=== asac_ is now known as asac [09:24] morning chrisccoulson, the LP guys have been working on a fix for the bug that caused us trouble building the Lucid langpacks. They're about to land it, but to fix Lucid (and Maverick) translations in Launchpad, we need to upload a new en-US.xpi template for FF 3.6. May I ask you to generate one for me from the src package and put it somewhere where I can fetch it and upload it manually? [09:32] dpm - yeah, i can do that. but i think i have to do a full firefox build to do that [09:32] i'll have a look [09:34] thanks chrisccoulson. Otherwise, was not a 3.6.14 update going to happen soon? If it is soon enough, I could use the en-US.xpi file generated in that build [09:34] dpm - yeah there will be a 3.6.14 release soon, but that's already built though [09:36] chrisccoulson, if that build generated an en-US.xpi file, that'd be even better, as I could then just grab that from LP. Was the code to generate the en-US.xpi file enabled on that build? [09:38] dpm - it would have generated one, but it wouldn't have been imported by launchpad [09:38] so, i guess it's been discarded now [09:38] chrisccoulson, why wouldn't have been imported by lp? [09:39] dpm - it's built in the security PPA [09:39] oh, I see [09:39] chrisccoulson, hi, you've got a video for your ff breakpad thingy ;) [09:40] fta2 - yeah, i saw that [09:40] that's weird, i can't recreate that though :( [09:40] perhaps the default is different in 3.6 ;) [09:41] chrisccoulson, could you point me to the security PPA (is it public?) where it was built? Perhaps I can ask the LP people to help me get it from the lp librarian [09:41] i'll look at that today though [09:41] dpm - https://edge.launchpad.net/~ubuntu-mozilla-security/+archive/ppa [09:42] chrisccoulson, i'd like to get breakpad in chromium quickly as i can't cope with the crash bugs [09:42] fta2 - yeah, that would be good. have you got anywhere to upload the symbols from, or do you want me to do that? [09:44] thanks chrisccoulson [09:46] chrisccoulson, as i said in the mail, i'm able to generate the symbol file, but don't know what to do with it. the builders can't upload them [09:46] chrisccoulson, i can ship them in a deb, but something else would need to extract the deb and upload [10:04] fta2 - yeah, that's what we do for firefox. we stick the symbols in to a firefox-mozsymbols package, and then i just grab that in our datacenter, unpack it and upload it [10:18] chrisccoulson, could that be done for all the ch debs? (i mean, the whole bunch) [10:18] fta2 - yeah, that shouldn't be a problem [10:18] cool [10:19] i'll go for a -breakpadsymbols deb then [10:19] should be 17MB each [10:19] 90MB uncompressed [10:38] chrisccoulson, how does it work? (how do you detect that you need to pull & extract the deb?) [10:39] fta2 - i just have a cron job that runs every hour (at the same rate as the publisher) that grabs the package lists for the archives [10:39] and it just parses them to get the new versions and compare those to what has already been uploaded [10:42] ok, good enough [10:42] you're doing it for the ff dailies too? [10:43] fta2 - yeah, i do it for the dailies as well [10:43] and also for thunderbird as well [13:22] chrisccoulson, firefox-4.0-4.0b12pre.en-US.linux-x86_64.crashreporter-symbols.zip where does that name come from? [13:23] fta2 - that filename comes from the mozilla build system (ie, i'm not setting it manually) [13:24] chrisccoulson, hm, in my case, it's just chrome.breakpad.x64 [13:24] i guess i should include the version & branch [13:24] that's probably ok. it gets unpacked on the server anyway [13:24] what do the filenames look like inside? [13:25] i don't remember, i have to rebuild :P [13:25] heh :) [13:25] there is also a manifest file in the firefox one too, but i'm not sure if that is a breakpad feature, or whether it's just added by mozilla [13:25] i know that the mozilla guys are using the manifest file for periodic cleaning of the symbol files [13:34] i just have a text file, no manifest === m_conley_away is now known as m_conley [14:40] fta2 - you're on nvidia aren't you? [14:40] did you see the conversation in #ubuntu-desktop? [14:58] ? [15:01] fta2 - new nvidia driver in xorg-edgers [15:01] (not sure if it's published juet yet though) [15:02] \o/ [15:03] brb [16:05] -rw-r--r-- 1 1234 1234 12M 2011-02-15 16:22 chromium-browser-breakpadsymbols_11.0.672.0~svn20110215r74909-0ubuntu1_amd64.deb [16:05] -rw-r--r-- root/root 92220948 2011-02-15 16:08 ./usr/lib/chromium-browser/chrome.breakpad.x64 [16:05] chrisccoulson, ^^ [16:07] MODULE Linux x86_64 97A80197B0F05649C6767D3FEA11A3770 chrome [16:07] FILE 0 ../css/tokenizer.flex [16:07] FILE 1 ./app/gfx/gl/gl_context.h [16:07] FILE 2 ./app/gfx/gl/gl_context_stub.h [16:07] FILE 3 ./app/sql/connection.h [16:07] chrisccoulson, ^^ [16:07] that's the file untouched [16:07] fta - thanks [16:08] i still need to tune a few things, like the project id [16:12] hm, not sure what this 97A8.. is about. if it's the md5 of the unstripped binary, it won't match [16:14] fta - it's not. i think the number is also embedded somewhere in the binary so that the server can match the binaries up to the symbols [16:14] it works for firefox, and we strip the binaries after generating the symbols too [16:15] ok, good [16:15] i will read the code though, i'm expecting some more magic is needed [16:16] that still won't solve the EULA issue [16:16] not sure what we should do here... [17:31] Is it just luck that my Minefield didn't crash once in months? :o [17:31] And is there any chance you will add PGO builds? [17:33] Also I heard that Firefox may include GStreamer as a backend on Linux, is this true? H264 could work then too, I suppose? [17:33] phsi: yes, gstreamer in Firefox has been a work in progress for a few years [17:52] Will chromium-codecs-ffmpeg-nonfree/extra continue to include H264? [17:52] phsi: no, not once it's dropped upstream [17:53] It wouldn't be too hard to get stuff back in though? If even manually [17:54] phsi: I don't know about that, it's open source, so in theory should be possible [17:54] Yeah, well. The buildscripts look a bit funky but it should be fine I suppose.. [18:22] chrisccoulson, is firefox using breakpad on arm? [18:23] fta - yeah. it's only enabled on i386, amd64 and arm [18:28] good [18:29] fta - i don't think breakpad even compiles on the other architectures [18:29] and in any case, google probably don't care about those ;) [18:33] sure, i was just curious about arm [18:33] i didn't find any arch restriction in the rules file for breakpad [18:45] Will there ever be PGO builds in the PPA? [18:48] Hi, I know this is a ubuntu team, but I guess you might know the answer even for windows. My question is about the notification bar in firefox or thunderbird. How can I change its color? [19:07] ich brauche hilfe [20:10] /usr/include/xulrunner-2.0b12pre/jspubtd.h:105: fatal error: jsproto.tbl: No such file or directory [20:10] anybody know why that's failing? [20:11] sounds like you packaged it incorrectly [20:12] * magcius pings chrisccoulson, micahg [20:12] yes? [20:13] did you forget to package jsproto.tbl with the headers? [20:13] possibly [20:13] I can go fetch it myself, but just wanted to make you aware. [20:14] jspubtd.h includes it, and jsapi.h includes jspubtd.h [20:15] yeah, i made some changes in the packaging last week, that must have just dropped off [20:15] thanks [20:17] magcius, http://bazaar.launchpad.net/~mozillateam/xulrunner/xulrunner-2.0.head/revision/645 [20:18] that will be picked up by the next nightly [20:21] Thanks. [20:27] chrisccoulson, you also need *.msg for js.msg [20:27] heh, thanks [20:28] after that it works [20:29] ok, pushed a fix for that too [20:29] Thanks. [20:29] I assume none of these things matter for firefox-4.0-trunk or the like? [20:35] magcius, they will do [20:36] but, we don't ship a shared libmozjs in firefox, so there shouldn't be anything including the jsapi headers really [20:36] unless they're crazy ;) [20:36] i'll copy the same fixes across for the firefox-dev package though === Nafallo_ is now known as Nafallo [21:02] fta: hey any clue how/if a11y is coming along in chromium? [21:04] jcastro, i didn't hear anything about it, but i didn't search for it either [21:04] jcastro: wouldn't that depend on the support in Webkit for it in part? [21:04] yes, that's part of the answer [21:05] fighting with breakpad right now, i want the crashers upstreamed asap [21:05] well, I'm just wondering if it's worth pursuing chromium-by-default at some point [21:05] but if it's as complicated as whenever we talked about it last then ... [21:06] jcastro, don't get me wrong, i just don't know. maybe it's moving, maybe not [21:06] i moved to unity yesterday, i'm lost [21:07] can you ask around? no rush or anything [21:07] heh [21:07] filed 10 bugs already [21:07] I saw your app mode bug [21:08] jason tells me it can be fixed, so I am hoping for a fix soon [21:08] doh, found another bug [21:08] drag a desktop file from nautilus to the side bar, get an empty black slot [21:10] what's the equivalent of the old applications menu? [21:12] clicking the application button on the bottom left [21:12] but that doesn't work sometimes [21:12] ohh [21:13] thanks [21:13] jcastro, http://code.google.com/p/chromium/issues/detail?id=28286 one bit fixed [21:13] http://code.google.com/p/chromium/issues/detail?id=27112 not fixed [21:14] http://code.google.com/p/chromium/issues/detail?id=24585 some clues in the last comment [21:16] https://bugs.webkit.org/show_bug.cgi?id=25531 ohh, starting to look good [21:16] bugs.webkit.org bug 25531 in Accessibility "[GTK] Metabug: Bugs blocking Orca support" [Normal,Unconfirmed] === Moot2 is now known as MootBot [22:07] jcastro, will there be an indicator with the world clock like before? [22:08] yeah someone's working on that [22:08] I don't know the status offhand [22:10] chrisccoulson: feel free to take any of the xulrunner porting stuff [22:12] jcastro, also the weather indicator is kind of weak, mph, no km/h choice, hard to read (dark on dark),.. [22:14] saw some good mokups though [22:17] fta, do i blame firefox or chrome/ium for the latter not actually importing my settings when it says it does? [22:20] LLStarks, ch just imports from ~/.mozilla/firefox/ so if your ff profile is called something else, it won't work === m_conley is now known as m_conley_away [22:21] it is firefox [22:21] grrr, with unity (or is compiz), xchat doesn't beep when asked to [22:22] +it [22:22] anyway to do import manually? [22:22] LLStarks, the b-m, yes, you have an import entry in the b-m manager [22:23] b-m? [22:23] bookmarks [22:23] you can ask ff to export, and ch to import [22:24] firefox can export? [22:25] i really don't care about bookmarks, i want chromium to have a fully saturated awesomebar that has browsing history parity with ff [22:43] fta, can chromium import addressbar/searchbar data? [22:44] i don't know. it has its own stuff already. right click the bar [23:01] god, chrome is refusing to import anything. firefox is closed and it procedes like it hsould. [23:01] i dun get it [23:02] oh i know, is it importing the profile i don't use at all? [23:03] still nothing... does chrome hate ff4? [23:06] very nice... i have to downpin to umd maverick to do this. [23:31] okay, i'm at a loss. how do you set up a firefox 3.6 environment on natty that uses a 4.0 profile [23:33] we don't support firefox 3.6 on natty, also, I don't know if the profiles are backwards compatible yet [23:33] the daily PPA should have a firefox 3.6 build, you could pin that higher than the archive [23:37] chrisccoulson, doesn't gtkmozembed use firefox rather than xulrunner? [23:38] Oh, who do I blame for "Restart Firefox" not restarting firefox in the breakpad crash reporter? [23:48] magcius: gtkmozembed uses xulrunner [23:48] hm, ok. [23:49] I'm not sure I understand what the purpose of firefox having its own copy of the libraries then [23:49] static linking causes it to load faster or something? [23:49] magcius: to be closer to the upstream build [23:50] FF on XR isn't tested [23:50] what is upstream's build rationalization? [23:50] they ship it as one unit [23:50] mozilla-central is where both "abrowser" (Firefox) and xulrunner get their code from? [23:51] yes, but they don't ship Firefox on top of xulrunner, they just ship firefox [23:53] OK. [23:54] the difference is that Firefox just has all the libs statically linked, and they use a binary entry point instead of a .js/.xul one? [23:56] the issue with using a system xulrunner is we don't want the pain of updating it in a stable release every few months, potentially breaking everything that depends on it [23:56] in addition to that, as micahg pointed out, it's unsupported [23:56] and there are startup time improvements that depend on not using the system xulrunner too [23:57] i'm not sure what you mean about binary versus js/xul entry point. they both start in the same way