[00:00] Jazzva: otherwise just tar everything up and upload it somewhere [00:00] It's upgrading good so far [00:00] Jazzva: already pushing? [00:00] just upgrading ... it should finish soon, I suppose [00:00] Copying content into repository. 3/4 [00:14] i think i should write a book "learning by suffering" :) [00:14] You wouldn't be the only one ;) [00:14] i guess in 11 days i am going into a sanatorium ;) [00:15] hehe [00:15] mental institution ;) [00:15] Shouldn't you be attending the ubuntu-dev conference then :P? [00:16] and write another book "get strong in 1 month - how to recover in a 6 month release cycle" [00:16] lol [00:16] Jazzva: yeah ... in prague ;) [00:16] i think its 3 weeks after release [00:16] 15th march or something [00:16] Oh, good... You have the time for your "recovery" then :) [00:16] hehe ... yeah, i hope so [00:16] More like may ;)... [00:17] now I see another reason for shipping unarchived sources - no need to wait for that when you're searching for something :) [00:18] right ;) [00:18] thats the point [00:18] how many ship .xpi files now? [00:18] well, none so fart [00:18] *far [00:18] :) [00:18] i had the feeling during review that most were disciplined enough to extract and write a build.sh at least [00:19] Two of them are shipping with .jar in chrome, (all-in-one-sidebar and mozilla-bookmarskftp), and lightning source is completely tarred [00:19] yeah those are evil ;) [00:20] especially if you have to wait for unpacking :) [00:20] but we shouldn't raise the until we have acquired more contributors :) [00:20] but i think fta's script might be a good start that would allow us to autopackage and update our extensions [00:21] so you really just need a qa contact that deals with upstream questions [00:21] Yeah... that script is really handy :) [00:21] and reads bugs :) [00:21] we could create a bzr branch where the qa contact could bump/sign off just the version info of his extension and the script would automatically notice that an update is required [00:21] import that in .upstream ... merge over ... and auto commit if no conflicts arise [00:22] that would be nice :) [00:22] * asac dreams [00:22] :) [00:22] i think for extensions it makes sense to provide the latest crack as backports everywhere [00:22] so we definitly would need a way to automize as much as possible [00:23] Yep... that would make packaged extensions useful :). [00:23] otherwise maintaining like 100*4 things can become quite painful [00:23] you wouldn't have to wait for 6 months just to get a new version [00:23] once again, such a script is easy to write. I even already have it, it's my buildbot [00:23] exactly. i think the lack of latest crack is one of the main arguments why extensions are considered not-so-great-for packaging [00:24] fta: ;) [00:24] Yay :) [00:24] * Jazzva admires fta's scripting powers [00:24] fta: can we get those tools in a form that we can use them? [00:25] providing some cleanup, sure [00:26] !version mozgest [00:26] Sorry, I don't know anything about version mozgest - try searching on http://ubotu.ubuntu-nl.org/factoids.cgi [00:26] !package mozgest [00:26] Sorry, I don't know anything about package mozgest - try searching on http://ubotu.ubuntu-nl.org/factoids.cgi [00:26] pif... [00:26] !package mozgest hardy [00:26] found it with apt-cache :) [00:26] hehe [00:26] thanks [00:27] i never remember how this thing works [00:27] !package mozgest/hardy [00:27] nm [00:27] Jazzva: so did the push succeed? [00:27] fta: reconnect? [00:27] asac: still doing it ... [00:27] fta: can we get those tools in a form that we can use them? [00:27] providing some cleanup, sure [00:28] yeah we got that. [00:28] asac: it's slow [00:28] yep, #!?$ daily dsl reconnection [00:28] Jazzva: do you use bzr+ssh? and have pack-0.92 ? [00:29] yep and yep [00:29] http://arstechnica.com/articles/culture/reinventing-gtk.ars is worth the read [00:29] gtk 3? [00:29] asac, mozilla-mozgest should be dummy package depending on mozgest [00:29] yeah [00:29] Jazzva: why? [00:29] did we have mozilla-mozgest before? [00:30] mozilla-mozgest is an older version (1.5.2), while mozgest is 1.9.something [00:30] I suppose they are the same... [00:30] Jazzva: if so we also need versioned replaces: conflicts: [00:30] ok [00:30] Jazzva: but lets first get data up [00:30] then file a bug against mozgest which i make a blocker so we can justify upload [00:31] ok [00:31] fta: did we close the packages-static review bug already? [00:31] you said you did [00:31] i did. ok [00:31] i thought so. just wasn't sure [00:31] (anymore) [00:32] hmm [00:32] i don't know [00:32] bug 194342 [00:32] Launchpad bug 194342 in xulrunner-1.9 "review make install target and package-static files" [High,Confirmed] https://launchpad.net/bugs/194342 [00:32] still open [00:32] fta: let me look at official build if compare was run in b5 [00:32] i think if its in build log we can really close this [00:33] it is now for sure: http://launchpadlibrarian.net/13413793/buildlog_ubuntu-hardy-i386.xulrunner-1.9_1.9~cvs20080413t1131%2Bnobinonly-0ubuntu1~fta1_FULLYBUILT.txt.gz [00:34] fr [00:34] and http://launchpadlibrarian.net/13414046/buildlog_ubuntu-hardy-i386.firefox-3.0_3.0~cvs20080413t1131%2Bnobinonly-0ubuntu1~fta1_FULLYBUILT.txt.gz [00:34] fta: http://paste.ubuntu.com/6944/ thats xul output [00:34] can we sign that off? [00:34] the added files are definitly ok [00:34] everything else looks like chatzilla + windows [00:35] whats going on with chatzilla [00:35] ? [00:35] why is it build and then not included? [00:35] fta: ? [00:35] i want to document that in the bug when closing [00:36] i haven't looked closely enough [00:37] fta: but chatzilla is not ment to be included anywhere i guess, right? [00:37] yeah i think so [00:37] we get it from debian [00:37] fta: ? is it a standalone xul package there now? [00:37] !info chatzilla hardy [00:38] Package chatzilla does not exist in hardy [00:38] !info mozilla-chatzilla hardy [00:38] mozilla-chatzilla (source: seamonkey): dummy upgrade package for SeaMonkey Chatzilla IRC client. In component universe, is optional. Version 1.1.9+nobinonly-0ubuntu1 (hardy), package size 10 kB, installed size 84 kB [00:38] fta: ok. i close the xulrunner-1.9 bug then. [00:40] hm, it's from sm1. so the debian one is gone now. the one from xul1.9 would be fresher, or even from AMO as a regular ff extension [00:40] but it's too late now [00:40] yeah. [00:41] the xul application isn't really feature full [00:41] so i think its ok to have just the extension now [00:46] asac, what's the behavior of g-a-i if the desktop file contains an Icon field, but there is no icon? I wasn't able to find some icons, so I'm wondering should I set the Icon field to no icon. [00:46] assign a default one, the puzzle piece for example [00:46] hmm ... thanks, fta_ :) === fta_ is now known as fta [00:48] * fta leaving. cu [00:48] Hmm, g-a-i assigns package icon, if Icon field is set to none... [00:48] cu fta [00:49] Jazzva: thats ok [00:49] fta: bye [00:49] Yeah. But I setted a few Icon fields, but there's no icon... Wondering if it can handle that well too... [00:51] Jazzva: ;) ... no idea :). why not leave it empty? [00:51] Offtopic... I'm getting my first domain tomorrow/tuesday :). jazzva.net [00:51] Jazzva: you know how you can try? [00:51] Jazzva: congrats [00:51] copy the desktop file to /usr/somewhate :)? [00:51] Thanks :D... [00:52] *somewhere [00:52] and then to launch g-a-i? [00:54] hmm ... cant you build the branch and install the .deb? [00:54] then you can start gnome-app-install --xul-extensions=firefox [00:55] I wouldn't really play with it until it finishes pushing :)... [00:55] Jazzva: yeah ;) [00:55] Jazzva: you copy the .desktop files to /usr/share/app-install/desktop/ and the icons to /usr/share/app-install/icons [00:56] tried that ... doesn't seem to work :/ [00:56] do I need to reload the database (if it has any)? [00:56] of course that would make your file system a bit dirty, but given that you will install a package soon with all those files that shold work [00:56] Jazzva: i don't think so [00:56] but not sure [00:56] :) [00:56] let me see [00:57] I think I saw a message during the upgrades, that it's rebuilding something related to g-a-i, but not sure [00:58] Jazzva: /usr/sbin/update-app-install [00:58] thats run by postinst [00:58] thanks, asac [00:58] you might want to run that after copy [00:59] damn thing [00:59] getting all .xpi's from mozilla server appears harder than expected [00:59] yay... pushed :) [01:00] ;) [01:00] great! [01:00] http://code.launchpad.net/~jazzva/app-install-data-ubuntu/ubuntu.mozilla-extensions [01:00] Jazzva: maybe test and if it works for you ill tes ttoo ;) [01:00] * asac branches [01:00] ok [01:01] working :) [01:01] it shows adblock-plus, even if Icon field is set and icon is missing. [01:01] That's good, this way we're able to set Icon fields now, and then add icons later [01:01] good [01:01] (though, it is easier to check which icons are missing if icon field is not set) [01:02] Jazzva: why is there no icon for adblock plus? [01:02] I didn't put it yet :) [01:02] ah now i remember ;) [01:02] yeah ... sorry [01:02] this was just an example ... :) [01:02] perfect ... nautilus doesn't open ftp in its own anymore [01:02] btw, I have 6 png/gif/jpg icons to be converted to jpg ... could you do it? If you have some time available :) [01:02] now i have to search for some other ftp client [01:03] as normal ftp chokes on ftp.mozilla.org [01:03] maybe I could try with imagemagick [01:03] Jazzva: is only jpg supported? [01:03] afaik png should work fine as well [01:03] or am i wrong? [01:03] well, no. You said it would be cool to ship in xpm format [01:03] Oops, I wrote jpg instead of xpm [01:04] :) [01:04] when did i say that? last relesae cycle? [01:04] Yep... [01:04] hmm ... i don't think its needed anymore [01:04] hmm ... oke [01:04] xpm support was even dropped in ffox [01:04] ;) [01:04] I need to rename mozilla-mozgest to mozgest :) [01:04] umm... ok [01:04] :) [01:06] Jazzva: is ubufox itself in that app-install-dat apackage ? [01:06] I think it is... [01:06] (just curious) [01:06] yeah fine [01:06] hmm branching your branch is really slow :/ [01:06] you sure its pack-0.92 now? [01:06] all the commits? [01:06] probably [01:06] just branch ;) [01:07] Standalone tree (format: pack-0.92) [01:07] Jazzva: what size does the .bzr directory have? [01:07] lemme see [01:07] oh, 49.7MB :) [01:07] and everything? [01:07] e.g. whats the raw data size? [01:08] 11.6 [01:08] ok thats ok then i guess [01:08] should be faster though imo (the initial branching) [01:08] but maybe launchpad is just slow today [01:08] let's hope that's the reason ;) [01:08] Ok, I'll add the rest of the icons now :) [01:17] icons pushed (rev 542) [01:19] Should I submit bugs for extensions shipped with xpi/chrome/*.jar? [01:19] *report bugs [01:19] asac^^ [01:24] Jazzva: not high prio but would be good [01:24] Ok [01:25] great. thanks. pulling update [01:25] It was a pleasure :) [01:26] asac, another offtopic ... Do you think I could apply for a motu soon? Do you think my skills have improved enough? [01:27] :) [01:27] good question. [01:28] Jazzva: i think what is missing is that you do a few merges, when the doors open - i think you dropped out when hardy started? [01:29] Well, I did five for hardy, I think [01:29] Jazzva: you skills in bzr and all this stuff are definitly good enough and probably far better than of lots of MOTUs ;) [01:29] Jazzva: i would vouch for you. but i am not alone :) [01:30] https://wiki.ubuntu.com/Jazzva , three merges, two syncs :) [01:30] Yay, thanks :D... [01:30] (I still need to update my wiki page, according to some rules...) [01:32] asac, what do you suggest? :) [01:33] I would like to continue working with mozilla team, too. I really like it here :). [01:40] Jazzva: feel free to. ill ask dholbach and show him your wiki page [01:41] Heh ... Ok, I'll update it then :) [01:42] btw, there's a little bug with mozgest I found now ... it's simple to fix, the contents of chrome dir is not jarred before packaging. I'll upload diff soon [01:42] Is it ok if I do it? [01:46] Jazzva: why is that a bug? [01:46] It's not loading chrome components... [01:46] Jazzva: thats ok if the chrome.manifest doesn't refer to .jar file [01:46] e.g. chrome is not required to be jarred [01:46] it's reffering [01:46] ok [01:47] Jazzva: at best push a branch [01:47] base don ~ubuntu-dev one [01:47] ok... [01:52] asac, subscribed firefox-extensions [01:53] bug 216494 [01:53] Launchpad bug 216494 in mozgest "all blank with mozgest" [Undecided,New] https://launchpad.net/bugs/216494 [02:07] Jazzva: ok thanks. ill do that tomorrow [02:07] now sleep :) [02:07] n8 [02:07] Ok... I'm off too... good night [02:34] asac, I forgot to upload the changelog for app-install-data. I pushed the new revision to http://code.launchpad.net/~jazzva/app-install-data-ubuntu/ubuntu.mozilla-extensions (rev 543) [02:35] Now I'm off to sleep :) [02:36] ...and corrected version number (rev 544) :) [03:44] asac: need the two-minute refresher on updating the upstream === asac_ is now known as asac [08:53] asac: hi [08:56] carlos: hi! [08:57] asac: did you have time to test your part of the language pack script with latest export? [09:01] carlos: not sure if its latest language pack export. i exported my own on sat and used that [09:01] carlos: did anything change? structure? [09:03] it changes from the tarball you got on saturday [09:03] and it includes .xpi files [09:04] for en-US [09:04] but it uses the layout we agreed [09:12] carlos: hmm ... is the _big_ tarball online yet? [09:12] asac: yeah, that's the one I asked whether you did tests with ;-) [09:14] carlos: let me upload ubufox. then i go back to translations :) [09:14] :-P [09:14] asac: https://translations.launchpad.net/ubuntu/hardy/'+latest-delta-language-pack [09:15] that link gives you latest language pack with updates since previous base package [09:15] ah [09:15] ok [09:15] sorry, bad link... [09:16] yeah i figure [09:16] https://translations.launchpad.net/ubuntu/hardy/+latest-delta-language-pack [09:16] * asac downloads base pack again :) [09:16] and this other: https://translations.launchpad.net/ubuntu/hardy/+latest-full-language-pack is when we do a full export, which includes all translations (for firefox this is not yet true given that we started exporting firefox recently) [09:17] oh s the delta would be enough? [09:17] hmm [09:17] asac: the idea is that the delta one will include only languages with changes [09:17] asac: right now, yes [09:17] carlos: but complete files or diffs? [09:17] asac: after next base package is exported, you will need both if you want a full export with latest updates [09:17] carlos: lol [09:17] asac: complete files [09:17] i mean that delat pack is not really a win ;) [09:17] its 185M [09:18] the other was 443M [09:18] ok fair enough [09:18] but still not a mini-delta ;) [09:18] asac: well, that's just because we need to do a new base refresh (which happens before latest release) [09:19] asac: btw, if you upload ubufox, will I get template + translations? [09:19] did you prepare that package already? [09:20] carlos: no ubufox will be in intrepid sorry. [09:20] ok [09:20] np [09:20] carlos: midbrowser might get template though [09:21] carlos: but i have to discuss that with intel folks. i think we will do that for hardy.1 [09:21] asac: ok [09:30] <[reed]> why do you all feel that you need to re-translate all the Firefox strings when we already localize in 40+ languages? how many locales does Ubuntu support? [09:31] [reed]: we don't retranslate. [09:31] we supplement [09:32] <[reed]> supplement what? the ubufox-specific strings? [09:32] no [09:32] new languages. [09:32] <[reed]> oh [09:32] <[reed]> ok [09:33] [reed]: thats the idea. in pratice we had some issues with figuring the right context during import so we have to manually approve a few strings per translation ... but the suggestions are from the upstream translation [09:33] <[reed]> k [09:33] and we use upstream translations for those that are not complete obviously [09:34] [reed]: https://translations.edge.launchpad.net/ubuntu/hardy/+source/firefox-3.0/+pots/firefox [09:34] so there are already _new_ languages started ;) [09:34] obvioulsy those won't be ready for release. but anyway ;) [09:35] <[reed]> neat [09:35] * [reed] goes back to doing homework [09:36] yeah. when everything works find we give them back of course ;) [09:36] have fun [09:45] carlos: wanna test to extract es.tar.gz on top level dir and see if tere are any bugs in es? [09:46] asac: sure [09:46] carlos: wait a sec :) [09:48] carlos: http://people.ubuntu.com/~asac/translations.tar.gz there is a es.tar.gz inside [09:48] you have to extract that in / [09:48] but look in to the content so you are sure you don't wipe anthing else .) [09:50] carlos: ok i started it ... everything appears to be properly translated. only things that are broken are accesskey [09:50] carlos: you think you can go through the UI and see where he accesskeys are (X) in the ui and try to choose the right one? [09:50] if thats done i can whitelist es as locale that will be launchpadized :) [09:52] :-P [09:52] asac: should I restart firefox? [09:52] yeah [09:52] carlos: works? [09:53] * carlos restarts... [09:53] carlos: is your desktop es ? [09:53] yes [09:53] good ;) [09:53] mine is not german ;) [09:53] it detects both language packs [09:54] and looks like Spanish application :-P [09:54] carlos: yeah [09:54] carlos: you see he broken accesskeys? [09:54] like next to archivo [09:54] grrr,, I really hate those translators that use File (_F).... [09:54] its still "F" [09:54] it should be just _File (in Spanish _Archivo) [09:54] and the Ctrl+W [09:54] carlos: thats not that way [09:55] carlos: its just displayed that way [09:55] access key is "F" ... and F is not found [09:55] I see [09:55] those mozilla displays it next to the word [09:55] if you set accesskey to "A" it would work [09:55] but seems like the problem is that we only get English access keys [09:55] carlos: well ... someone confirmed them [09:56] carlos: not sure about "only get english access keys" i had the feeling that a few were correct in german [09:56] but not sure [09:56] i definitly had to go over all accesskeys and command keys and fix them properly [09:56] asac: let me see whether I can get a Spanish translator to fix it... [09:57] carlos: is there _any_ way a translator could search for lets say "accesskey" ? [09:57] not in the UI [09:57] carlos: maybe by exporting the .po and importing a .po as translation again? [09:57] does that work? [09:57] but they can get the .po file, fix it and upload it back [09:57] yeah [09:58] carlos: anyway ... es is a good whitelist candiate [09:58] asac: anyway, I only see that problem 5 times in the UI [09:59] carlos: thats good. but it doesn't mean that the others are correct and match the ones used in upstream .xpi [09:59] (not that we can't live with that for now) [10:00] well, once we kill the problem that not all upstream translations are accepted automatically [10:00] those divergences should be easy to detect and fix [10:00] carlos: hopefully :) [10:01] carlos: there is also a broken accesskey in context menu of download manager [10:02] and in the "organizer" menu of the bookmakr manager [10:02] (deshacer and eliminar have a broken key there) [10:04] asac: given that the text is displayed, I think is "easy" to know the key to press [10:04] but yeah, I will try to get that fixed by a Spanish translator ASAP. Anyway, if we deploy the xpi files, they will see it and fix it also as soon as they see it ;-) [10:05] carlos: if that key is wrong then there is no way to press the right key - but i think thats what you ment :) [10:05] carlos: well ... ill blog about it ... i will not whitelist any language that still has broken ui elements [10:06] because of trademark deals we have with firefox [10:06] ill blog and post intructions how translators can test themselves [10:07] so if lets say 4 more languages are ready in hardy + 1 month i can whitelist them at that point. [10:07] and so on [10:07] ok [10:07] carlos: but spanish looks good so far :) [10:07] except the accesskeys, but i am sure we get that sorted for release [10:07] german is definitly good as well. [10:08] ill look at swedish, chinese and a few more that are already "finished" [10:08] obviously chinese is hard for me to guess ;) [10:49] bug 188492 [10:49] Launchpad bug 188492 in ubiquity "console-setup/layoutcode settings are ignored" [Medium,In progress] https://launchpad.net/bugs/188492 [10:59] ArneGoetje: do you know if jimmy found a solution for his broken website? :-P [11:10] asac: nope [11:41] * asac lunchtime [12:08] hmm [12:10] bug 42019 [12:10] Launchpad bug 42019 in grub-installer "installer: grub password not verified" [Medium,Confirmed] https://launchpad.net/bugs/42019 [12:56] asac, the url from yesterday is not crashing on ati and intel. just nvidia. kind of strange. [12:59] fta2: pleaes try XaaNoOffscreenPixmaps [12:59] fta2: i have the feeling that thats the problem :) [13:04] fta2: there have been other wierd crashers reported with nvidia 8400 and 8600 [13:04] do you have such a card? [13:04] 7600 GS [13:04] hm [13:05] fta2: bug 212648 [13:05] Launchpad bug 212648 in linux-restricted-modules-2.6.24 "[nvidia-new] a visit to http://www.themareks.com/xf/ in firefox hardy causes X to restart" [Medium,Confirmed] https://launchpad.net/bugs/212648 [13:05] do you see that crash as well? [13:05] i will test tonight, my nvidia is at home [13:05] yeah sure === rzr is now known as rZr [13:53] carlos: you see why no.po is not exported for firefox translation domain? [13:53] carlos: its in xulrunner though [13:54] asac: actually.. no.po is a mistake [13:54] carlos: yeah [13:54] asac: no.po must not be exported ever, it's deprecated [13:54] carlos: ok ... whats its langcode? [13:55] nn and nb are the new ones, seems like I did a mistake [13:55] nb-NO ? [13:55] asac: the original one, yes [13:55] carlos: how do you map those? [13:55] however, they should move to use nb and nn directly... [13:55] but I understand we should be compatible with upstream [13:55] i now face the problem that the xpis have nb-NO and the export only nb :( [13:55] asac: didn't we agree that you will do such mapping in your side? [13:56] carlos: can't remember. but i can certainly map that. [13:56] carlos: however i need a complete mapping file or something :( [13:57] carlos: how do you deal with that when importing translations? [13:57] es -> es-ES, fy -> fy-NL, ga -> ga-IE, gu -> gu-IN, nb -> nb-NO, nn -> nn-NO, pa -> pa-IN, pt_BR -> pt-BR, pt -> pt-PT, sv -> sv-SE, zh_CN -> zh-CN, zh_TW -> zh-TW [13:57] hmm [13:57] but can we really map in that direction? [13:58] asac: except for pt-BR, en-* and zh-* the rest of languages don't need a country difference [13:58] the use of a country code is a waste of resources [13:59] carlos: so now i give pitti a es.tar.gz with es-ES inside? [13:59] will that work? [13:59] GNOME, KDE and other big projects share translations between countries (like Spanish and French) except for languages that are really different, like Chinese and Brazilian Portuguese [13:59] or will we miss some es-XX folks [13:59] ? [14:00] asac: well, that depends on firefox i18n infrastructure [14:00] in gettext [14:00] if you ship es_ES.po, only people from Spain get those translations [14:00] while es.po is shared between all countries [14:00] let me see [14:00] I don't know how firefox handle that, though [14:01] ok let me install the es-ES.xpi and see if it works when i start with LANG=es [14:02] carlos: ok appears to work [14:02] uff [14:02] finally i have luck ;) [14:02] i can start with LANG=es-XX firefox [14:02] and still get the es-ES translation [14:03] apparently firefox tries to be smart and tries more specific options if they exist [14:03] carlos: whats a valid es country code? [14:03] es-AG ? [14:04] es-MX [14:04] well [14:04] es_MX [14:04] for libc [14:05] carlos: can we remove the no thing then? [14:05] from the export? [14:05] ok let me try if mexican folks get english or spanish :) [14:05] ok they appear to be spanish ;) [14:05] at least thats good [14:06] :-) [14:06] yeah, I'm going to open a task in Launchpad translations to remove it completely from the UI [14:06] and the database [14:07] carlos: ok, but it probably won't hurt if i produce a no.tar.gz for pitti, right? [14:07] i mean he should to use it? [14:07] instead he takes the nb.tar.gz + nn.tar.gz [14:07] s/he should to/he should not/ [14:07] :) [14:08] right, there should not be any language pack for 'no' code [14:08] but better if you confirm it with him [14:27] carlos: i have es-AR as .xpi [14:28] do we have a langpack for that? [15:03] carlos: ok i have blacklisted ^no$ and have this mangling function in place ... http://paste.ubuntu.com/6994/ [15:29] carlos: help! [15:29] what is en.po ? [15:29] just blacklit? [15:29] blacklist? [15:30] asac: yeah [15:30] a side effect of the tricks we are doing... [15:30] about es-AR... that's the kind of waste efforts I told you [15:31] btw, from the link you gave me [15:31] asac: are you going to generate es.xpi files ? [15:31] carlos: i generate es.tar.gz :) [15:31] or is just a way to say 'look for es.po' ? [15:31] and currently i also have es-AR.xpi [15:31] aeh [15:31] tar.gz i mean [15:32] shouldn't hurt if there is not langpack for that imo [15:32] carlos: for things that come from launchpad we will automaticall yhave the right codes (launchpad codes) [15:32] isn't that any problem with the agreement we have with Mozilla? [15:32] carlos: for things that use upstream .xpi directly (everything not whitelisted) i am doing the mangling above [15:33] asac: I prefer to do it in that way, but just want to be 100% sure ;-) [15:33] carlos: its like this: [15:33] lets say we don't whiteliste es [15:33] then we would ship es-ES in es.tar.gz [15:34] carlos: i am not sure about mozilla ... if they don't complain we should not make a fud out of it imo. [15:35] ok [16:12] willguaraldi: fta complained about miro trunk taking about 25 seconds to startup ... might this be due to a big sqlite db? [16:12] for me latest 1.2 starts up quick (like 2-3 seconds) [16:13] however fta2's database was about 10-15 times larger [16:19] asac: yeah. the larger your database of channels and items, the longer it takes to start up. [16:19] asac: it's a problem we need to fix at some point, but will likely require some rearchitecture. [16:21] willguaraldi: ok ... so not just a missing index? [16:24] asac: well... the problem is that we've got an object database. so we load the whole database into memory at startup. [16:25] i wasn't around when they implemented that, so i don't really know why it was built that way. [16:25] willguaraldi: okay ;) [16:26] it's definitely a problem. i think bdk was looking into rearchitecture options, but was sidetracked into a different rearchitecture problem we have. [16:26] which is? [16:27] rendering items in the item list takes a long time. [16:28] many problems, not enough devs to deal with them. [16:29] asac: bdk figured out a patch for the drag and drop problem with the gtkx11 platform with xulrunner 1.9. i'm going to check that in today. [16:30] asac: i have a question for you, though. it looks like in your patches, you renamed one of the files. so the 80_xul patch involves deleting a file and re-adding it. it makes it really hard to figure out what changes you guys have for that file. [16:30] asac: any idea why that's there? [16:31] .c vs .cpp [16:32] i think it's a .cc file, not a .c file. [16:33] i'm pretty sure that .cc is a valid extension for C++ files. [16:33] willguaraldi: its because the standalone glue now needs to include gtkmozembed_glue.cpp [16:33] willguaraldi: but i think its not required anymore as 1.2 worked for me without changes [16:33] asac: what's the standalone glue? [16:34] also, that sounds like a new file altogether... not a renamed file. [16:34] willguaraldi: if you embed gecko you have to use the standalone glue ... if you write a component that gets embedded in gecko you use the dependent glue :) [16:34] asac: funky. [16:35] willguaraldi: gtkmozembed symbols are hidden nowadays ... and you can only get them by #include in your file that startsup mozilla [16:35] willguaraldi: howver, our python uses the standalone glue so if you first run into python and then load your C libs you should be fine [16:35] without using standalone glue [16:35] * willguaraldi nods. [16:35] i guess that you changed to order and thats why its now working with just libxul [16:35] startup order [16:36] mmm... i think on gtkx11, it's always been a python application. [16:36] willguaraldi: in intrepid we will provide two options for python: mozembed_standalone and _dependent ... so you don't run into those issues anymore. [16:36] willguaraldi: sorry i might be out of context. don't you mix python with native code= [16:36] ? [16:37] thought miro was one of the cases that did this ;) [16:37] my nomenclature might be wrong. [16:37] i'm a little new to this side of the app. [16:38] actually ... i though that you have a python application that also has some native bindings (e.g. for gtkmozembed) [16:38] yes, i think that's right. [16:39] right. i think in 1.1 the native binding was loaded _before_ the python module was included. [16:39] in 1.2 its different ... as it now works [16:39] not sure about trunk [16:39] fta2: do we still have a patch? [16:39] 1.2 is derived from trunk. [16:40] asac, nope. not anymore [16:40] my branch is on lp [16:41] so... are .cc files with C++ code in them bad? [17:11] willguaraldi: no .cc files _are_ c++ files [17:11] :) [17:12] its a bit confusing [17:12] CC stands for C compiler in make [17:12] CXX for C++ :) [17:12] but .cc extension is definitly c++ [17:15] ok, that's what i thought. i don't have the patches in front of me, but i'm pretty sure the file that the 80_xul patch is renaming is MozillaBrowserXPCOM.cc . i think you can take the renaming out of the patch. [17:16] willguaraldi: we don't need to rename anyway anymore :) [17:16] willguaraldi: if i renamed it then because you compiled .cc file with gcc not g++ [17:16] funky. [17:16] so the confusion might have started on your side [17:17] but better forget about this ... apparently we don't need any patch anymore [17:17] oh, i'm not placing blame anywhere. i just wanted to talk about what's in the patches so i can figure out what i can absorb into the source here to reduce the work you folks have to do. [17:19] i don't see "gcc" or "g++" in the setup.py file. maybe it was a change in distutils between python 2.4 and 2.5? [17:20] no biggie, though. i'm going to try to put out a 1.2.3 rc0 today. do you want me to send you the revision numbers for the "fix" that fixes drag and drop and then you can absorb it into your packages? [17:39] asac: ping? bug #202343 updated. [17:39] Launchpad bug 202343 in opensc "mozilla-opensc firefox plugin not visible in FF3 (bad install directory)" [Undecided,Incomplete] https://launchpad.net/bugs/202343 [17:42] mlind: lookig [17:43] mlind: mime-type is important [17:43] asac: I though so, what would be the correct value? [17:43] mlind: the mimetypes you see in about:plugins [17:43] asac: ah, okay [17:48] asac: Security modules supposed to show in about:plugins ? [17:51] mlind: no idea [17:51] if they serve any mime type then probably yes [17:51] otherwise its not really a plugin imo [17:56] mlind: is it not in there? [17:57] asac: nope. [17:57] mlind: does it work? [17:58] asac: I don't actually have a device to test this, but security module loads and registers without errors [17:58] how do oyu see that? [17:58] asac: in FF's "Security Devices" dialog [18:00] asac: /usr/lib/opensc-pkcs11.so loads okay, but card reader status is "not present" [18:02] asac: maybe leave the Xb-Npp headers stuff out as the module doesn't seem to serve any mime type? [18:04] mlind: yes thats ok then [18:06] asac: okay, I'll submit a new debdiff and leave the Xb-Npp stuff out. Thanks. [18:07] mlind: thanks [18:07] mlind: let me know when ready [18:25] asac: okay, debdiff in bug #202343 should be good to go [18:25] Launchpad bug 202343 in opensc "mozilla-opensc firefox plugin not visible in FF3 (bad install directory)" [Undecided,Incomplete] https://launchpad.net/bugs/202343 [18:32] mlind: ask for an uploader in -motu ... if there is none ill upload later today [18:35] asac: will do [20:11] asac: whose job it's to put the translated versions of the firefox start page (that are available in ubuntu-docs and pointed to by ubufox translations) in place to start.ubuntu.com? [20:17] (since the translations were now lost with the hybrid start page) [20:30] Mirv: which translation got lost? [20:45] asac: Firefox start page translations, ubufox now switched to some hybrid start page model and ignores the ubufox URLs that pointed to ubuntu-docs' translated versions of the index.html [20:46] Mirv: yes [20:46] Mirv: do you have a langpack installed? [20:47] like finnish? [20:48] ok, I had that on an another computer, will try with it and the new ubufox [20:48] Mirv: what do you see in location bar? [20:48] (quickly chrome://something and then start.ubuntu.com/8.04) [20:48] Mirv: you will only get the translated page according to the translation you use [20:48] Mirv: yes, thats the online page [20:48] it will always use the language you have in firefox [20:49] but even the local page should be detected properly [20:49] to see that you have to disable disc caching in firefox about:config [20:49] stop firefox [20:49] disable networking in network manager [20:49] start firefox [20:49] that should go to index-LANGCODE.html [20:50] ArneGoetje: scim is popping up again :( [20:50] ctrl+pgup -> scim pops up :/ [20:50] now it pops up whenever i push ctrl [20:51] and exit does only prevent that temporarily [20:51] asac: ok. retried. offline start page good, online start page bad. so my question was actually still relevant, who should get the translated home pages up to start.ubuntu.com (and do some language detection somewhere)? [20:54] asac: anyway, good that the offline start page is still working as it should be [20:55] Mirv: this is all a "last minute action" ... the guy told me that he now also has index.html.fi [20:55] can you retry? [20:55] Mirv: we are establishing a process to maintain that home page in a bzr reop [20:56] asac: yeah, I see. it's easier to fix the online part anyway than the offline part. [20:56] Mirv: right. does it work now? [20:56] asac: retried, English still. I can manually open the index.html.fi, though, but some configuration is still not there. [20:56] online? [20:57] Mirv: is firefox in finish now? [20:57] asac: yes, and offline page is shown as finnish and Finnish is selected in Firefox's preferred languages [20:57] (ie. other pages that are aware of multiple languages work) [20:57] ok i asked him to come here [20:58] newz2000: oh, it's you, hi :) [20:58] hi newz2000 ! [20:58] hey Mirv [20:58] long time no chat [20:58] newz2000: yep, doing so great at canonical's data centre... [20:58] newz2000: anyway, browser doesn't automatically show index.html.fi yet [20:58] Mirv: just to be safe, can you try a shift+reload? [20:59] newz2000: yes. [20:59] I also disabled (again) disk cache and started browser again. [20:59] Mirv: can you install http liveheaders extension and tell us whats in accept-language [20:59] ok, just a sec [20:59] (warning: unqualified comment) [21:00] Mirv: the package ... no idea if upstream has updated the maxversion yet [21:00] nope, they haven't. what was the key to override it? [21:01] Mirv: install the package ;) [21:01] Mirv: mozilla-livehttpheaders [21:02] apt is my friend, yes. Accept-Language: fi [21:02] newz2000: ^^ [21:03] sounds like .fi should match then [21:03] yeah, testing now [21:03] Content-Location: index.html.en [21:03] crazy ... i wondered why i have so much space on my 12" screen ... now i noticed that gnome-panel is gone :/ [21:04] lol ... full screen mode ;) [21:04] ok started in english ... [21:04] verified that cache is enabled [21:05] now trying german :) [21:05] works [21:05] so its not a general regressoin at least [21:06] yeah, let me trying bypassing the frontend proxy [21:06] newz2000: maybe the cache now remembers (because we already tried) and the server is broken :) [21:06] s/cache/proxy/ [21:06] :) [21:06] right, that's my theory [21:08] nope, not so simple apparently [21:09] asac: did you enable 'complex script support' in language-selector? [21:10] ArneGoetje: i did do nothing ;) ... what i did was a reboot after the latest upgrades ;) [21:10] asac: nothing has changed in scim [21:10] yeah ;) ... no idea when i rebooted last time [21:11] let me check language selector [21:11] asac: what does 'im-switch -l' say? [21:11] ArneGoetje: ok i start language selector and get "the language support is not installed completely" [21:11] looks like i never used that menu [21:12] ArneGoetje: yeah [21:12] ArneGoetje: its enabled [21:12] don't ask me why [21:12] asac: disable it and relogin [21:12] ArneGoetje: sure. i managed to get rid of it for this session (dont ask me how) [21:12] ArneGoetje: is there a key combination that would enable that? [21:12] (by accident)? [21:13] * newz2000 contacts server admins for ideas [21:13] asac: crtl+space, alt+grave, crtl+shift, crtl+pg up/down [21:13] ArneGoetje: he? that will switch "complex scripts" on? [21:14] i use ctrl+space in emacs all the time ;) [21:14] asac: no. they enable scim when 'complex scripts' is enabled. [21:14] ok [21:14] ArneGoetje: why is that cluttered on all kind of key combinations? [21:14] instead of one? [21:15] i use ctrl+pgup/down also frequently in my daily workflow [21:15] asac: FYI: crtl+space and alt+grave toggle scim on/off. crtl+space is the chinese way and alt-grave is the japanese way. [21:15] i wonder how much pain scim users support ;) [21:16] newz2000: if I manually go to index.html.de, I see in headers Content-Language: de, de, but if I go to index.html.fi I don't see Content-Language [21:16] asac: crtl+shift / shift+crtl and crtl+pgup / crtl+pgdown flip through the available input methods in scim. As the default (when turned off) is your native keyboard, flipping one up or down will automatically enable scim and select the next/previous input method. [21:17] Mirv: yeah looks like fi is not known in the default lang mapping file most likely installed ;) [21:17] asac: CJK users expect that behaviour, so no pain here. ;) [21:17] thats my guess [21:19] ArneGoetje: is there an easy way for me to see if that i am looking at http://start.ubuntu.com/8.04/index.html.zh_CN vs. http://start.ubuntu.com/8.04/index.html.zh_TW? [21:19] asac: ? [21:20] ArneGoetje: lets do it simple ;) ... install both language packs please ... and start firefox a)with TW ...and b) with CN ... and tell us if the online startpage shows the right language ;) [21:20] asac: ah... [21:20] http://releases.mozilla.org/pub/mozilla.org/firefox/releases/3.0b5/linux-i686/xpi/ [21:20] there is where oyu can install ffo xlangpacks ;) [21:27] Mirv: is https://help.ubuntu.com/6.06/ in finish or english? [21:27] wow... There's a serbian langpack. Didn't know it existed. First time I see FF in serbian :). [21:28] (sorry for interruption) [21:30] Jazzva: from upstream? [21:31] whats its langcode? [21:31] sr [21:31] Mirv: I think the start page now shows up right for fi [21:32] Followd the link you provided [21:32] *Followed [21:32] asac: hmm... even when I start firefox in zh_TW locale, it still shows up the english page by default. [21:32] newz2000: help.ubuntu.com English [21:32] newz2000: yes, start.ubuntu.com is now Finnish! thanks! [21:32] sweet. I'll check the list of languages and compare to apache's config and see which ones apache doesn't like. [21:32] ArneGoetje: is firefox in zh_TW? [21:32] (Ng gets credit for this fix) [21:33] newz2000: ^^ [21:33] let me test [21:33] newz2000: maybe check the config first ;) [21:33] newz2000: ok, great, and thanks to Ng too. should help.ubuntu.com be fixed similarily too (in some future)? [21:33] faster than test ;) [21:33] Mirv: I'll report it [21:33] asac: I switched the whole system to zh_TW in langauge-selector. [21:33] ArneGoetje: well ... that means nothing [21:34] ArneGoetje: important is that firefox is translated [21:34] zh_TW is en [21:34] :) [21:34] newz2000: ok [21:34] asac: looking in firefox settings-> preferred languages lists zh_TW at the top like expected. [21:34] asac: thanks for invoking newz2000. he's the guy who (had to) spent (way too) much time on keeping ubuntu-fi.org running when we were on worse servers than nowadays [21:34] ArneGoetje: how does the UIlook like? [21:34] Mirv: ;) [21:34] asac: chinese [21:34] Mirv: this is a fun project. Wait until you see the final result. ;-) [21:35] ArneGoetje: ok ... i think its not yet supported server side then :) [21:35] asac: the UI is correctly translated into zh_TW [21:35] asac: probably [21:35] yeah ;) [21:37] asac: so, what did you want to test with the zh_* Ubuntu pages? [21:37] ArneGoetje: just figure if they are properly mapped. newz2000 wasn't sure, but now we have this config problem to sort out first [21:38] asac: config problem? [21:38] apache doesn't like some languages [21:39] ah.. so you mean the apache config problem? [21:39] ArneGoetje: the reason why its in english for you still ;) [21:39] yeah [21:39] ok. [21:39] ArneGoetje: what is your preferred language? [21:39] en [21:40] but when I start FF in zh_TW locale, the setting in FF lists zh_TW as preferred language [21:40] and zh_TW isn't working? [21:40] nope, it shows the en page [21:40] gotcha [21:41] there are 68 langs so I'll script it to see which are working and which aren't. [21:41] oh [21:42] newz2000: we have 68 translated? or 68 in apache? [21:42] there are 68 different files, some are duplicates, for example .pt .pt_BR [21:42] but they still need to be tested [21:42] newz2000: interesting [21:43] yeah, that's why I was wondering if you knew how the tralnslated start pages were generated because I definitely want to figure it out and use it to generate the new start pages [21:43] newz2000: we only have 42 lang packs for ffox [21:43] that's probably about right [21:44] aeh 46 [21:44] ;( [21:44] and 6 more blacklisted or so [21:44] (people just started to translate ffox in launchpad) [21:47] do you guys still need my help or can I go to sleep now? [21:47] ArneGoetje: sleep well [21:47] thanks [21:48] thanks... almost 5:00 am here [21:56] asac, please check https://wiki.ubuntu.com/Jazzva if you have a minute... I have added some new info, so I'm interested if it's ok... Thanks :) === fta_ is now known as fta [23:52] asac, http://paste.ubuntu.com/7056/