-gtk deal with xkb issue again?
15:30:22 <juhp> okay
15:30:26 <juhp> that would be good
15:31:21 <juhp> pravins: in the mean time you could write an input map for emoticons, etc
15:31:28 <juhp> might catch on in irc... :)
15:31:41 <pravins> juhp: surely i will try for that
15:31:45 <juhp> great
15:31:51 <pravins> but :) and :( these are very common
15:32:02 <juhp> yeah
15:32:03 <pravins> and we can add these in existing mim files
15:32:16 <juhp> and things like arrows, and what not
15:32:34 <juhp> pravins: yeah but not everyone may want them that i the problem
15:32:40 <juhp> is
15:32:43 <pravins> hmm
15:32:46 <phuang> pravins, yeah. maybe we could create a common mim file for it. And let others include it. Is it possible?
15:33:13 <juhp> or some conditional configure way of combining m17n maps
15:33:19 <pravins> phuang: yeah, that is also nice
15:33:20 <juhp> configuration
15:33:43 <pravins> yes we can start something like that
15:33:57 <juhp> pravins: thanks for bringing it up
15:34:12 <pravins> juhp: welcome :)
15:34:38 <juhp> # default vfat fs encoding in F11 -- User:Alsadi
15:34:49 <juhp> default vfat fs encoding in F11 kernel (should the utf8 patch be removed? and be sure it's fixed in [hal or devicekit rhbz #470592])
15:35:06 <juhp> is Alsadi here?
15:35:17 <juhp> I think he wanted to bring attention to this issue
15:35:58 <juhp> it is really something more for the desktop team I guess but we can help to try to get a satisfactory solution for international users
15:37:13 <juhp> just to summarise: I believe a fedora kernel patch to change the default encoding for vfat mount to utf-8 was reverted since it changed the behaviour for the fedora kernel compared to upstream and caused problems
15:38:25 <juhp> I Alsadi is suggesting that the hal should nevertheless be made to mount with utf8 by default so that filenames in utf8 appear correctly
15:38:34 <juhp> filename encoding are fun...
15:38:53 <juhp> will try to keep an eye on it
15:39:21 <juhp> # Fonts package renaming
15:39:31 <juhp> this is in progress
15:39:42 <juhp> how are people here doing with it
15:40:04 <juhp> the latest version changes the naming convention from *-fonts-* to *-*-fonts
15:40:44 <juhp> rahul_b, pravins, Kaio, tagoh3, juhp all have fonts
15:41:00 <Kaio> juhp☺ yes
15:41:13 <juhp> I confess to not having updated thaifonts-scalable will do that this week
15:41:28 <tagoh3> not yet finished for {sazanami,hanazono,vlgothic}-fonts.
15:41:50 <juhp> tagoh3: so vlgothic is getting downcased too?
15:41:57 <Kaio> talk to nim-nim if need anyth
15:42:01 <rahul_b> I have lot to do with culmus and amy be even lohit fonts
15:42:03 <juhp> tagoh3: including the base package?
15:42:11 <pravins> not finished yet for samyak-fonts and smc-fonts
15:42:27 <juhp> rahul_b: ok better do lohit first I guess
15:42:29 <rahul_b> have been waiting for final guidelines for that
15:42:39 <juhp> good thinking
15:42:48 <juhp> they seem to be settling down now finally ;)
15:42:54 <tagoh3> juhp: no, not at all
15:43:12 <juhp> tagoh3: ah ok so still VLGothic-fonts?
15:43:19 <tagoh3> right
15:43:21 <juhp> good
15:43:36 <juhp> nim-nim was keen earlier to downcase ;-)
15:43:51 <juhp> maybe another say ;-P
15:44:18 <juhp> I hope by next meeting we will have finished the fonts stuff anyway :)
15:44:40 <juhp> # Language tagging in toolkits of input text for correct rendering and spelling-checking -- nim-nim
15:45:08 <juhp> mailhot put forward an interesting proposal which behdad also seems to buy
15:45:25 <juhp> first described on the ibus feature talk page
15:45:28 --> lfoppiano (n=lfoppian@fedora/lfoppiano) has joined #fedora-meeting
15:45:59 <juhp> but after discussing it earlier on #fedora-devel I think I understand it now too and agree that it sounds attracture
15:46:08 <juhp> erm attractive
15:46:24 <rahul_b> how does it work?
15:46:55 <rahul_b> from input method, or apps will handle on their own?
15:46:56 <juhp> basically the idea is that through the important method (probably) one specifies the language of input
15:47:08 <-- mether has quit (Read error: 113 (No route to host))
15:47:21 <juhp> rahul_b: well the exact details are still open under discussion
15:47:50 <pravins> rahul_b: might be adding language property field
15:47:53 --> mether (n=sundaram@122.169.28.77) has joined #fedora-meeting
15:48:02 <juhp> im will then communicate this to the app as a language hint which it can use to set the rendering language and which dictionary to use to spellcheck the input
15:48:37 <asgeirf> so this will in other words be like e.g. OpenOffice does, but on a desktop-level?
15:48:48 <juhp> so even if one is not running English desktop say one could still have english spellchecking easily in apps when one wants
15:48:57 <juhp> asgeirf: right
15:49:17 <juhp> well openoffice allows setting fonts I guess?
15:49:44 <tagoh3> asgeirf: at least that sounds like not something can be fixed in one.
15:50:00 <juhp> or importantly for CJK users can specific which Han language (script) they are using
15:50:30 <juhp> well doing it for gtk/gnome might be possible but it sounds complicated
15:50:48 <juhp> it needs to be discussed upstream for anything to happen
15:50:56 <juhp> mclasen was also listening
15:51:28 <asgeirf> KDE4 had a feature called 'sonnet' which was supposed to do language detection and stuff... not sure if it was ever included though...
15:51:43 <juhp> aha interesting
15:52:35 <juhp> ok
15:52:52 <juhp> any other last topics before we draw to a close?
15:53:06 <-- mether has quit (Read error: 104 (Connection reset by peer))
15:53:37 <juhp> I propose to have another meeting here in two weeks time - should be a lot of fedora 11 topics by then I guess after testing alpha :)
15:54:55 <juhp> okay if there are no more comments or topics then let's close the meeting
15:55:06 --- mbonnet is now known as mbonnet__
15:55:07 --> mether (n=sundaram@122.169.28.77) has joined #fedora-meeting
15:55:10 <juhp> thanks to everyone who joined in
15:55:28 <juhp> # (fedora-i18n meeting closed)