[00:01] <[reed]> Linux people will always hate huge packages [00:01] <[reed]> except the kernel [00:12] hmm [00:12] right he is in launchpad-dev team [00:22] commented [00:22] fta: ^ ;) [00:37] yep === micahg1 is now known as micahg === ubott2 is now known as ubottu [02:04] asac: xulrunner-1.9.1 1.9.1.5+nobinonly-0ubuntu0.9.10.1 and firefox-3.5 3.5.5+nobinonly-0ubuntu0.9.10.1 from the security ppa work fine from my testing (new 3.5.5 profile, existing 3.5.4 profile, existing 3.0.15 profile) [02:05] ugh, 3.5.5 is a update release, not a security release... [02:06] * micahg needs to add something to Lucid Ideas wiki page [02:07] dtchen: thx a bunch [02:07] micahg: yes. its a stability only release [02:07] asac: it gave me another topic for the wiki [02:08] :) [02:08] great [02:08] asac: when do we get to discuss the topics? [02:09] micahg: right. thats what we talk about for ages now [02:09] i started a wiki on PPA reorg [02:09] * asac searches wiki [02:09] https://wiki.ubuntu.com/MozillaTeam/PPAs [02:09] the pinning examples are a bit messed up [02:10] still have the feeling that the pinning will be a red herring for the transitional package use-case ... which is the only reason not to also split per-branch [02:11] asac: I don't use it like that [02:11] what do you mean? [02:12] I pin all PPAs at 450, if you install a version from a ppa, you get updates from the ppa until an archive version supercedes it [02:13] I pinned my moz-beta ppa at 475, so it's ahead of the dailies [02:14] so assume you have firefox-backports and firefox-daily ppas enabled [02:14] and want [02:14] a) firefox-3.5 from archive [02:14] b) firefox-3.6 from -backports [02:14] c) firefox-3.7 from -daily [02:14] how would you pin that? [02:14] and maybe [02:15] a) firefox-3.5 from archive [02:15] sounds like me :) [02:15] b) firefox-3.6 from -backports [02:15] c) firefox-3.7 from -backports (in theory) [02:15] micahg: can you post your pin files? [02:15] no [02:15] would be no ff3.5 [02:16] I would pin -backports higher than -dailies [02:17] so how do your current pins look like? [02:17] * micahg has too many [02:17] e.g. how do you do the per-version pinning [02:17] do you list all package names explicitly? [02:17] I'll pastebin the relavent ones [02:17] thx [02:18] http://pastebin.ubuntu.com/314734/ [02:18] ok [02:18] I have 3.7 from dailies, 3.6 from micahg-mozilla-beta, and FF35 from -security [02:18] thats just seleting based on PPAs [02:18] not by version [02:19] so if you want to run 3.5 from dailies but 3.6 from archive ... will not work [02:19] sure it will [02:19] how? [02:19] ppa is on demand if higher version [02:19] you install it, you get updates [02:19] you don't, you don't get it [02:19] http://pastebin.com/f5a3bcbb8 [02:21] well. that depends on a specific layout of what is where, right? [02:21] not really [02:21] if you assume that all packages are everywhere, etc. ... this wouldnt work, right? [02:21] as long as the ppa is higher than the archive, you can choose which one you want [02:21] like you have a firefox-3.7 in the archive, beta and daily ppa [02:21] how? [02:21] how do you choose that? [02:22] the user-experience i want is that user says: [02:22] "i want firefox-3.5 as dailies" [02:22] pin beta at 475 [02:22] daily will always be higher [02:22] "i want firefox-3.6 as backports/milestones" [02:22] apt will not make you install a lower version [02:23] then user runs apt-get dist-upgrade ... and all should be as he opted-in [02:23] dist-upgrade disables all ppas :) [02:23] no [02:23] not here ;) [02:23] did for me [02:23] well.. you know what i mean [02:23] :) [02:23] apt-get dist-upgrade? [02:23] that works for me [02:23] i always run that [02:24] you probably mean update-manager -d [02:24] oh, I guess I'm remembering from debian :) [02:24] you want it totally automatic [02:25] ugh, yeah, you'll probably have to pin individual packages then [02:25] what i want is to build a UI that allows users to just select what they want from where ... [02:25] which is a mess [02:25] that UI does not interfaces with apt/update-manager etc. to reduce complexity [02:25] just adds the PPAs and the pinning [02:25] once you install a version higher than the archives, you'll get updates [02:25] that would be the perfect user experience i would hope for :;) [02:25] mozilla-configurator? [02:25] micahg: right. but then in case archive gets a higher version for a minute you bounce back to archive [02:25] right [02:25] or daily-tester-tools [02:25] or something like that [02:26] well, if we're going to do that, then we can certainly pin higher and the individual pacakges [02:27] like you have on the wiki [02:27] yes. but we need working examples [02:27] for the two cases i mentioned (not sure what the UI would use, but probably the second): [02:27] 1. For users that want to run dailies or backports for just a specific version, we suggest lowering the priority of the repository using apt pinning and then increasing the score for the app that is wanted. [02:27] 2. To opt-in only for specific application versions we provide a wizard script puts the proper pinning in place. [02:28] sorry for the confusion ;) [02:28] thats just one approach. i think i removed the second way of approaching this ;) [02:28] from the wiki [02:28] i'm confused... can someone please direct me towards the firegpg team? [02:28] I think all ppas should be < 500 by default [02:29] maybe that should go in lucid... [02:29] but does that matter for the mozilla configurator thing? [02:29] i would think thats independent. [02:29] no, we can handle it independently [02:30] right [02:30] something that might be worth looking into for the whole distro [02:30] Out_Cold: in Ubuntu or in general? [02:30] we should discuss with mvo [02:30] asac: yeah, that's what I'm thinking [02:30] micahg: also i hoped by using a low prio like 10 that packages might even get downgraded if user opt-out later on [02:31] Out_Cold: http://getfiregpg.org/s/help [02:31] well i'm using ubuntu so it may be related to that [02:31] the main problem is that we seem to need to add explicit entry for _all_ binary packages for a source [02:31] Out_Cold: we don't package it yet, so I suggest following the link I posted [02:31] like user says: "firefox-3.5 from beta" -> we need to somehow get all package names and put a pin with prio like in the wiki or something [02:31] thanks [02:31] micahg: gnomefreak made a package [02:32] it should get uploaded for lucid finally [02:32] it had license issues in the past [02:32] i /think/ those are now fixed [02:33] micahg: what i mean is that "Package: firefox-3.6* abrowser-3.6* xulrunner-1.9.2* [02:33] " [02:33] does not work i found out [02:33] asac: well, we can code the package names into the app [02:33] we need one entry for each binary package [02:33] micahg: yeah. but thats also maintenance pain ;) ... would prefer if it was automatic somewhat ... [02:33] that leads to another question: how to keep track in case more packages pop up etc. [02:34] don't one of the files in the bzr archive have the package names? [02:34] control [02:34] debian/control [02:34] can we parse that? [02:35] we probably could. but when? when user enables/disables it, that might be simple ... but what if suddenly there is one more package avail. [02:35] we can also check the package DB ... [02:35] thats not that hard [02:35] asac: that would be after they enabled though [02:36] yeah. so probably needs to be just implemented and then tested [02:36] we'd have to add the source, pin, update again [02:36] micahg: not necessarily [02:36] sorry, add source, update, pin, update [02:36] micahg: in ubufox i use python apt lib to do our own package db tempdir [02:36] asac: a regular user wouldn't have any package info on 3.7 without a daily [02:37] you can do that for arbitrary sources.list snippets that would be shipped in the configurator app [02:37] bzr branch lp:ubufox [02:37] cd ubufox/pfsdb/ [02:37] ah, we could build a temporary apt db, pin, update [02:37] ok [02:37] pfs/db/ [02:38] nppapt.py plugindb.py recreatedb.sh sources.list sources.list.7.10 sources.list.8.04 sources.list.8.10 sources.list.9.04 sources.list.9.10 [02:38] micahg: yes. still we would need to react on when new package appear [02:39] point is that if we start to have our own apt db ... [02:39] we will be out of sync of what the real apt db knows [02:39] so we could do our own db in mozilla configurator ... and also add a trigger maybe [02:39] apt trigger [02:39] not sure how ;) [02:39] yeah [02:40] have to check with mvo ... [02:40] asac: etckeeper has an apt trigger [02:40] http://pastebin.com/f211b5517 [02:41] in anycase ... before we start putting work into something like that we should be sure that we have the right generic pinning approaches [02:41] strategies [02:41] and that they work as expected [02:41] otherwise we do lots of work and then find that the pinning is a mess ;) [02:42] and that we need to split ppa per version too [02:42] asac: PPA per version seems crazy [02:42] but in that case cant do transitional packages etc again ... but here again i am not sure about how the transition would behave [02:43] micahg: the only technica reason (besides the guts feeling of bloet) to not have ppa per version is that we need to do transitions like i described [02:43] in wiki [02:43] for technical reasons we don't have per-version daily repositories as this would cause problems when shipping partial transitions such as moving the firefox meta package from 3.5 to 3.6 package [02:43] https://wiki.ubuntu.com/MozillaTeam/PPAs#Using%20PPAs%20for%20just%20specific%20application%20versions [02:44] maybe it would be easier to just implement the package* syntax in apt for pinning [02:44] oh, I wanted to ask, why are we taking over default browser in 8.04 - 9.04 dailies? [02:44] in lucid [02:44] (doesnt help for all the older releases though) [02:45] micahg: you mean: why we transitioned to 3.5? [02:45] thats because dailies are double dailies for us [02:45] in the dailies for hardy - jaunty [02:45] otherwise things would become realy complex [02:45] we would need -daily-stable ... -daily-backports ;) [02:45] so antoher ppa set [02:46] like daily, but using the release branches [02:46] firefox-3.5.karmic ... firefox-3.0.hardy etc. [02:46] and daily, but using .head branch [02:46] the latter is what we define as dailies atm [02:47] which automatically comes with the take over you asked for [02:47] and also with the semi transition that is the technical blocker for splitting ppas per version [02:47] 03:43 < asac> for technical reasons we don't have per-version daily repositories as this would cause problems when shipping partial transitions such as moving the firefox meta package from 3.5 to 3.6 package [02:48] asac: you're up late :) [02:48] yes [02:48] too much holiday -> too much stuff to do this week [02:48] next week is UDS etc. [02:48] well, it would seem like a lot of work to maintain per version ppas [02:49] thats the same work imo ... fta would just upload them to the right repo [02:49] we can also automate the milestone spinnings i hope in the same way [02:49] just a manual push i guess rather than something that riuns fully automated [02:51] asac: is ff3.5.5 going through -proposed? [02:51] no [02:51] its a regression release [02:51] for the previous security update [02:52] and just has a hand ful of changes [02:52] like a minibranch [02:52] -updates? [02:52] it goes to -security ... and we always copy that to -updates too to help mirroring security updates [02:52] same procedure as security updates basically [02:52] just a regression fix for a security udpate [02:53] (thats why i uploaded to the security ppa) [02:54] asac: you don't seem to make new snapshot commits, am I supposed to do that? [02:54] i dont? [02:54] oh [02:54] i forgot ... my bad then [02:54] sometimes [02:54] i usually do [02:54] i try to do [02:54] like today [02:55] so i should at least have bumped the changelog version [02:55] ugh, TB3 cairo bumped to 1.8.8 [02:55] and include that its a new snapshot in the same commite message [02:55] did i do that at least? [02:56] i think i ended up doing that in one commit if its just one change for that snapshot ... but splitting up in pure snapshot bump and multiple commits [02:56] yes [02:56] if there is more than one thing required [02:56] i think both is ok [02:57] ok i failed here [02:57] let me uncommit :/ [02:57] ugh [02:58] oh shit. i think that was really a bad idea because fta already did a new run :/ [02:58] * asac feels dirty [02:58] * asac hides [02:59] I didn't see a new run since you fixed that [02:59] good [02:59] asac: you're safe [02:59] then i might be lucky :-P [02:59] he reran after I committed ff3.7 [03:00] did you do overwrite? [03:00] yes. [03:00] i have a bound branch [03:00] that automatically applies it [03:00] thats why i couldnt stop it after bzr uncommit ;) [03:00] it just did it [03:00] oops [03:00] wrong version :) [03:01] New upstream snapshot: 1.9.0.6 HG 20091109r26560 [03:01] help me ;) [03:01] I think only sleep can help at this point ;) [03:01] ok i think its now ok ;) [03:02] and yes. i should not touch branches after 4am ;) [03:02] thats usually a bad idea [03:02] ok i nthat sense out ... hope i didnt bust something else in that last commit ;) [03:02] ok, I'll bump libcairo for TB3 later otnight [03:03] yeah go ahead [03:03] ttyt [03:03] ok === ]reed[ is now known as [reed] [14:25] bdrung: also: http://www.asoftsite.org/s9y/archives/152-ubuntu-mozillateam-minutes-14th-sep-08-....html#c748 [14:27] k [14:28] bdrung: so the other problem is that we cannot really share anything with debian because of trademark implications. so the only practical way would be for debian folks to join ubuntu mozillateam (e.g. flip it upside down) [14:29] which would make much more sense [14:29] especially because we have all the users and get real feedback on QA etc. [14:30] asac: the feadback depends on the package. some packages get more feedback in debian. [14:33] yes. but for mozillas [14:33] i am only speeking on mozilla stuff ;) [14:34] yeah, that true. simply compare the popcon stats [15:00] kenvandine: not online with jabber? [15:00] asac, i am [15:00] hmmm ... do you still have asac@jabber.ccc.de in there? [16:39] hello, my lizard friends [16:40] I am attempting to use xulrunner 1.9.1 from here: https://launchpad.net/~ubuntu-mozilla-daily/+archive/ppa [16:40] it looks like it needs a newer nspr, which is not in that ppa [16:40] a little help? :) [16:40] snorp: which package? [16:41] micahg: it's not a package dep issue (more of a missing dep) [16:41] or rather which releasE? [16:41] but libxulrunner-unstable pkg-config require mozilla-nspr [16:41] which I can't fine [16:41] are you trying to build it? [16:42] no [16:42] trying to build against it [16:42] which release are you on? [16:42] err, libxul-unstable.pc I mean [16:42] hardy [16:42] ah [16:42] s/fine/find/ :) [16:43] snorp: xulrunner-1.9.1 is not meant to be built against in hardy [16:43] oh :( [16:43] that makes me immeasurably sad [16:43] we don't officially support it at this time [16:43] we're working on backports for mozilla software right now [16:43] right, I understand [16:45] snorp: is there a reason you're using 1.9.1 instead of 1.9.0? [16:45] or rather 1.9 [16:45] micahg: just want to move to 1.9.1 for some things [16:45] I work for a partner of canonical [16:45] they are going to get us "official" packages [16:45] asac: ^^^ [16:45] but just trying to get something going in the mean time [16:45] yeah, if asac could help that would be good :) [16:45] pbuilder? [16:45] micahg: any news on that nvidia/firefox issue? [16:46] yoasif: I've seen more bugs :) [16:46] no, unfortunately not [16:46] is 3.5.5 out yet? [16:46] snorp: afaik we should have nspr 4.7.5 in all ubuntu releases ... isnt that the case? [16:46] no master bug? [16:46] !info nspr [16:46] Package nspr does not exist in karmic [16:46] i just got 3.5.4 from lucid repos [16:46] !info libnspr1 [16:46] Package libnspr1 does not exist in karmic [16:46] !info libnspr1d [16:46] Package libnspr1d does not exist in karmic [16:47] asac: yeah, 4.7.5 is [16:47] https://edge.launchpad.net/ubuntu/+source/nspr [16:47] snorp: ^^ [16:47] so maybe you just need to enable -security -updates [16:47] to get 4.7.5 [16:47] that should be recent enough for 1.9.1 [16:47] asac: win. [16:48] asac: it is requiring 4.8.2 [16:48] ugh, we don't have 4.8.2 in anything [16:48] crap [16:48] err [16:48] no [16:48] snorp: are you sure it's xulrunner that requires that [16:48] i dont think that 1.9.1 really needs 4.8.2 ... maybe latest dailies [16:48] but i havent seen that [16:48] asac: the latest in karmic is 4.8 [16:48] Requires: mozilla-nspr >= 4.8.2 [16:49] from libxul-unstable.pc [16:49] yeah, this is dailies [16:49] is there another source for 1.9.1 I could use? [16:49] debian has 4.8.2 [16:51] i am in meeting ... will check with you ater that [16:51] cool [16:56] snorp: I don't see anything that would require that [16:56] so maybe I can just hack it to use a lower version [16:56] do you have nspr 4.8.2? [16:56] nope === micahg1 is now known as micahg [17:49] asac: fta: sparc workaround latest comment at mozilla bug 448658 [17:49] Mozilla bug 448658 in Phishing Protection "nsUrlClassifierDBService has bad alignment, causes SIGBUS" [Critical,New] http://bugzilla.mozilla.org/show_bug.cgi?id=448658 [17:50] can i ask a random question guys as im seeing armin76 mention sparc. how does one test stuff for other processor architectures if they dont ahve the hardware [17:51] eagles0513875, they use a qemubuilder for building purposes [17:51] but for testing is har [17:51] * hard [17:51] gotcha [17:52] also setting up a qemubuilder is not that easy [17:52] there is no emulated environment or anything of the sort? [17:52] qemubuilder is an emulated environment [17:52] it uses particular kernel images to build the chroot [17:52] and then you can run some archs like sparc or amd64 [17:52] if you are on i386 [17:53] ahhhh [17:53] or i386 if you are into amd64, sparc whatever [17:53] but sometimes it's broken [17:53] ahhh ok [17:53] so ... [17:53] im gonna update my laptop to lucid and begin helping u guys out [17:54] cool [17:55] eagles0513875: you don't have to be on lucid to package for lucid [17:55] micahg: working on extensions for ff [17:55] right [17:55] as long as you have the latest m-devscripts, you should be ok [17:55] are you talking about working in a chroot environment? [17:56] well, pbuilder can build the lucid package for you [17:56] pbuilder-dist in ubuntu-dev-tolls [17:56] *ubuntu-dev-tools [17:56] ahhh ok [17:56] im a novice to all this [17:56] that's why I don't think you should move to lucid [17:57] unless you are willing to do full system recovery if something breaks [17:57] if need be i am willing to reinstall [17:57] or recover [17:58] latest mozilla-devscripts are in the daily PPA [17:58] nice thing for me since im duel booting on boot camp partition on my mac is all i need to do is boot onto the live cd and reinstall grub doesnt complain about it and osx boots just fine [17:58] eagles0513875: you could run virtualbox also === micahg1 is now known as micahg [17:59] if it's an intel mac [17:59] which it has to be I guess [17:59] micahg: i have noticed though that bugs might not surface in a vm like they would in a physical machine [17:59] well, hw bugs wouldn't [17:59] ok, gtg [18:00] bbiab === asac__ is now known as asac [18:09] * eagles0513875 waves to asac [18:45] hi eagles0513875 [18:45] snorp: so what are you working on? karmic? [18:46] asac: hardy [18:48] snorp: ok. for that you definitly need to pull in a nspr on your own if you want to build using system-nspr [18:48] i would think that the karmic package should be enough [18:49] ok, I'll give that a shot [18:50] snorp: so basically to build the karmic packages for xulrunner-1.9.1 firefox-3.5 and nspr and nss [18:50] s/to // [18:50] nod [18:51] asac: membership meeting is Wed Nov 18 at 2100 UTC [18:52] ok ... what time is that in dallas? [18:54] 3PM :) [18:54] I'll be several hundred miles north of you ;) [19:04] asac: should I comment on the blueprint? [21:01] anyone using ff3.7 here? [21:02] Error: uncaught exception: [Exception... "Component returned failure code: 0x80520012 (NS_ERROR_FILE_NOT_FOUND) [nsIXPCComponents_Utils.import]" nsresult: "0x80520012 (NS_ERROR_FILE_NOT_FOUND)" location: "JS frame :: chrome://browser/content/browser.js :: delayedStartup :: line 3590" data: no] [21:03] asac, [reed]: ^^ [21:03] <[reed]> on startup or what? [21:03] yes [21:04] <[reed]> guess it can't find the new components.list file? [21:04] i'm trying to find why my session is no longer restored [21:04] that's the only error in the console [21:05] <[reed]> do you have components/components.list [21:05] <[reed]> in your firefox directory? [21:05] <[reed]> should be in the same directory compreg.dat is [21:07] <[reed]> fta: ^ [21:08] nope [21:10]