[00:01] it was one of the leaks [00:01] lets see [00:01] mozilla bug 405032 [00:01] Mozilla bug 405032 in Preferences "Changing feed handler to web handler leaks nsGlobalWindow" [Normal,Assigned] http://bugzilla.mozilla.org/show_bug.cgi?id=405032 [00:01] mozilla bug 388573 [00:01] Mozilla bug 388573 in General "ForecastFox 0.9.5.2 extension leaks 2 DOM Windows on Trunk but not on Branch" [Normal,Assigned] http://bugzilla.mozilla.org/show_bug.cgi?id=388573 [00:02] mozilla bug 388577 [00:02] Mozilla bug 388577 in General "FoxClocks 2.1.93 extension can leak nsGlobalWindows on trunk (but not on branch)" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=388577 [00:07] hmm mozilla bug 270159 still exists on trunk :( [00:07] Mozilla bug 270159 in Download Manager "Download manager adds extension regardless of file's own extension" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=270159 [00:13] hmm, cannot find it anymore. will look tomorrow [01:53] jetsaredim: ok mozgest dev landed license.txt [02:21] asac: how did you find that out? [02:22] i see it too [02:22] ok - i'll work on a package later tonight [04:50] asac: fyi - uploaded those mozgest code branches [05:23] wb [08:05] jetsaredim: thanks [08:05] the discussion went on and he apparently removed you from loop. finally he agreed to what we see now [08:08] * asac yawns [12:21] asac: I figured that [12:22] yeah. but in the end he agreed ... which is a success i guess [12:38] asac: with the UI/string freeze tomorrow, what's the plan with ubufox translations? bug #139380 would now have fi-FI, hu-HU and ru-RU translations. can you add those to the source tree, or have you heard something from the Launchpad folks? [12:38] Launchpad bug 139380 in ubufox "Untranslated strings in ubufox" [Undecided,Confirmed] https://launchpad.net/bugs/139380 [12:39] Mirv: yes, we will add them. we should also call for translations. [12:40] Mirv: https://code.edge.launchpad.net/~ex/ubufox/main thats the branch where i see ru-RU and fi-FI [12:48] asac: ok, great. the branch has old version of fi-FI translation, the bug's last comment has a newer version which also has the startup_homepage_*_url:s "translated" (I assumed that they're supposed to be translated to eg. index-fi_FI.html from ubuntu-artwork folders) [12:48] but it'll be good to have a post to ubuntu-translators list about the subject, so that translations are gotten and also people know what/how to translate [12:50] Mirv: doesn't the localized hompage work without tweaking hompage? [12:51] asac: not in the current version at least (/usr/share/ubuntu-artwork/home/index.html) [12:54] hmm [14:21] asac: hi [14:21] carlos: hi ... i need to get some food now :) [14:21] 30 minutes i guess? [14:22] sure === asac_ is now known as asac [14:52] carlos: now that we know that we won' get regular export for xpi i wonder if we can export something else? [14:53] carlos: we won't? [14:53] danilos: not for Hardy. [14:53] not for hardy [14:53] well, for Mozilla we are not importing anything yet [14:53] ah, right [14:53] asac: the only thing we could give you is .po files, is that useful for you? [14:54] sorry for jumping in the middle of conversation [14:54] thats why i am asking [14:54] we would need to setup the import, though [14:54] asac: that should be possible, and we could start exporting to XPI later when we have that working. [14:54] asac: I invited jtv and danilos to join us [14:54] if we could get .po files i would figure out if we could create the proper format somewhere else [14:54] hi all :) ... much appreciated [14:54] hi asac :) [14:54] asac: hi :) [14:55] ok ... so what can we import atm? [14:55] asac: translation-toolkit from pootle is supposed to produce .po files from xpi files and get them back from the .po files [14:55] hi asac [14:55] asac: but we were not able to get it running for Firefox [14:55] carlos: but they don't exactly work with any PO files [14:55] we use it for OO.org though [14:55] carlos: yes, i would look at that [14:55] danilos: I know, that's why I was trying to point that the whole process needs to be done with pootle [14:55] carlos, asac: I can look at the old code I have written for xpi exports, maybe we can reuse parts of that [14:56] asac: our .po files would not be compatible with translation-toolkit layout [14:56] danilos: to produce a script out of Launchpad? [14:56] thats ok. i just would like to take a look how they are assembled [14:56] carlos: well, to make a script to merge stuff from PO files and using en-US.xpi file [14:57] danilos: yeah, that was what I was thinking [14:57] getting that working might not be so much easier than getting exports working [14:57] ok so the idea is to import po files produced by pootle [14:57] right? [14:57] carlos: of course, that might include parts of LP code (like parsing en-US.xpi files) [14:57] (lets stick to the import side for now) [14:57] asac: also, we are waiting for a solution to validate .xpi files from Mozilla developers, that would help us to validate the output either from Launchpad or with some temporary script [14:57] jtv: not really, I had that working, only bits like install.rdf files and similar would have to be hacked around, and with non-LP code, we can hack around as we please [14:58] danilos: that's pretty cool [14:58] jtv: right, ugly hacks are faster, although is a hell to expand ;-) [14:58] but for this case, would work [14:58] jtv: and I had no regression tests either :) [14:59] carlos: jtv: i don't want any hacks. I would to figure what to do to make use of existing features in launchpad [14:59] danilos: some testing would be appreciated though, to prevent Firefox breakages [14:59] asac: well, using translation-toolkit is a hack too [14:59] s/any hacks/any hacks on LP side/ [14:59] ok [15:00] ok :) [15:00] asac: this would not be on LP side, it'd be on your side :) [15:00] asac: what danilo suggests is just provide a script to do such hacks outside Launchpad ;-) [15:00] so what import can you process as of now? [15:00] asac: gettext and xpi. [15:00] ok ... what happens if i throw an xpi at you? [15:00] asac: the template needs to be handled manually, but translations are handled already [15:00] asac: should work, though could do with some real-world testing. [15:01] asac: if I prepare a hand made .xpi to import en-US strings, users will be able to translate Firefox right now using Launchpad [15:01] carlos: ok. i could produce that during build i guess [15:01] asac: but, I'd prefer it if en-US.xpi was created along with the package build [15:01] what format do you expect? [15:01] asac: yeah, that would help [15:01] danilos: "me" [15:01] any special layout or can you parse any .xpi ? [15:01] danilos: ;-) [15:02] asac: we don't care about manifest files in our current implementation [15:02] carlos: why do you say "hand made" ? [15:02] asac: well, firefox doesn't provide en-US.xpi by default since 1.5.0.6 or something [15:02] so we have as a restriction that everything inside a .xpi files is for the same locale [15:02] (just to understand the current state: why couldn't we import an .xpi from the mozilla mirrors) [15:03] danilos: ok thanks. [15:03] carlos: ok thats good to know [15:03] asac: an xpi file with the en-US.jar file would be enough [15:03] carlos: no other meta information required? like install.rdf? [15:03] asac: afaik you could take a full one, delete some directories, and re-pack. [15:03] asac: the translations from mozilla mirrors are good for translations, but as long as I know, en-US.xpi doesn't exist at all [15:03] asac: not used right now [15:03] e.g. just en-US.jar in top level directory? [15:04] jtv: ok ... and import translations. would that work? [15:04] asac: hmm, having one is not harmful [15:04] although we don't use it at all right now [15:04] asac: it should, yes [15:04] do you look at any .jar inside the .xpi? or just those named en-US.jar ? [15:04] asac: any, iirc [15:05] asac: what danilos said [15:05] thats good. so you look for .properties files and .dtd ? [15:05] asac: yes [15:05] any .properties? [15:05] (most likely just those in a .jar) ? [15:06] asac: to provide us .xpi files, you will need to ask pitti to update his script that extracts translations to extract .xpi files too [15:06] carlos: can't we do this manually once to evaluate? [15:06] asac: any [15:06] asac: sure [15:06] and if we figure things out, do it automatically? [15:06] ok great. [15:06] for the testing phase, we could [15:06] and should, really [15:07] ok ... so what happens next? do you keep meta info about the key used by the xpi for each translation snippet? [15:07] asac: we get all .properties and .dtd files because we don't parse the .manifest file, but with this concrete case, it shouldn't be a problem [15:07] asac: yes [15:07] asac: we keep using those as message identifiers, but show the en-US "translations" in the UI [15:07] but i guess that meta info is not included in a .po exported, right? [15:08] ah good [15:08] that works already? [15:08] asac: yes [15:08] carlos: you'd have to export an en_US PO file, right? [15:09] jtv: we don't need to export en_US i guess, because we usually have that [15:09] otherwise, yes. [15:09] jtv: why would I like to do that? [15:09] ok, so could we throw just http://people.ubuntu.com/~asac/ubufox.xpi in? [15:10] carlos: phrased that badly... I mean, we also keep the symbolic msgids in the exported PO files, right? [15:10] asac: the problem with that, is that it doesn't use xpi files, if you produce fake ones per language, yes [15:10] jtv: yes, we do [15:10] carlos: ubufox.xpi is equiv to en-US.xpi [15:10] jtv: maybe, but that's very easy to break (that's what pootle/translate-toolkit do, which is why it's so fragile) [15:11] there are only en-US entitties in [15:11] asac: it contains translations [15:11] asac: really? [15:11] is that new? [15:11] yes. i didn't include any translation yet [15:11] last time I checked it had some translations [15:11] hmmm [15:11] lets see [15:11] if so, i can fix that [15:11] hmm [15:12] there is only locale/en-US [15:12] right, it doesn't include any translation... [15:12] how do you plan to deploy translations? [15:12] and contains one .dtd + one .properties [15:12] so a decent target to test [15:12] carlos: what do you mean? [15:12] how to package? [15:13] how to deploy translations [15:13] yes [15:13] I can investigate writing a script or two to actually construct xx.xpi's from a PO file exported from Launchpad and en-US.xpi (well, ubufox.xpi if needed :) [15:13] we could include them in language packs [15:13] danilos: that would be wonderful [15:13] not thought in detail about that. either one deb for each language or one deb for all (except en-US naturally) [15:13] asac: ok, so language pack is its place [15:13] no need to add new packages [15:13] no idea where it belongs :) [15:14] but most likely yes [15:14] asac: I do ;-) [15:14] carlos: that would make a lot of sense, but we need to know how would they be "activated" in firefox [15:14] at least that's the plan for firefox [15:14] danilos: there's a good one. Let's hope it's scriptable. [15:14] carlos: ok, then lets move on with the theory. ubufox.xpi has just a few entities, so i guess its a good try. can we do that? [15:14] danilos: install them as an extension in /usr/lib/firefox/extensions/ [15:15] carlos: danilos: new place is /usr/lib/firefox-addons/extensions/ [15:15] asac: sure [15:15] or /usr/lib/xulrunner-addons/extensions/ [15:15] asac: is that enough for firefox/anything else to register them [15:16] if it is, excellent [15:16] ? [15:16] (just completing danilos' sentence there) [15:16] jtv: :) [15:16] yes, if its a complete .xpi (unzipped) it will work [15:16] asac: I guess we should use IDs like: language-pack-LL-CC@ubufox.ubuntu.com or something like that for each language pack [15:16] but lets first work about getting an .xpi ... the rest will be easy imo [15:16] carlos: yes ... we can do all that [15:16] we can try it out on staging [15:17] ok, can we get an import round and than see what happens if you export en_US? [15:17] or do we need an initial translation? [15:17] jtv: that's going to be a problem, each import will require Tom around [15:17] jtv: what about using demo? [15:17] ok so Tom is the bottle-neck now? [15:17] ok figure out first ;) [15:17] carlos: does it run scripts and send email though? [15:18] asac: Tom is needed to simulate cron on staging. [15:18] jtv: not by default, that's why Tom is the bottle-neck [15:18] asac: we don't need any translation [15:19] asac: if you are able to provide me with an xpi for ubufox, we can try to get it imported today (Tom should be around right now) [15:19] carlos: take the one above ... or wouldn't it work? [15:19] carlos: code on demo is pretty old [15:19] the problem is with file exports and any other updates [15:20] http://people.ubuntu.com/~asac/ubufox.xpi [15:20] jtv: if no one is using it, we may get it updated [15:20] asac: let me try in my local computer.. [15:20] \o/ [15:20] carlos: this would certainly be worth requesting that [15:20] jtv: in worse case, we could use production [15:20] but hiding that template [15:22] carlos: okay, let's try on demo, and if anything breaks that we know is fixed already, ask for an upgrade. Failing that, we can look at production. [15:22] carlos: why hiding :) ... i need translations anyway ;) [15:22] asac: for testing [15:22] sure, just kidding [15:23] asac: well, I'm fine to leave it there... but would be confusing if we are not able to deploy it :-) [15:23] let's not get lost in this "what-if" case though. [15:23] ok ... all set? [15:24] import .xpi asap ...then try to export en_US [15:24] ? [15:24] (in the current default format) [15:24] asac: one thing: I think the name of the XPI file needs to be en-US.xpi [15:24] i assume i can export "per-package", right? [15:24] jtv: just rename to test then [15:24] jtv: I can do some 'magic' to avoid that [15:24] carlos: sounds exciting. [15:24] i don't thinks its a problem to get the name right before import if we do it automatically at some point [15:25] jtv: but only for testing [15:25] jtv: edit the path for the import queue entry :) [15:25] danilos: sssh, that kills the magic!! [15:25] carlos: ok, ok :) [15:25] :-D [15:26] Okay, now that Carlos' deep dark secrets have been revealed... [15:26] ok, how does the current export work? i ask launchpad: "export .po files for package XYZ"? [15:26] and then i get a tarball with all translations? [15:26] asac: you request an export in the normal way, and it'll ask you which format you want [15:26] asac: but PO and MO are the only options for now. [15:26] what is "the normal way" ? [15:27] asac: "Download translations" link. [15:27] actually, my real question is if launchpad remembers the entities needed for each package? [15:27] It's in the actions menu on the left. [15:28] so ... if i provide an en-US.xpi for xulrunner, i can get .po files for all the entities found in that .xpi? [15:28] (i think its too simple to be really a question:)) [15:29] carlos: it all comes out as one big PO file, right? [15:29] asac: it's a bit more complex (entities -> english strings -> PO files, and then you can basically get translations for all english strings which were linked to entities :)) [15:29] carlos: I mean, per XPI file per language. :) [15:29] one PO file per language? [15:29] asac: one PO file per template per language. [15:29] asac: IIRC an XPI file is treated as a single template. [15:29] asac: but that's essentially what I'm asking carlos. [15:30] asac: just a minute, I'm preparing an export for you... [15:30] ok lets assume i want to export entities for openoffice? [15:30] do i need to provide all entities i want to export or just the name "openoffice" ? [15:30] to get the translations for that package? [15:31] hmm, I found a missing feature with the .po export... [15:31] asac: https://pastebin.canonical.com/2858/ [15:31] that's current output, Launchpad UI shows the English strings, though [15:31] hmm paste.ubuntu.com? i currently have no pass for that [15:31] asac: look at #canonical channel [15:31] * asac looking [15:31] hmm, they removed it from the topic... [15:32] hmm its not there ... looking i wiki [15:32] asac: it's the same we use for our wikis [15:32] ok lets try ;) [15:33] yeah works [15:33] ok so it doesn't export the actual en-US texts? [15:33] carlos: the missing feature would be to add the English strings as comments? [15:33] would it work for translations? [15:33] asac: this is the Spanish translation, not the en-US "translation" that Carlos exported. [15:33] (as i am not interested in en-US exports atm) [15:33] he? [15:33] jtv: actually, to export the English string in the msgid [15:33] is there a spanish translation yet? [15:34] asac: no, but I requested its export, and given that there are no translations, it's a file without translations [15:34] let me show it with some translations... [15:34] carlos: makes sense [15:34] can you add one or two translations [15:34] ? [15:34] i could then start to code the .xpi producer for that ;) [15:34] looks rather straight forward [15:35] asac: https://pastebin.canonical.com/2860/ [15:35] that includes a translation [15:35] yeah [15:36] ok ... so what kind of script do you need? [15:36] lets start different: [15:36] how is the language-pack maintained? [15:36] is that something automatized? [15:36] carlos: should we have had this be exported with English messages as msgid's? [15:37] danilos: yeah, that's why I said I found something missing for the .po exports [15:37] danilos: but for what we want right now, it should be enough [15:37] danilos: inserting en-US msgstrs wouzld be safe, but i think we can try without them [15:37] carlos: right, ok [15:37] asac: of course [15:38] asac: what we do right now is to export the files to a tarball that Martin downloads and process automatically to produce the final .deb [15:38] ok ... this automated process is coded in a source package ? [15:38] so if i do apt-get source language-pack i can take a look? [15:38] what we would need is to add there all the scripts to produce an extracted .xpi that is installed at /usr/lib/firefox-addons/ [15:38] asac: langpack-o-matic, it's on LP [15:39] asac: I don't know all details, you will need Martin Pitt help [15:39] he does some preprocessing in our servers [15:39] carlos: actually he just documented it! [15:39] asac: I have some existing code I can share, though it might not be considered free software yet (I did make it while on company time and while working on LP :)) [15:39] ok if this turns out to be too difficult, we could most likely export all mozilla translation separately? [15:39] and the conversion from .po file into the native format should be done in that stage, given that it will depend on non free code coming from Launchpad [15:40] (e.g. i could do a mozilla-lang-pack source package if we otherwise need changes on launchpad side) [15:40] asac: sure, although at some point, it needs to be deployed in language packs, so better if we do it now [15:40] Pitti's docs for the language pack process: https://wiki.ubuntu.com/TranslationLifecycle [15:41] let me look [15:41] asac: not that I think of, Launchpad has all the import process working, is the export what is missing (and improve the imports, but basic functionality is there) [15:41] carlos: i mean: "export as .po files" [15:42] i can't believe that thats missing because you showed me an export ;) [15:42] :) [15:43] asac: .po export is not missing ;-) [15:43] I was talking about .xpi exports :-) [15:43] yeah ... so in worst case we could get a separate export of .po files for all mozilla packages? [15:43] asac: anyway, all this depends on Danilo being able to do that hack to produce the .xpi from the .po files [15:43] asac: yes [15:43] asac: that should work, and either separate or all together in a big tarball. [15:44] you can request them when you want to [15:44] carlos: he? you don't need to produce that hack. i can write a script for it [15:44] looks easy [15:44] asac: ok, if you think you can handle it, perfect ;-) [15:44] all information i need is: .po file + language + package [15:44] asac: ok, if you feel like doing it, be my guest, but I've already had working code for this :) [15:44] asac: your experience patching up the rdf etc. could be useful to us, in fact [15:45] danilos: can you ask your boss if you can free the code you currently have, so i can look at that [15:45] :) [15:45] jtv: boss, can I... :) [15:45] danilos: I already asked our boss [15:45] He's talking to his boss [15:45] asac: it's in process :) [15:45] wow ... i thought we had flat hierachies [15:46] asac: it's about the same depth as my gym in Bangkok, which has about a dozen staff. [15:46] asac: anyway, if you feel like it's simple, I guess going through administration is pita :) [15:46] i hate redoing things [15:46] asac: seriously though, can't make any guarantees that it will be approved, but I'll add this as a case in point. [15:47] who is your boss? [15:47] asac: Kiko. [15:47] ok [15:48] danilos: can you outline how you do it? [15:48] so i can see if i am thinking about a similar approach right now? [15:48] asac: sure [15:48] (though this is a pretty old code) [15:49] just a brief description (looking at https://pastebin.canonical.com/2860/) [15:49] asac: basically, recursively walk through a zip file and modify contents in-place as soon as I run into any messages needing translations (according to a PO file which is read at the same time) [15:50] ok there is a far simpler solutoin i guess (please think about it): [15:50] asac: anyway, you have a NDA signed with Canonical as a Canonical employee, so you should be able to see the code :-) [15:51] just add entities for to .dtd files and simple properties to .properties file [15:51] iterate through the .po file and do that for each msgid [15:51] the file to append this to can be parsed from the comment [15:51] done [15:52] asac: that could work, but I suggest doing this with an en-US.xpi file so you can actually modify existing files, and keep the English strings where there are no translations in the PO file [15:52] well ... i can implement that easily by looking at en_US.po file as a fall back [15:53] it looks easier for me to just append to files than replace (given that there might be entity ids somewhere in real texts) [15:53] maybe reuse the install.rdf from en-US.xpi though [15:53] asac: it needs some modifications AFAIK [15:53] yes reuse == copy + patch :) [15:54] maybe ffox will auto fall back to en-US on its own though. have to test that again [15:55] Sounds plausible. [16:00] asac: are you testing? [16:09] asac, hello ;) [16:12] jtv: not yet. got distracted [16:14] asac: so try uploading on demo, and letting us know so we can push it through. [16:14] me? [16:14] why would i upload anything? [16:15] (sorry ... i miss some context here. what is demo)? [16:15] and how can i access that [16:17] asac: I'm sorry, I thought you were planning to do that. [16:18] asac: demo is demo.launchpad.net [16:18] asac: for longer-running demo use of launchpad, without polluting production with test data. [16:18] ah right ... i am still in evaluation phase :) have to figure out langpack-o-matic [16:19] :) [16:19] so i know how to hook in [16:19] the import and export (well more or less) is clear for me atm [16:19] asac: then anything else you need? [16:22] not yet. on friday i will know more i guess [16:22] thanks! [16:23] asac: thank you. I'll quietly sidle away from the channel now. :-) [16:23] jtv: one more thing [16:23] asac: I knew it :) [16:23] in the .po file i don't see the lang code (e.g. spanish) [16:24] would the real code need to guess that from .po filename? [16:24] asac: right, it's really only in the file name. [16:24] ok ... how does the filename read? [16:24] $PROJECT-en_US.po ? [16:24] asac: there are several patterns. [16:25] what would we get? [16:25] asac: ideally (and if you'll be controlling the project, that can be reality) they'll have names like fr.po and en_GB.po [16:25] but they end up in a directory that reveals the project name? [16:25] Another pattern is $PROJECT-$LANGUAGE.po [16:26] asac: no, you're supposed to know the project when you request the export. [16:26] asac: that request is local to a product or distroseries/package [16:26] ok ... so langpack magic requests export for each individual project? [16:28] asac: there's a separate way to do it for the entire distroseries, but I'm assuming you don't need that here. [16:29] asac: if you do, well, we do generate them. [18:32] hi [18:34] hi [18:55] Ubulette: they applied the pyxpcom build patch. good. but we still see python startup issues, don't we? [18:55] yes [18:56] damn, i keep forgetting to push my commits [18:56] did that one yesterday [19:05] fail [19:05] btw tb is out [19:08] asac already did it [19:09] i wanted to do it but he beat me to it ;) [19:09] !info thunderbird hardy [19:09] thunderbird (source: thunderbird): mail/news client with RSS and integrated spam filter support. In component main, is optional. Version 2.0.0.12+nobinonly-0ubuntu1 (hardy), package size 10703 kB, installed size 32020 kB [19:10] [reed], lots of commits today, i thought it was code freeze since yesterday ?!?!? [19:11] <[reed]> lots? I only see 4 [19:11] <[reed]> all commits after freeze require explicit driver approval beyond what is normally required [19:11] <[reed]> mostly regression fixes [19:12] <[reed]> but some large patches that didn't land yet [19:12] <[reed]> so, freeze was extended to 3am PST [19:12] <[reed]> anything after that needs approval [19:12] <[reed]> does that help? :) [19:12] oh, maybe the rush was just before that [19:12] <[reed]> yeah [19:13] <[reed]> everybody was rushing to get in [19:38] bug 196230 [19:38] Launchpad bug 196230 in firefox-3.0 "crash on youtube" [Undecided,New] https://launchpad.net/bugs/196230 [19:39] [reed]: did you get my answer from yesterday? [19:39] s/answer/question/ [19:39] :) [19:40] Ubulette: python invalid-crash.py 196230 [19:40] ? [19:41] Ubulette: http://people.ubuntu.com/~asac/invalid-crash.py [19:41] replace my name with yours [19:41] Ubulette: oh i ran it :) [19:48] asac, btw, at work, i don't have sound in ff3 (flash) if totem is running [19:48] ALSA lib pcm_dmix.c:866:(snd_pcm_dmix_open) unable to open slave [19:48] ALSA lib pcm_dmix.c:866:(snd_pcm_dmix_open) unable to open slave [19:49] tried to tweak FIREFOX_DSP env setting? [19:49] nope, but this is often reported in the forums too [19:49] maybe we should do something about that [19:51] hmm. its most likely a flash issue we cannot really cope with. they should use gstreamer, but they probably wont, so we can only use sound server, but that is known to be buggy as well [19:51] is that a regression at all? [19:51] or rather a long standing issue? [19:56] started last week for me [19:56] or maybe a bit more [20:03] hmm ... remember what upgrade you got at that time? [20:03] kernel? [20:03] donno, about a hundred per day [20:03] <[reed]> asac: check #extdev on moznet [20:04] <[reed]> they can more easily answer than than I can :) [20:05] asac, http://ubuntuforums.org/showthread.php?t=709111 look at comment #3 [20:05] asac, http://ubuntuforums.org/showthread.php?t=709280 [20:06] http://ubuntuforums.org/showthread.php?t=709203 [20:06] http://ubuntuforums.org/showthread.php?t=708907 [20:07] ok, not the same as me [20:07] http://ubuntuforums.org/showthread.php?t=708928 [20:07] seems to be fixed now [20:07] bug 196147 [20:07] Launchpad bug 196147 in python2.5 "ImportError: No module named _bsddb" [Undecided,Fix released] https://launchpad.net/bugs/196147 [20:07] (the miro issue) [20:08] http://ubuntuforums.org/showthread.php?t=706788 [20:08] http://ubuntuforums.org/showthread.php?t=707488 [20:08] the .cfg file issue we should review [20:09] we should consider to put firefix.cfg to /etc/ hierarchy again ... and don't encrypt it [20:09] http://ubuntuforums.org/showthread.php?t=707823 [20:09] and provide a grepref.js in /etc as well [20:11] yeah i noticed the performance issues with XAA [20:12] no idea if we can fix that [20:16] no idea about the compiz issue [20:16] commented on the other threads [20:30] moz cvs is soooooo slow tonight [20:30] half an hour to fetch xul [20:42] asac, did you use the bzr branch to do tb2 today ? [20:43] i don't understand what happened to https://code.edge.launchpad.net/~mozillateam/thunderbird/thunderbird.dev [20:44] Ubulette: yes https://code.edge.launchpad.net/~asac/thunderbird/ubuntu-2.0.0.x [20:45] Ubulette: yes, i rolled back the last revision [20:45] its still here on my disc [20:46] i received bad feedback from debian about that change [20:46] compare with http://paste.ubuntu.com/5055/ [20:46] where are my commits ? [20:46] seems you own everything now [20:48] of course i've diverged [20:55] Ubulette: they are in sync till rev 60 afaics [20:55] 61 will be merged on top once i figured out why debian had so much issues with it [20:57] unless i'm blind, you redid each of my commits manually [20:57] iirc, that was long ago and we already argued about that incident [20:58] really ? [20:59] maybe, we argue a lot :) [20:59] yes for sure. [21:02] can you then update the mt branch so I can trash/resync mine [21:03] lets see [21:11] Ubulette: ok i pushed the current dev to a feature branch to keep it and popped top most revision to get back in sync with my branch [21:11] https://code.edge.launchpad.net/~asac/thunderbird/thunderbird.dev.gnome-autodetect [21:13] ok used --remember so i don't get back to my branch [21:13] and changed branch detail for my branch to "Abandoned" [21:13] thx [21:21] oh i pushed to ~asac :) [21:22] you have to --remember for each command, push, pull, merge [21:23] i did ;) [21:23] now all is fine [21:33] i sometimes have font size issues in ff3 (not new). it seems

are sometimes rendered much bigger than they should be, a reload always fixes it [21:33] did you ever get that? [21:34] happens a lot on lp [21:36] no never [21:36] can you reproduce? [21:36] maybe that happens if the .css file fails to load for some reason? [21:37] does it always happen on the same server? [21:38] i happens on the 1st load only so it's rather random [21:38] it [21:38] yes, sounds a bit like a server load issue ... or race [21:39] if you first visit a page, ffox loads every .css + .js ... otherwise just those not yet loaded [21:39] so on reload you might have luck to get the initially missing bits [21:40] jetsaredim: i commentred on the mozgest bug [21:41] yes - i see that [21:41] i'm uploading a change that will exclude the debian dir [21:42] great [21:43] asac, where are user data stored in tb2 ? i didn't change anything in tb3 and a user asked me how to migrate data and rules from tb2 to tb3 ? (i've kept the moz profile name) [21:43] yeah ... thats a good coincident :) [21:44] previously it was .mozilla-thunderbird, but we want .thunderbird [21:44] in past it was difficult to migrate because of old and rotten profiles [21:44] but for 3.0 branch that should be ok [21:45] ? [21:46] what?? [21:46] obviously that user doesn't see his data so there's something wrong [21:47] yes, because it doesn't look at .thunderbird [21:47] would be beneficial to migrate that [21:47] migrate back to .mozilla-thunderbird ? [21:47] or keep using .mozilla-thunderbird for now [21:47] no ... migrate back to upstream dir: .thunderbird [21:48] but i didn't patch that [21:48] so i assume i have the same as upstream [21:49] yes. tb3 uses .thunderbird ... <3 used .mozilla-thunderbird [21:49] so migrate users when the run tb3 for first time from .mozilla-thunderbird to .thunderbird :) [21:49] got it? [21:50] yep [21:50] i thought tb2 was patched in ubuntu already, so i'm troubled [21:54] [reed], is there a list of tb3 changes since tb2 somewhere ? kind of changelog summary.. [21:55] <[reed]> not sure, but then again, I don't follow TB [21:55] (bonsai is not the answer i'm expecting) :P [21:55] i know, but you may have seen something [21:56] or know someone who does know better ;) [21:57] hm, japanese is no longer anti-aliased :( [21:58] i think there is a font issue atm [22:02] a few days ago the whole text was aa: http://www.sofaraway.org/ubuntu/tmp/japanese1.png now, kanji are no longer aa [22:04] Ubulette: if the problem persists for another day prod ArneGoetje in -devel i guess [22:04] waaaa, amazon is ugly now: http://www.sofaraway.org/ubuntu/tmp/japanese2.png [22:04] far too big [22:04] dpi? [22:05] no, everything else is okay [22:05] he is your guy then :) [22:05] i'm in 1440x900 [22:07] yeah looks broken [22:07] 89x76 dpi [22:07] 0 in about:config [22:08] k [22:13] looks fine in sm1 [22:15] http://www.sofaraway.org/ubuntu/tmp/japanese2-good.png [22:22] same for hardy release? [22:23] or just b4? [22:36] w8, i'm doing a big upgrade [22:36] but you can also try: http://www.amazon.co.jp/b/ref=topnav__gw?_encoding=UTF8&node=3210981 [22:41] for me kaze and ffox look equal ugly [22:42] (kaze uses xul 1.8) [22:42] no seamonkey here atm [22:42] midori (webkit) doesn't work at all [22:43] ok looks better in konqueror [22:43] wierd [22:43] cairo? [22:43] pango? [22:43] can you please start seamonkey with pango enabled? [22:43] MOZ_DISABLE_PANGO=0 seamonkey [22:43] MOZ_ENABLE_PANGO=0 seamonkey [22:43] no idea [22:45] ENABLE...=1 of course [22:46] tried both, no change [22:46] do we still have anything in common with debian iceape? [22:47] (well you know what i mean) [22:48] yes, most is still the same [22:48] for me sm looks as ugly as ff3 [22:48] most likely a pango thing then [22:48] hmmm the fonts on the left look a bit aa'ed [22:49] http://www.sofaraway.org/ubuntu/tmp/japanese2-epiphany.png [22:50] looks uglier here ... almost like ffox 3 [22:51] yes its pango [22:51] MOZ_DISABLE_PANGO=1 seamonkey [22:51] fixes it for me [22:54] Ubulette: see if older pango fixes you [22:54] for me, sm1 looks the same either ways [22:56] pango has been updated today [22:56] well, for me [22:56] previous was Fri Jan 25 [22:57] Feb 18 00:03:00 hmm, in b4pre, Japanese is now anti-aliased, fine :) [22:58] amazon is clearly not antialiases at all [22:58] sed [22:58] amazon looks the same with the version in hardy [22:59] I need to reboot (new kernel).. maybe that will fix this [23:00] brb [23:18] waaa, gnome/X frozen during startup [23:19] sounds like fun [23:19] still bad resolution, I have to kill X once [23:22] and I still have a blocking error popup for en_US not supported (got that for a while) [23:23] now ff3 is no longer starting [23:23] killall esd fixes it [23:23] like the gnome freeze [23:23] so it's a sound mess [23:27] oh, my japanese fonts in ff3 prefs were messed up [23:29] damn, ubufox killed my middle click again => uninstall [23:34] changed again today: http://www.sofaraway.org/ubuntu/tmp/ff3-b4-Add-ons.png [23:43] <[reed]> go Tango team ;) [23:44] Error: Components.classes['@mozilla.org/plugin/host;1'] is undefined [23:44] Source File: chrome://mozapps/content/extensions/extensions.js [23:44] Line: 847 [23:44] what does that mean ? [23:45] Error: uncaught exception: [Exception... "Component returned failure code: 0x804b000a [nsIIOService.newURI]" nsresult: "0x804b000a ()" location: "JS frame :: chrome://global/content/contentAreaUtils.js :: openURL :: line 894" data: no] [23:46] <[reed]> that sounds like you broke something [23:46] * [reed] blames ubufox [23:46] <[reed]> :) [23:48] that's tb3, I had to create the installer as there was only one for windows, maybe i've missed sometinh [23:48] g [23:48] http://bazaar.launchpad.net/~mozillateam/thunderbird/thunderbird-3.0.head/annotate/fta%40sofaraway.org-20080227211335-6otghfwqt3a7gj3f?file_id=fix_unix_installer.p-20080222125447-i8t6i8gsf4kohvcl-1 [23:49] (damn url) [23:49] you lack the plugin host component [23:54] http://paste.ubuntu.com/5062/ [23:54] I don't really understand the magic behind those .xpt files. [23:54] seems they are all merged into components/mail.xpt [23:56] and the other files are not installed whatever i do in packages-static [23:57] bin/components/libgkplugin.so is named twice in that file [23:59] oops