From Fedora Project Wiki
2009-07-08 0500 UTC on #fedora-i18n at Freenode
Agenda
previous meeting I18N/InputMethods/Meetings/2009-07-01
- keyboard layouts
- next meeting I18N/InputMethods/Meetings/2009-07-15
Summary
discussion about consolidating ibus handling of kbd layouts in IMEs
Log
juhp | hello | Jul 08 15:03 |
---|---|---|
juhp | ping phuang, tagoh3, dingyichen, pravins, et al | Jul 08 15:04 |
dingyichen | pong juhp | Jul 08 15:04 |
pravins | hi all | Jul 08 15:04 |
phuang | hi | Jul 08 15:04 |
juhp | hmm did we talk about using #fedora-meeting? for me here is fine but we have the meeting bot there and more visibility ;o) | Jul 08 15:04 |
juhp | maybe here is ok but if people we should move that is cool too :) | Jul 08 15:05 |
dingyichen | I don't mind either. | Jul 08 15:06 |
pravins | same here | Jul 08 15:06 |
<--yshao has quit ("what a day!") | Jul 08 15:06 | |
-->yshao (n=yshao@60.247.97.97) has joined #fedora-i18n | Jul 08 15:06 | |
juhp | phuang: how are you? :) | Jul 08 15:08 |
phuang | juhp, fine. | Jul 08 15:08 |
phuang | juhp, Last week, Shao and I participated a conference about Chinese Standards for China Mainland and Taiwan | Jul 08 15:08 |
juhp | right | Jul 08 15:09 |
phuang | juhp, In the conference, we introduced ibus to people from Mainland and Taiwan. They are interested | Jul 08 15:09 |
juhp | cool | Jul 08 15:10 |
phuang | juhp, Taiwanese also want to help us to improve ibus for Taiwan users | Jul 08 15:10 |
juhp | great! | Jul 08 15:10 |
juhp | more IMEs? :) | Jul 08 15:10 |
dingyichen | Good to hear that. | Jul 08 15:10 |
phuang | juhp, more test for ibus, | Jul 08 15:11 |
juhp | ok great | Jul 08 15:11 |
tagoh3 | hi | Jul 08 15:11 |
phuang | and introduce ibus to more people | Jul 08 15:11 |
juhp | aha | Jul 08 15:11 |
juhp | nice | Jul 08 15:11 |
phuang | They said they are maintaining some tables (like cangjie). Maybe we could use them. | Jul 08 15:12 |
juhp | ok | Jul 08 15:13 |
phuang | Except it, I also fixed some ibus bugs, and continue developing qt im module | Jul 08 15:13 |
juhp | :) | Jul 08 15:13 |
phuang | that's all from me | Jul 08 15:13 |
juhp | phuang: been thinking a bit more about kbd layouts | Jul 08 15:14 |
juhp | I think it needs to be per mode not IME - so maybe IME's should do that not ibus | Jul 08 15:15 |
phuang | no | Jul 08 15:15 |
phuang | juhp, you mean let IME handle the keycode? | Jul 08 15:16 |
juhp | phuang: right | Jul 08 15:16 |
dingyichen | The current ibus-1.2 allow IME developers to do that. | Jul 08 15:17 |
juhp | phuang: eg chewing apparently does something like that in its English submode | Jul 08 15:17 |
juhp | and anthy should only do it for kana input mode | Jul 08 15:17 |
juhp | tagoh3: wondering though is kana input usable on non-jp kbd? | Jul 08 15:17 |
juhp | (at least by default) | Jul 08 15:18 |
dingyichen | juhp: Not yet. ^.^b. | Jul 08 15:18 |
tagoh3 | juhp: dunno - I could do because I haven't looked at any printing at the key top :) | Jul 08 15:18 |
juhp | dingyichen: thought you showed me some code related to that? | Jul 08 15:18 |
<--noriko has quit (Read error: 110 (Connection timed out)) | Jul 08 15:18 | |
juhp | tagoh3: heh - I mean in terms of missing keys | Jul 08 15:19 |
dingyichen | juhp: That's for how to use IBusKeymap. :-) | Jul 08 15:19 |
juhp | dingyichen: ah | Jul 08 15:19 |
-->Kaio_Strassus (n=Kaio@fedora/Kaio) has joined #fedora-i18n | Jul 08 15:19 | |
juhp | dingyichen: but that is what we want no? | Jul 08 15:19 |
juhp | for anthy anyway | Jul 08 15:19 |
juhp | and for normal chewing mode I guess | Jul 08 15:20 |
juhp | but not the English submode.... | Jul 08 15:20 |
tagoh3 | juhp: it may depends how much keys are missing except the extra keys. | Jul 08 15:20 |
juhp | phuang: so actually IMEs are already doing it maybe? I need to look at anthy | Jul 08 15:21 |
tagoh3 | like win, menu keys etc | Jul 08 15:21 |
juhp | tagoh3: ok I was more thinking about kana keys like the extra keys on the right side of the kbd | Jul 08 15:21 |
phuang | juhp, ibus will send keysym and keycode to engines, engines can handle keysym or keycode byself. | Jul 08 15:21 |
juhp | phuang: ok cool | Jul 08 15:22 |
juhp | so probably ibus is fine | Jul 08 15:22 |
tagoh3 | juhp: symbol keys you mean? true. is it so different among keyboards? | Jul 08 15:22 |
juhp | phuang: but I am not sure the ibus-setup config is really necessary? | Jul 08 15:22 |
juhp | tagoh3: no - jp106 has more keys than us | Jul 08 15:22 |
juhp | 2 more I believe | Jul 08 15:23 |
juhp | + the others of course | Jul 08 15:23 |
phuang | juhp, it fixes some bugs, 'ibus can not works with no-us' keyboard | Jul 08 15:23 |
juhp | tagoh3: next to shift and Enter I think | Jul 08 15:23 |
tagoh3 | juhp: hmm, just missing 4 keys, no? | Jul 08 15:23 |
tagoh3 | erm, 5 keys | Jul 08 15:24 |
juhp | tagoh3: nope | Jul 08 15:24 |
phuang | tagoh3, and some keys' position are different | Jul 08 15:24 |
juhp | that too | Jul 08 15:24 |
tagoh3 | phuang: right | Jul 08 15:24 |
juhp | tagoh3: I am talking about punctuation | Jul 08 15:24 |
juhp | anyway | Jul 08 15:25 |
tagoh3 | different position may be more serious. it's important to kana input | Jul 08 15:25 |
juhp | I think kaio said scim-anthy maps "ro" to ` | Jul 08 15:25 |
juhp | for example | Jul 08 15:25 |
tagoh3 | aha | Jul 08 15:26 |
juhp | tagoh3: yep that can be fixed by layout but not missing keys :) | Jul 08 15:26 |
juhp | anyway | Jul 08 15:27 |
juhp | just wondered... | Jul 08 15:27 |
tagoh3 | well, probably Kanji, Mukenkan, Henkan, Katakana_Hiragana keys - interesting which one more key are missing | Jul 08 15:27 |
Kaio | phuang☺ could I ask who in the conference said they are maintaining some tables like cangjie? | Jul 08 15:27 |
juhp | and guess not many people use kana input without jp kbd anyway... | Jul 08 15:27 |
phuang | Kaio, wait | Jul 08 15:28 |
Kaio | phuang☺ ok | Jul 08 15:28 |
juhp | tagoh3: http://en.wikipedia.org/wiki/Keyboard_layout - I think US kbd model is missing equivalent to ] and \ on jp kbd fwiw (is what I meant:) | Jul 08 15:29 |
phuang | Kaio, I think you could ask Lieh-neng Wu, lnwu@mail.tca.org.tw | Jul 08 15:29 |
juhp | anyway | Jul 08 15:30 |
juhp | phuang: well I am saying/wondering if those bugs should be fixed in the IMEs solely | Jul 08 15:31 |
Kaio | if they release as GPL or so, then that's cool :) | Jul 08 15:31 |
-->noriko (n=noriko@nat/redhat/x-f0bfd4524bf6cd7d) has joined #fedora-i18n | Jul 08 15:31 | |
juhp | if so better not to have the config option if possible IMHO | Jul 08 15:31 |
juhp | phuang: anyway as noted in bz anthy should not default to jp layout | Jul 08 15:31 |
phuang | Kaio, It is free. I asked | Jul 08 15:31 |
tagoh3 | juhp: is it US101 layout keyboard? | Jul 08 15:31 |
Kaio | phuang☺ thx | Jul 08 15:32 |
phuang | juhp, if ibus does not handle it, all ime should handle it by self. | Jul 08 15:32 |
juhp | phuang: yup | Jul 08 15:32 |
juhp | phuang: in general ibus can't know when to do layout conversion only IME can know that | Jul 08 15:33 |
tagoh3 | juhp: did they use 98th keys before windows launched? I see two win keys and menu key there though.. | Jul 08 15:33 |
phuang | juhp, I think ime don't know it too | Jul 08 15:33 |
phuang | just users know it | Jul 08 15:33 |
juhp | tagoh3: yeah I am not worrying about the bottom row but anyway :) | Jul 08 15:34 |
juhp | of course we know US does not have conversion keys | Jul 08 15:34 |
tagoh3 | hmm | Jul 08 15:34 |
juhp | phuang: so who knows ;-) | Jul 08 15:35 |
juhp | phuang: seriously the IME has to know | Jul 08 15:35 |
juhp | hmm but I see what you mean | Jul 08 15:35 |
juhp | if you use the "correct" kbd no conversion needed... hmm | Jul 08 15:35 |
<--Kaio_Acer has quit (Read error: 110 (Connection timed out)) | Jul 08 15:35 | |
juhp | perhaps I need to think more about it | Jul 08 15:36 |
dingyichen | juhp: Does our l10n team have a spare Jpn keyboard? | Jul 08 15:36 |
juhp | dingyichen: maybe or possibly QE | Jul 08 15:36 |
juhp | phuang: but I think IME can know, no? | Jul 08 15:37 |
phuang | juhp, how do IME know it? Using IME's configure UI? | Jul 08 15:37 |
juhp | if I am using azerty to input chewing - chewing can convert to qwerty no? | Jul 08 15:37 |
juhp | phuang: that might be way? | Jul 08 15:38 |
juhp | phuang: or what would break if chewing say always use scan codes for input? | Jul 08 15:38 |
juhp | well something may break... ? | Jul 08 15:38 |
juhp | dingyichen: ? | Jul 08 15:39 |
dingyichen | juhp, I think there is at least German keyboard around, maybe I should test with it. | Jul 08 15:39 |
dingyichen | juhp, currently ibus-chewing force en-us keymap. | Jul 08 15:39 |
phuang | juhp, dingyichen, chewing is OK with keycode | Jul 08 15:40 |
juhp | dingyichen: ok - so everywhere, right? | Jul 08 15:40 |
juhp | phuang: so any examples of cases that would not work? :) | Jul 08 15:40 |
phuang | juhp, but in chewing, it must convert keycode to keysym too | Jul 08 15:40 |
dingyichen | juhp: I've test with querty and dvorak, seems work for me, | Jul 08 15:40 |
juhp | phuang: yep | Jul 08 15:40 |
juhp | dingyichen: so actually chewing currently ignores the ibus setting? | Jul 08 15:41 |
phuang | juhp, If some cases don't work, I think chewing will not work with xkb too | Jul 08 15:41 |
dingyichen | juhp: Exactly. :-) | Jul 08 15:41 |
juhp | phuang: why not? | Jul 08 15:41 |
phuang | converting in ibus, ime or xkb. they are almost same | Jul 08 15:42 |
juhp | dingyichen: but english mode should use keysyms I guess | Jul 08 15:42 |
dingyichen | juhp: right. | Jul 08 15:42 |
juhp | phuang: I think biggest problem is m17n | Jul 08 15:42 |
phuang | juhp, why? | Jul 08 15:43 |
juhp | some maps need qwerty others not | Jul 08 15:43 |
juhp | so we need some metadata for that or maybe hardcoding in the meantime | Jul 08 15:43 |
phuang | juhp, how do we know which need qwerty or qwertz, or others? | Jul 08 15:44 |
juhp | phuang: I can't think yet of any cases that need the config in ibus-setup but we'll see I guess | Jul 08 15:44 |
juhp | phuang: from the map layout whether it is layout based or phonetic :) | Jul 08 15:44 |
juhp | there is no automatic way of course | Jul 08 15:44 |
phuang | juhp, it is defined in map files? | Jul 08 15:44 |
juhp | not yet :) | Jul 08 15:45 |
juhp | phuang: eg Indic inscript maps (the default for Indic) are all layout based | Jul 08 15:45 |
juhp | as in assume qwerty in the maps | Jul 08 15:45 |
juhp | probably more than qwerty (ie us) | Jul 08 15:45 |
juhp | parag is not here... | Jul 08 15:46 |
phuang | But in the map, it handles the keysym | Jul 08 15:46 |
phuang | not keycdoe | Jul 08 15:46 |
phuang | keycode | Jul 08 15:46 |
juhp | if everything just works for the user that would be great :-) | Jul 08 15:46 |
juhp | phuang: indeed | Jul 08 15:46 |
juhp | that is the problem | Jul 08 15:46 |
juhp | so those maps are really "wrong" | Jul 08 15:46 |
juhp | so another way might be to fix the maps... | Jul 08 15:47 |
juhp | or use xkb ;) | Jul 08 15:47 |
juhp | hmm | Jul 08 15:47 |
phuang | no | Jul 08 15:47 |
juhp | ? | Jul 08 15:47 |
phuang | juhp, we just need define layout in map files | Jul 08 15:47 |
phuang | And ibus will convert keycode for it | Jul 08 15:47 |
juhp | but there is the thought we should really use xkb for trivial maps like inscript | Jul 08 15:48 |
juhp | and let ibus handle the layout switching | Jul 08 15:48 |
juhp | is it too radical? | Jul 08 15:49 |
juhp | phuang: defining layout + map sounds like an xkb map though :) | Jul 08 15:49 |
phuang | Can it work with all xkb layouts? | Jul 08 15:50 |
juhp | phuang: I dunno - all is lot :) | Jul 08 15:50 |
phuang | juhp, I mean just need define layout=us or layout=jp in maps files | Jul 08 15:50 |
juhp | we would probably have to define a subset | Jul 08 15:50 |
phuang | juhp, I think it can not work with kana xkb layout | Jul 08 15:51 |
juhp | phuang: what about no layout? :) | Jul 08 15:51 |
juhp | so layout would optional? | Jul 08 15:51 |
phuang | layout="" or layout=xkb | Jul 08 15:51 |
juhp | ok | Jul 08 15:51 |
phuang | ibus will use system keysym | Jul 08 15:51 |
juhp | yeah that makes sense I think | Jul 08 15:51 |
phuang | not convert for it | Jul 08 15:51 |
juhp | it might take time to get it upstream though | Jul 08 15:52 |
juhp | so we could hardcode it first in fedora perhaps | Jul 08 15:52 |
phuang | yeah | Jul 08 15:52 |
phuang | if not define the layout, we don't have good way to resolve this problem | Jul 08 15:53 |
-->krishnababu (n=kkrothap@nat/redhat-in/x-58543cc070766e1d) has joined #fedora-i18n | Jul 08 15:53 | |
juhp | sure | Jul 08 15:54 |
juhp | and after that I am hoping we don't need the global layout config in setup? | Jul 08 15:54 |
juhp | just worry a little about synergy and virt and stuff, but hopefully no worse than now? | Jul 08 15:55 |
juhp | ok anyway running out of time.... | Jul 08 15:56 |
juhp | anything else to discuss today? :) | Jul 08 15:56 |
juhp | from anyone | Jul 08 15:56 |
dingyichen | juhp: Yes, synergy might post problem. | Jul 08 15:56 |
dingyichen | But we might be about to come out with workaround anyway. | Jul 08 15:57 |
juhp | dingyichen: yeah but don't see any solution to remote kbd problem | Jul 08 15:57 |
juhp | dingyichen: ok - maybe we can discuss that later :) | Jul 08 15:57 |
juhp | ok let's finish here then | Jul 08 15:59 |
juhp | thanks for the meeting! | Jul 08 15:59 |
phuang | juhp, btw, currently, i am working on several tasks for ibus (kbd layout for ibus, qt im module, and fix bugs), | Jul 08 15:59 |
juhp | phuang: aha | Jul 08 15:59 |
phuang | anyone want to help me? | Jul 08 16:00 |
-->ifelix (n=felix@nat/redhat-in/x-887b2d97d8d5c5a8) has joined #fedora-i18n | Jul 08 16:00 | |
juhp | phuang: sounds good | Jul 08 16:00 |
juhp | phuang: any particular bugs we should reassign? | Jul 08 16:00 |
phuang | juhp, I planed to assign ibus-qt to an intern. I don't know when the interns can on board. | Jul 08 16:01 |
juhp | ok | Jul 08 16:02 |
juhp | phuang: I think qt is important for ibus | Jul 08 16:02 |
phuang | yeah | Jul 08 16:03 |
juhp | maybe dingyichen can help with layout stuff | Jul 08 16:03 |
phuang | dingyichen, do you have time? | Jul 08 16:03 |
juhp | phuang: or we can go through the buglist together if you like | Jul 08 16:03 |
juhp | (or both:) | Jul 08 16:03 |
dingyichen | juhp: phuang: Agree, I think in brisbane we have some non-US keyboard to test on. | Jul 08 16:04 |
juhp | I need to work on yum plugin but when that is finished hope to have a little time for ibus too (but may be optimistic this month) | Jul 08 16:04 |
phuang | dingyichen, juhp, thanks | Jul 08 16:04 |
juhp | phuang: anyway please feel free to assign bugs to me if you need assistance and I can try to find resources to work on some of them | Jul 08 16:05 |
dingyichen | juhp: I think we do need someone to take care of ibus-anthy | Jul 08 16:05 |
juhp | yep | Jul 08 16:05 |
dingyichen | juhp: You are very busy, apparently. | Jul 08 16:06 |
juhp | hope also kaio and others can lend a hand too | Jul 08 16:06 |
juhp | trying to encourage people to take more bugs proactively :) | Jul 08 16:06 |
phuang | juhp, ok | Jul 08 16:08 |
Kaio | sure | Jul 08 16:08 |
juhp | phuang: we 3 sometimes look at the bug list together on Monday morning | Jul 08 16:08 |
phuang | cool | Jul 08 16:08 |
juhp | phuang: let's have a chat about it later anyway | Jul 08 16:08 |
juhp | phuang: yeah cooler if we actually take some bugs ;) | Jul 08 16:09 |
juhp | ok I will include these comments in the log as "extra time" :) | Jul 08 16:09 |
phuang | thanks | Jul 08 16:10 |
juhp | phuang: also good to file bugs for such issues if they are not there so easier to track :) | Jul 08 16:11 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!