[04:41] asac: im going to bed finally its almost 11 :( firefox-trunk we need to change the page it opens by default its a 404 file not found link it opens is http://www.mozilla.org/projects/firefox/3.0a5pre/whatsnew/ but its low on priority list as far as i am concerned but just letting you know before you upload to universe in gutsy. === tonyy [n=anthony@ubuntu/member/tonyyarusso] has joined #ubuntu-mozillateam === asac_ [n=asac@debian/developer/asac] has joined #ubuntu-mozillateam === gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-mozillateam [10:26] hey all [10:27] morning [10:35] morning (which is still somehow valid here) :) [10:45] its early === gnomefreak was thinking about dapper and firefox. is it possible to wait till next point release for dapper and sneak it into main (2.0.0.x) [10:46] if point release matches around the same time as EOS [11:34] no idea yet [11:34] will have to talk to other distributors [11:34] is the dapper rebuild done? [11:35] (btw, we have about 2 month from now) [11:35] to figure out if we update to 1.5.0.13d [11:38] asac: everything but OO.o afaik [11:39] brb have to reboot === gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-mozillateam === Starting logfile irclogs/ubuntu-mozillateam.log === ubuntulog [i=ubuntulo@ubuntu/bot/ubuntulog] has joined #ubuntu-mozillateam === Topic for #ubuntu-mozillateam: Home of Ubuntu Mozilla Team - https://wiki.ubuntu.com/MozillaTeam | Bug Triagers please read: https://wiki.ubuntu.com/MozillaTeam/Bugs/ | Firefox trunk package source : https://code.launchpad.net/~asac/firefox/trunk | Mailing List: ubuntu-mozillateam@lists.ubuntu.com === Topic (#ubuntu-mozillateam): set by asac at Wed May 2 14:46:32 2007 [02:14] its gonna be one of those days isnt it :( [02:25] gnomefreak: ? [02:25] gnomefreak: if you refer to power of live ... yes I feel exhausted [02:25] X is broken see -devel. guy just told someone the only way to update repos is to restart :( [02:26] oh and i have to clean out gutsy today [02:36] X is broken? [02:36] you mean on -devel channel? [02:37] asac: X wont upgrade [02:37] we were talking about it in #ubuntu-devel [02:38] hmmm ... ingutsy? [02:38] then i don't care ;) [02:39] yes and you should care my gutsy goes down i lose all my work :( [02:39] its all good i can do it in tty just as well [02:40] Bug 96267 [02:40] Launchpad bug 96267 in firefox "MASTER gecko/epiphany crash [@nsBlockFrame::ReflowFloat] [@nsBlockReflowState::FlowAndPlaceFloat] [@nsBlockReflowState::AddFloat] " [High,Confirmed] https://launchpad.net/bugs/96267 [02:40] gnomefreak: do you see the "new" testcase [02:40] in epiphany? [02:40] ill look as soon as i get off phone [02:40] k [02:40] gnomefreak: i guess X will be fixed soon [02:40] you are probably not the only one hit by this [02:43] im not everyone is [02:44] you mean https://bugs.launchpad.net/ubuntu/+source/epiphany-browser/+bug/96267/comments/4 [02:44] Launchpad bug 96267 in firefox "MASTER gecko/epiphany crash [@nsBlockFrame::ReflowFloat] [@nsBlockReflowState::FlowAndPlaceFloat] [@nsBlockReflowState::AddFloat] " [High,Confirmed] [02:52] yes [02:55] ill see if i can reproduce it [02:56] damn ffox bughelper page is really huge [02:56] 160K [02:56] http://daniel.holba.ch/bugs/firefox.html [02:57] Admiral_Chicago: for those "this might be a duplicate of" ... can we add bug-links as well [02:57] Admiral_Chicago: there are currently only links to the backtraces [02:59] asac: it does not crash in feisty chroot [03:00] yes it did [03:00] hmmmmmmm thats odd [03:00] it crashed second time i did it [03:02] gnomefreak: so you can reproduce? [03:03] gnomefreak: can you try with ffox as well? [03:03] Floating point exception (core dumped) [03:03] maybe it needs multiple tries as well? [03:03] im about to try ffox [03:06] 3 times no crash on ffox 2.0.0.3+1-0ubuntu2 nor 2.0.0.3+2.ng-0mt.6 ill try 2.0.0.4 in a minute [03:07] The following packages have been kept back: firefox this is odd [03:16] installing atm ill be back [03:23] Admiral_Chicago: did totem clue files adapt our tag/state combination clues? [03:23] Admiral_Chicago: http://daniel.holba.ch/bugs/totem.html [03:42] asac: i cant get it to crash on ffox 2.0.0.4 but epiphany i was able to. i have a few things to take care of for a few minutes than i will comment on the bug [03:43] k [03:43] btw, the bug does not claim that ffox crashes [03:43] yes it does [03:43] look at the first comment [03:43] it does? [03:44] or 2nd === asac looking [03:44] me too as soon as it opens [03:45] i don't see that anyone claims that firefox crashes [03:45] look at the stack trace [03:45] rom /usr/lib/firefox/components/libgklayout.so [03:45] yeah... that is used by epiphany as well [03:45] oh ok [03:48] from what upstream gnome says its mozilla issue. [03:49] yeah [03:49] the probably use it wrongfully [03:49] upstream gnome never admits their own faults [03:49] i could crash epiph as well btw [03:50] i think they miss to cleanup post preview properly [03:55] not sure but i would push it to epiphany maintainers. we dont have enough crashes to worry about ;) [03:55] does mozilla even support it? [03:55] they dont list it as a package afaik [03:56] who? [03:56] if its a bug in mozilla code-base then they will care [03:56] ah ok [03:56] if its just a false assumption of gecko embed users ... then they won't [03:56] asac: than who does it fall on? [03:57] no idea [03:57] can i reject ffox? [03:57] btw if it was gecko wouldnt ff have same issue [03:58] unless we patched it already and epiphany never took our patch [03:58] maybe ... which is why i think that epiphany uses it wrongfully === gnomefreak gonna reject ffox task as it doesnt crash in ffox. [04:00] gnomefreak: keep it open [04:00] add epiphany-browser task [04:00] and add gnome upstream task for now [04:00] set firefox task to Confirmed/mt-eval [04:00] why are we keeping open? [04:02] because we don't know [04:02] :) [04:04] mt-eval is the name of the tag? [04:04] full name even [04:05] all fixed [04:12] brb i hope === gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-mozillateam [04:18] X works fine even though failed to update [04:24] hmm [04:24] but it should be fixed in a few hours anyway. he found problem and im assuming hes fixing it === gnomefreak going to lay down 4am came way too early this morning [04:35] yeah [04:35] sw [04:35] ty [05:02] anyone here sees memory leaks as well? [05:03] would be interesting if trunk builds work better [05:03] e.g. have less leaks [05:36] Good evening asac gnomefreak :) [05:36] ola hjmf [05:36] just for the record, about bug 96267 as you were talking [05:36] Launchpad bug 96267 in firefox "MASTER gecko/epiphany crash [@nsBlockFrame::ReflowFloat] [@nsBlockReflowState::FlowAndPlaceFloat] [@nsBlockReflowState::AddFloat] " [High,Confirmed] https://launchpad.net/bugs/96267 [05:36] yeah [05:36] this morning when I did the duplicate stuff I tried to find in mozilla's bugzilla an upstream match [05:36] I wasn't able. [05:36] the nearest thing so far I've found is bugzilla 37937 comment #12 [05:36] which incidentally is a ubuntu's guy, and that its backtrace matches ours 0-9 stacks but in his #14-#23 stacks. Too weak for anything :) [05:36] so I think too that is epiphany only [05:36] yeah [05:37] as i said ... epiphany have a hard job to understand how gecko engine is ment to be used [05:37] which is why its likely that they failed [05:37] bugzilla 37937 [05:37] thats an ancient bug [05:38] bugzilla bug 37937 [05:38] Launchpad bug 37937 in Ubuntu "Linux headers problem when update the kernel" [Medium,Rejected] https://launchpad.net/bugs/37937 [05:38] ?? [05:38] at that time ubuntu didn't even exist i guess [05:38] shit! vnc copy and paste [05:38] wait :) [05:38] bugzilla bug 379037 [05:39] mozilla bug 379037 [05:39] it is ;) [05:39] Mozilla bug 379037 in GFX: Gtk "Firefox crash on visiting wellsfargo.com [@ nsFontMetricsXft::CacheFontMetrics] " [Critical,Unconfirmed] http://bugzilla.mozilla.org/show_bug.cgi?id=379037 [05:39] yeah :) [05:39] probably don't related at all [05:40] is just that part of the stack posted on comment #12 matched the first 9 stacks in our bug [05:41] hmm [05:41] we don't crash in fonts? [05:43] no, no match in maloe for nsFontMetricsXft:.... [05:43] *malone [05:43] asac: I'm off now [05:44] yeah then its different crash [05:44] maybe I'll be back at night [05:44] ok cu [05:44] cy [05:44] cu! [05:44] :) [05:44] have fun! [06:26] asac: let me review the clue files now. [06:27] :) [06:27] Admiral_Chicago: please take a look what the torrent clue does ... if its not an exact copy of our "tag/state" clues then its fine and has been adapted by someone [06:28] asac: if people are using our work, then thats great [06:28] sure [06:28] however if they just copy and try to match mt-XXX tags :) then its stupid :) [06:29] yea, i'll review it. [06:30] Admiral_Chicago: ... can we get the bug url to the duplicate hints? [06:30] or is that not possible without fixing bughelper core code? [06:31] yes, i'm reviewing now. it can be done, but let me check somethings first [06:34] asac: this tags business is due to the firefox in totem [06:35] possibly a bug? [06:35] i'll talk to danel [06:35] oh [06:36] yeah :) [06:36] ask him [06:38] just did [07:40] i think http://pastebin.ca/537191 is the relevant part of the file no? [07:42] Admiral_Chicago: looking [07:42] Admiral_Chicago: i think it start there [07:42] Admiral_Chicago: yes [07:42] asac: i'll have to review it some more though just getting some ideas about the code in my mind still [07:42] Admiral_Chicago: you have to wade through cluefiles [07:43] aeh through .cloes [07:43] iterate through the .clues [07:43] and only use those that have inheritable="true" [07:43] inheritobj.clues ... this is probably an array [07:44] or something [07:44] i see how it's done... [07:44] good [07:44] inheritlist.update(self.get_inherited_clues(inheritobj.srcpkg, \ verbosity)) [07:44] that line.. [07:44] Admiral_Chicago: no ... the line above [07:45] applies the .clues ... afaict [07:45] you have to filter that array [07:45] and pass another array that only contains inhertiable clues [07:45] ah yea, i see it. the line i posted passes the inherited array to the running clue fiel [07:45] inheritlist.update(inheritobj.clues) <- that line applies [07:46] Admiral_Chicago: the line you posted is a recursion [07:46] it ensures that all inherits of the inherited one get applied as well :) [07:46] actually there is no loop check which means you can kill bughelper :) [07:46] oh i see very good [07:46] by some circle in inherits ... ouch ;) [07:47] kind of DoS attack [07:47] given that lots of people can commit clues [07:47] anyway, the recursion ensures that you will end up with a stack-overflow [07:47] so at least the process will be ended :) [07:48] haha [07:48] or maybe even recover if that execption is catched somewhere "accidentially" [07:48] ... so really an interesting piece of code :) [07:48] #FIXME: Thanks Matthias Klose: fix endless recursion! [07:48] is even a comment [07:48] i didn't see that :) [07:48] yea i laughed at that... [07:49] comments in the code:##HEY FIX THIS BUG..KKTHXBYE [07:49] Admiral_Chicago: i don't think we need to fix it for our new feature [07:49] but should keep an eye on that [07:50] should we first file a bug report so people know what we are doing? [07:50] no [07:50] we develop that feature [07:50] then submit bug report with patch [07:50] okay. [07:50] unless you feel that it takes lots of work [07:51] e.g. there might be the likelyhood that markus says: "hey we want to do it different" [07:51] just to be fair :) [07:51] but actually i think its ok to add inheritable="xx" [07:51] to individual clues [07:51] srcpkginfo = self.get_info_file(srcpkg, verbosity) [07:51] hmm, okay [07:51] -> # [07:51] -> inheritlist.update(set(self.get_info_file(srcpkg, [07:51] # verbosity).clues)) [07:52] --> inheritlist.update(set(srcpkginfo.clues)) [07:52] ??? [07:52] huh? [07:52] are you asking me? [07:53] yes :) [07:53] i think you can fix that [07:54] e.g. use srcpkginfo.clue [07:54] s [07:54] but since you have to filter that array anyway, lets look at it afterwards === Don [n=dpowers6@ip68-102-225-218.ks.ok.cox.net] has joined #ubuntu-mozillateam [07:56] hmm... [07:57] Admiral_Chicago: i am currently getting feedback from thekorn [07:57] in -bugs [07:57] i'm following the discussion [07:57] k === JenFraggle [n=jen@host86-142-4-113.range86-142.btcentralplus.com] has joined #ubuntu-mozillateam === red_herring [n=rj@unaffiliated/redherring/x-220354] has joined #ubuntu-mozillateam [08:02] hey there JenFraggle [08:19] asac: sorry i stole someone from our loco with good python skills... [08:19] bringing him up to date [08:19] Admiral_Chicago: k [08:19] no problem [08:19] Admiral_Chicago: get him in the mt :) [08:19] he is here. [08:20] red_herring: engine? [08:20] asac: that's him... [08:20] yeah :) [08:20] i know [08:20] well [08:21] engine/whatever you wanna call it [08:21] welcome red_herring ;) [08:21] hi asac [08:21] red_herring: do you have the latest code? [08:21] i just apt-get installed it, contemplating grabbing the real source [08:21] you'll need to be in BugSquad and an ssh key to grab the latest (real) code [08:22] red_herring: you want to get involved with distro stuff a bit? [08:22] uh oh, i dont think i have either. [08:22] nope [08:22] asac: ubuntu? a *bit* [08:22] asac: i keep trying to get him involved... [08:22] im around just rarely helping ;) [08:22] red_herring: ssh-keygen [08:23] red_herring: maybe thats because you have no exciting tasks at hand :) [08:23] upload the id_rsa.pub to lp [08:23] kk [08:23] and join https://launchpad.net/~bugsquad [08:24] k updated it [08:25] done [08:25] it'll take a bit but follow the guide in https://wiki.ubuntu.com/BugHelper/doc/getting-started [08:25] bzr: ERROR: Not a branch: sftp://rjmarsan@bazaar.launchpad.net/~bughelper-dev/bughelper/bughelper.main/ [08:25] red_herring: we have a rather important spec open which would be widely used: https://blueprints.launchpad.net/ubuntu/+spec/firefox-distro-addon-support [08:26] red_herring: we need someone to code the "rather" simple web service ;) [08:26] red_herring: you join bugsquad too? [08:26] Admiral_Chicago: yeah [08:26] you need to or it won't grab it. took me and xjdriver 45 minutes to figure that one out. [08:26] red_herring: might need to wait for LP to update [08:27] asac: so was Admiral_Chicago when he mentioned python? [08:27] *lying [08:27] think it is the same message i got when trying to get bughelper sorted [08:27] excuse me, im using a dell laptop so everything i say sounds like a 12 year old kid playing counterstrike [08:28] red_herring: why? [08:28] red_herring: for the time being just grab the source with this command "bzr checkout http://bazaar.launchpad.net/~bughelper-dev/bughelper/main/ bughelper" [08:29] asac: red_herring is a python guy, there is plenty of work for him in Ubuntu [08:30] yeah ... probably :) [08:30] let us know when you have the source [08:30] got it [08:30] but why do something when there are specific things todo :) [08:30] JenFraggle: i have an assignment. just move the clue file dealing with bugs from the top to the bottom of the clue file [08:31] so we can just do bugxml-a in the future [08:32] done, shall I commit? [08:32] so the tags related clue is up at the top [08:32] let me see it first JenFraggle please [08:32] i was just about to say that :o) [08:33] hehe. [08:33] so the bughelper uses XMLOperations to read the file [08:34] problem is ive got no idea how that works =p [08:34] off! to learn! [08:35] red_herring: look at infoFile.py as well [08:35] infoFiles? [08:36] oh god regex [08:37] ok i am off for some time [08:37] seee you soon [08:37] later [08:38] if i may babble a bit [08:38] your infoFile or infoFiles class does not support a way to not run code if its been inhereted [08:38] JenFraggle: that would be only for a new file [08:38] if you want me to hax at it to see if i can make it, sure [08:38] a changed one requires bzr commit [08:39] red_herring: huh? [08:39] Admiral_Chicago: the program that parses the cluefiles has no way of distingushing between if it's been inhereted or not [08:41] Admiral_Chicago: its not built in anyways, so you can either try and fix the clue file or lemme try my hand at tinkering with bughelper [08:42] well i'm just not sure I know what you mean [08:42] there is no way to know if a clue file is inherited? [08:43] actually i think the OO model of bughelper should be reviewed [08:43] OO? [08:43] object orien.. [08:43] got it [08:43] imo Object model [08:43] yeah [08:43] i mean if you look at the patch that thekorn suggested [08:43] yeah..... [08:43] actually [08:44] theres so many tools that do exactly what bughelper tries to do [08:44] beautifulsoap for one [08:44] he? [08:44] in what way? [08:44] corba and soap are ways of turning objects into XML files [08:44] ah ... i am not at that point atm :) [08:45] have no idea what lib to use for O/XML serialization [08:45] because i don't know python [08:45] granted i never looked into them with much depth [08:45] my point was if you look at the patch thekorn suggested to prevent inheritable clues to apply [08:45] he codifies this hard in the selection algorithm [08:46] yea, not very elegant from what i see... [08:46] Admiral_Chicago: the problem is that he has no real "clue" object [08:46] yep [08:46] why are we inhereting clue files anyways if the inhereted clues don't apply? [08:46] he has even no abstract interface so you can implement your own matcher logic [08:46] red_herring: because some might apply [08:47] red_herring: i think its a corner case and default should be "false" [08:47] red_herring: the idea is to find bugs that are filed against wrong package [08:47] red_herring: at least thats the most obvious use-case i can identify [08:47] yeah [08:48] asac: he == bughelper team ... :) [08:49] Admiral_Chicago: yes ... wrong said ... sorry ... actually I suspect that its the bugClue that we have at hand in the code snippet you suggested [08:49] but thekorn did not comment on it :/ so i am not sure [08:49] well thekorn is just working as this SoC project... [08:50] Admiral_Chicago: he knows the code best afaik [08:50] Admiral_Chicago: because he does most work atm :) [08:51] wow, bugClue is the most useless object ive seen [08:52] red_herring: yeah [08:52] actually i think all the suffering starts in [08:52] well no, ive seen more useless objects, but thats close [08:52] bugHelper/infoFiles.py [08:52] there you see that there is add_simple_clue [08:52] yeah i know, ive been browsing it [08:52] which means that the layer on top knows of every detail of the low level clues [08:52] it should just pass the clue xml snippet [08:52] and let the parsing/interpreting happening on a lower level [08:53] hrm [08:53] ... so we can have multiple clue implementations for instance [08:53] that carry their own match actions et al [08:53] this is all very complex [08:54] its completely coupled ... e.g. you have to understand all before you can do useful [08:54] ... which is likely due to bad OO design :) [08:54] eh, not entirely === JenFraggle [n=jen@host86-142-4-113.range86-142.btcentralplus.com] has joined #ubuntu-mozillateam [08:56] red_herring: why not? [08:56] except XMLOperations.NoClueFoundError, e: [08:56] print >> sys.stderr,e [08:56] holy crap! he's using c++ syntax! [09:06] syntax is not what bothers me ... rather extensibility/maintainability of code :) [09:06] asac: yeah definently [09:07] i love it when someone makes a big down drawing of their code explaining how it all works [09:07] of course syntax can contribute to that (negatively) [09:07] it makes my life sooooo much easier [09:07] *top-down drawing [09:07] actually bugehlper has no top-down design [09:08] it wasted class concept for code splits :) [09:08] more or less :) [09:08] well... yeah, hell *any* drawing helps [09:08] looking at 3000 lines of code is hard to visualize [09:08] a pictuere? great! [09:09] yeah ... but 3000 lines of code is rather small code-base :) [09:09] if you have problems to get the idea in such sized project than its not properly done :) [09:10] which i meant with: "all is tightly coupled" ... [09:10] in my opinion bughelper tries to achieve a three layer design [09:10] 1st layer: bughelper main program [09:10] 2nd layer: matching code [09:10] 3rd layer: parsing [09:11] so well [09:11] however then the mixing starts :) [09:11] yeah [09:11] cuz when you have things like "Don't use this when inhereted" [09:11] it gets complicated cuz it falls between 2 and 3 [09:12] imo bughelper should be a crawler :) [09:12] that has conditions to match subtrees of bugs [09:12] and then applies clues on those :) [09:13] actually the model is pretty simple ... you start with a "virtual" set of all bugs and then refine your matching [09:13] and can recurse after each refinement [09:14] then you have tags that allow you to specify what is actually spit out ... and you can use xpath expressions if you want to put some content of the current bug set to the output :) [09:16] well [09:16] :) [09:16] actually to be honest [09:16] i don't even understand how clues work [09:16] have you looked at a clue file? [09:16] its pretty self explanatory ... i mean the basic concept [09:16] you can specify conditions that are tried on each bug [09:17] if all conditions match then the clue fires and the output is dumped to the output channel [09:25] heh [09:25] sorry, one sec [09:33] asac: yeah i know, im just confused on what conditions there are [09:33] yes ... its explicitly coded somewhere in level 2nd (above) [09:33] not really extensible .. nor encapsulated [09:33] not extendable (gasp), its not using real xml! [09:33] he? [09:33] what is not real for that xml? [09:33] its valid xml, isn't it? [09:35] red_herring: conditions boil down to be just an expression that gets evaluated somewhere (e.g. like parsing python code and running) [09:36] then there are special cases encoded for AND, OR, NOT [09:37] at least that was the state when i last looked .... maybe there have been xpath conditions added by now [09:37] well [09:37] are these objects supposed to be editable by humans? [09:37] which objects? the clue/xml is editable by humas [09:38] yeah [09:40] red_herring: look at ./bugHelper/XMLOperations.py for parsing of conditions [09:40] clueCondition(and_cond, or_cond, simple_cond, not_cond, cond_field) [09:40] those parameters are just "expression strings" === gnomefreak sort of here. on phone [09:46] grr [09:47] this program gets annoying [09:48] we may want to mail bughelper-dev ML at one pont [09:49] red_herring: yes it is :) [09:50] red_herring: as i said :) my webservice task is probably much more exciting :) [09:50] asac: what was the other project you were talkin about? [09:50] red_herring: its something completely from scratch :) [09:50] its a rather small and funny thing :) [09:50] whats it require i know? [09:51] web page programming mostly :) [09:51] xml parsing [09:51] asac: have you heard from mike? i looked on svn and nothing yet just checking in [09:51] meh, not my strong point but sure [09:51] maybe later use of some python lib to introspect packages [09:51] red_herring: you know firefox? [09:51] :) [09:51] i know how to use it. thats about all. [09:51] red_herring: yeah thats enough ... if you have a missing plugin [09:52] and you click [09:52] k [09:52] there opens a plugin finder dialog wizard [09:52] ever saw that? [09:52] i don't know any fancy XUL or whatever programming [09:52] yeah [09:52] red_herring: its not about XUL ... ist just XML webservice [09:52] (well not a fully flegded webservice) [09:52] red_herring: open about:config [09:52] in firefox [09:52] yeah yeah, the bajillion variables [09:53] red_herring: there is a pfs.datasource.url [09:53] setting [09:54] if you open that in url [09:54] i mean the url [09:54] you will see an xml answer [09:54] ah yeah [09:54] basically you can pass a MIME-type and get a list of plugins that support that kind of type [09:54] the idea is to write a webservice that takes those requests [09:54] asks the mozilla webservice [09:54] huh [09:54] and adds "other plugins" [09:55] yeah [09:55] if that is done we will ship firefox with "our" webservice [09:55] so people will find "our" packages [09:56] ... if we have plugins packaged that support that mime type [09:56] which are made for linux... right? [09:56] yes ... there are multiple problems why we do that [09:56] yeah [09:56] we can talk about that issue later :) ... did you ever wrote some webpage in python/php ? [09:57] php plenty [09:57] never in python [09:57] really? [09:57] granted im not very good at php other than the bare basics [09:57] i thought about php first ... probably easier to host for testing [09:57] red_herring: yeah its not that difficult i guess [09:57] i can't imagine python as being a veyr good webpage language [09:58] to be perfectly honest [09:58] yeah ... there are different opinions on that :) [09:58] its got too much overhead [09:58] anyway ... if you have a webserver at hand ... point the about:config setting to your php script [09:58] and see what gets passed [09:58] if for instance click on a missing flash icon [09:59] asac: would that need to be dynamic for various DEs? [09:59] DEs? [09:59] i think we could add caching ... e.g. if we know that we have these plugins we could precache results [09:59] so we don't need to call mozilla webservice for each and every request [10:00] but i think its good to start completely dynamic and add a cache afterwards [10:00] asac: it doesn't pass anything special [10:00] it should pass some parameters [10:00] did you keep the %XXX labels in url [10:00] ? [10:01] none by default [10:01] no [10:01] i knwo those pass parameters =p [10:01] https://pfs.mozilla.org/plugins/PluginFinderService.php?mimetype=%PLUGIN_MIMETYPE%&appID=%APP_ID%&appVersion=%APP_VERSION%&clientOS=%CLIENT_OS%&chromeLocale=%CHROME_LOCALE% [10:01] so you should get mimetype [10:01] appID [10:01] yeah i know [10:01] and CLIENT_OS [10:01] and so on :) [10:02] so all we do is write a simple php script to read these values and built a corrisponding xml file [10:02] isnt the hard part packaging? [10:02] ah cool ... so next step would be just to proxy that call to the original url :) [10:02] red_herring: yes ... 1st ... invoke official service (e.g. pfs.mozilla.org) [10:02] 2nd add some more entries [10:02] red_herring: the second part will be a bit tricky [10:02] Desktop Environment.. [10:03] as we have to figure out how to find "what plugins" are available et al [10:03] asac: so how much have you done so far? [10:03] red_herring: nothing ... i drafted the spec and am looking for someone who is willing to help on the server part [10:03] ... as the spec involves mozilla side part as well [10:03] heh ok [10:04] as well as coordinating that ll plugins get info about "what mimetype do we provide [10:04] so we can find the plugins on the webservice [10:04] though i know a lot of FOSS guys will get upset if we make it easy to install badpeople apps like flash ;-) [10:04] heeh [10:05] red_herring: actually we offer gnash as well [10:05] oh so we could offer both? [10:05] which is "more" then now [10:05] we just add the info to the XML [10:05] the user gets a list of available options [10:05] true true [10:05] god i hate gnash [10:05] but sure [10:05] yeah its not just gnash ... its for everything [10:05] yeah [10:06] <3 gash [10:06] gnash* [10:06] wait, is the java plugin oss yet? [10:06] its for things that are packaged as well (e.g. nonfree-flashplayer) ... so people can install "ubuntu" managed package instead of following wierd install instructions of adobe [10:06] red_herring: dunno ... but there exist choice in java as well [10:06] yeah [10:06] i wish there was shockwave..... [10:06] red_herring: its most important that we can offer and educate the user that there are packages available [10:06] oh, also [10:06] asac: what about 64 bit users? [10:07] we are working on that too ... but actually its why we want to push free flash solutions [10:07] as they work on any architecture [10:07] i run 64 bit firefox, plugins hate it so terribly [10:07] yeah [10:07] red_herring: yeah [10:07] red_herring: we work on nspluginwrapper [10:07] i am leading that spec as well === red_herring strangles nspluginwrapper [10:07] yes i need to evaluate that first [10:08] its never worked for me [10:08] e.g. how can we package it in such a way that it really works well === gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-mozillateam [10:08] well it does, but 15 minutes into a youtube vid it *will* crash [10:08] if its not possible then we can't do a thing, but help free software impls [10:08] yeah [10:08] red_herring: yes ... but maybe we can find a proper flash version that works and package that :) [10:08] yeah for real [10:09] actually, the only real problem with nspluginwrapper is that firefox won't restart it until i restart firefox [10:09] unlike opera or konqueror, which starts a new instance with every tab/window [10:09] red_herring: ah ... so nspluginwrapper crashes? [10:09] yeah [10:09] thought ffox crashes ... or flash :) [10:09] anyway ... would be great if you could help a bit :) [10:09] on the webservice :) [10:09] naw nspluginwrapper [10:10] yeah sure [10:10] ive got my own server [10:10] can you give me the parameters you got? [10:10] ? [10:10] from the URL [10:10] e.g. clientOS [10:10] https://pfs.mozilla.org/plugins/PluginFinderService.php?mimetype=%PLUGIN_MIMETYPE%&appID=%APP_ID%&appVersion=%APP_VERSION%&clientOS=%CLIENT_OS%&chromeLocale=%CHROME_LOCALE% [10:11] clientOS ... chromeLocale ... appVersion, appID, mimetype [10:11] e.g. if you try to install flash :) [10:11] umm [10:11] i just want to figure out if we have to parse HTTP headers to get the "distribution" info ... e.g. are we on feisty :) [10:11] heres the http header when i try to search for shockwave [10:12] 192.168.1.1 - - [04/Jun/2007:15:11:11 -0500] "GET /plugins/PluginFinderService.php?mimetype=application%2Fx-director&appID={ec8030f7-c20a-464f-9b0e-13a3a9e97384}&appVersion=2007060115&clientOS=Linux%20x86_64&chromeLocale=en-US HTTP/1.1" 404 229 "-" "Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.8.1.4) Gecko/20061201 Firefox/2.0.0.4 (Ubuntu-feisty)" [10:12] good [10:12] the mimetype looks wierd though [10:12] imetype=application%2Fx-director [10:12] ?? [10:12] yeah [10:12] application/x-director ? [10:12] x-director = shockwave [10:12] yeah [10:12] crazy [10:13] why do people always use zillions of different mimetypes for the same [10:13] ok ... so parsing the header might allow us to see if we are on feisty et all [10:13] ok ... i think thats enough for now [10:13] however ... people can change that [10:14] no idea. [10:14] ... but than bad luck :) [10:14] hrm [10:14] going through my logs [10:14] lots of people are visiting my lolcode site === red_herring makes counter [10:14] ;) [10:15] red_herring: ok ... if you come up with the base script to proxy the invocation to pts.mozilla.org I will try to get the info how we can find packages that match those parameters :) [10:15] anyone else having issues with LP? [10:15] sure, gimme a bit [10:15] red_herring: hey :) nothing to hurry ;) [10:16] true [10:17] red_herring: actually i will be off today :) ... will you be here sometime tomorrow/this week? [10:17] off for today [10:17] i ment [10:17] its my summer [10:17] freetime = all time [10:17] red_herring: lucky man :) === red_herring bows [10:18] asac: night [10:18] gnomefreak: night! [10:19] red_herring: if you tweak the clientOS when proxying requests to pfs.mozilla.org you might even get valid content back (e.g. to see how the xml looks like) ... [10:19] what is debians log in screen called? [10:19] log? [10:19] login [10:19] heh [10:19] xdm? [10:20] gnomefreak: i don't know what log you refer to :) [10:20] gnomefreak: you mean xorg log? [10:20] asac: the login screen [10:20] ah [10:20] where you type name and password [10:20] gdm [10:20] ;) [10:20] oh it uses gdm [10:20] xdm is the basic x one [10:20] you can probably choose [10:20] but i think debian default is gnome [10:20] k ty (filing bugs [10:20] ) [10:22] ok ... off for real [10:22] cu === ubotu [n=ubotu@ubuntu/bot/ubotu] has joined #ubuntu-mozillateam [10:33] ok ... i will be back in 20 minutes ... TV sucks and I even cannot play UT2004 because it recently freezes my system for whatever unknown reason [10:33] :) [10:33] your ati drivers are causing the freezing :( [10:33] yeah [10:34] and i was too dump to get the official ati drivers installed [10:34] now i don't have hardware accelleration :( [10:34] !ati [10:34] To install the Ati/NVidia drivers for your video card, see https://help.ubuntu.com/community/BinaryDriverHowto [10:34] tell you how to install the ones from ati iirc [10:34] tells [10:36] firefox has mem leaks OMG like this hasnt been known for years [10:37] Some people might experience random hangups. I heard that this might help: Edit /etc/X11/xorg.conf and add the following options in the corresponding section: [10:37] :) [10:37] maybe i should just add those zillions of options :) [10:38] https://help.ubuntu.com/community/BinaryDriverHowto/ATI === gnomefreak has never installed ati with any rate of success === gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-mozillateam [10:49] JenFraggle: we need to build a list of every bug that has a duplicate [10:49] i think there is a way to do that in bughelepr [10:49] helper* [10:49] not sure though [10:52] it really sucks .... this ati driver thing [10:53] i should have never tried to install through ati.com installer [10:53] damn [10:54] maybe i should wipe everything because of this [10:54] it just polluted something i cannot find anymore [10:55] asac: maybe it built its own l-r-m package like nvidias installer does? [10:58] i have to sue ati [10:58] this is completely unacceptable [10:58] i mean they polluted my system [10:59] YES [10:59] i found the intruder [10:59] at least i have a proper kernel module now again [10:59] hmm ... but actually i want to have amore modern one [11:00] lets see if i can drop something into the source package :) [11:00] lol [11:10] hehe [11:10] i just drop a patch from ati source :) [11:10] hehe [11:10] why do we patch the ati source at all? [11:10] who knows [11:11] got someone wants new ati drivers but he doesnt awant to help build them lol [11:11] yeah ... maybe because of the "not-invented-here" syndrom [11:11] did we add anything to tbird in gutsy that isnt in tbird in the repo? [11:11] in the repo? [11:11] what do you mean? [11:12] MT repo [11:12] in bzr? [11:12] dunno ... depends on what the last version in there is [11:12] afaik i uploaded latest tbird once [11:12] dunno if you did for i386 [11:12] i did [11:12] ~3 [11:12] or something like that === cjwatson_ [n=cjwatson@82-69-40-219.dsl.in-addr.zen.co.uk] has joined #ubuntu-mozillateam [11:14] y [11:14] then there is nothing new [11:14] k [11:16] ok press thumbs [11:16] i installed my hand made ait packages :) [11:16] mozilla supports releases for 6months only? [11:17] asac: does it work? [11:17] no idea ... moving things to save place then stopping X reloading new module et al [11:17] gnomefreak: yeah ... they support major release for at least one year [11:17] problem is they say first release is 1.5.0.0 [11:18] so 1.5.0.12 is the last :) [11:18] and at least 6 month after next major release is out [11:18] which is why they dropped ffox now ... as 6+ month ago 2.0 was out [11:18] that means 2.0 will be EOS before dapper is EOS [11:18] thats bad [11:26] cool [11:26] it appears not to freeze anymore [11:26] Yau [11:27] :) [11:27] ill be back later maybe i need to get something done today (other than phone calls and meetings)