[02:43] hi all. [10:57] Jazzva: i got a fail to build mail about foxyproxy? [10:57] whats going on? [10:58] didnt that ever build? [10:58] and nobody complained? [11:03] debian bug 497491 [11:03] Debian bug 497491 in icedove "icedove: Icedove inappropriately sets file-/MIME-type associations in .desktop database" [Critical,Open] http://bugs.debian.org/497491 [13:03] asac, I think we talked about that a month ago. It failed back then too, and I think that you pushed 2.7.5-0ubuntu2 [13:03] I'll check if you did [13:05] asac, https://code.edge.launchpad.net/~jazzva/firefox-extensions/foxyproxy.ubuntu/+merge/586 [13:05] Hmm, I thought that you merged that :) [13:06] hmm, you did [13:06] asac, https://code.edge.launchpad.net/~ubuntu-dev/firefox-extensions/foxyproxy.ubuntu [13:06] By Alexander Sack on 2008-07-29 * RELEASE 2.7.5-0ubuntu2 to ubuntu/intrepid [13:17] Jazzva: hmm [13:17] Jazzva: maybe it was just late coming mail then? [13:17] (it was about ubuntu1) [13:17] Maybe [13:18] Jazzva: oh [13:18] ubuntu2 wasnt uploaded ;) [13:18] but then again, -0ubuntu2 isn't showing [13:18] ah :) [13:18] back [13:19] Volans, welcome back [13:20] Hi Jazzva :) [13:22] asac, when will abrowser be included in ubuntu? Jaunty? Or Intrepid? [13:22] Jazzva: uploaded [13:22] Jazzva: next upload [13:22] asac, thanks :) [13:23] all is done in the branch. i just need to rename everything to abrowser now :) [13:32] Jazzva: ok abrowser-brnading uploaded to mt archive [13:32] hmm without orig ;) [13:33] heh :)... off to get smokes [13:33] ok reuploding [13:36] hmm ... the orig was already there as it seems [13:36] anyway ... should be building now [13:37] * asac preparing for lunch [13:41] Jazzva: ok the i386 build finished for intrepid [13:41] https://edge.launchpad.net/~mozillateam/+archive [13:41] could you test whether it works nicely for you? [13:41] the idea is to install abrowser -> you get the abrowser branding [13:41] if you then install "firefox" .... you ge tthe firefox brandin [13:42] if you have firefox installed and just upgrade you should still have firefox branding [13:54] asac, ok [13:54] so, to remove firefox and get abrowser? [13:54] Jazzva: just install apt-get install abrowser [13:54] that hsould do the trick [13:59] asac, abrowser is not showing in mt ppa... [14:05] ah, it's in firefox-3.0 [14:07] ok, it's here now... when I was updating sources, it fetched the old list [14:10] asac, seems the menu entry for A Browser is wrong [14:10] it says it can't run webbrowser command [14:10] maybe it should be abrowser? :) [14:11] @time [14:11] Current time in Etc/UTC: September 11 2008, 13:11:30 - Next meeting: Ubuntu Java in 48 minutes [14:11] asac, menu entry says "A Browser"... title bar says "Web Browser" [14:35] Jazzva: right. makes sense [14:36] Jazzva: icons ? [14:36] icons are ok [14:36] neonish light-blue glob [14:36] *globe [14:36] though, it uses the mozilla globe when you go to help -> about -> credits [14:37] and, afaics, the rest of the icons are taken from gnome theme. [14:40] ok, lunch. see you later [15:08] back [15:30] Jazzva: ok. the .desktop thing is still broken [15:30] e.g the menu [15:30] asac, why don't you just change the call to abrowser? [15:30] :) [15:32] Jazzva: in .destkop? [15:32] thats the idea [15:32] mhm [15:32] its just that i forgot it [15:32] ok, that's what I did locally [15:32] oh, ok :) [15:34] Jazzva: ok committed and uploaded a new version to ppa [15:34] yay :) [15:34] I'll upgrade it, once it's built [15:35] Jazzva: have you tried to switch back and forward? [15:35] e.g. install firefox, then install abrowser? [15:35] well, I had firefox, then installed abrowser [15:35] Now I have it it, and firefox is missing [15:35] :) [15:35] I mean, it's menu entyr [15:35] *entry [15:36] yes [15:36] thats expected [15:36] you should have A Browser in the menu ;) [15:37] I have it :) [15:37] this is strange. i use the same .desktop file name for both [15:37] but still when i put the menuentry into the gnome panel [15:37] it doesnt change [15:38] it sticks to whatever pic and command was used in the old thing [15:38] it doesn't? [15:38] well, it changed for me [15:38] no. [15:38] from Fx -> A Browser [15:38] Jazzva: the menu entry or a quick starter? [15:38] menu entry [15:38] Jazzva: yes that works [15:38] but the quick start icon in the panel doesnt [15:38] hmm [15:38] maybe it needs to be reloaded [15:38] killall gnome-panel [15:38] and let's see what will happen [15:38] :) [15:39] indeed [15:39] thanks [15:39] hmm [15:39] Jazzva: ok so in .postinst -> killall gnome-panel ;) [15:39] it has a different icon (the A Browser's globe) [15:39] but the title still says Firefox Web Browser [15:39] it does? [15:39] maybe I manually added it... [15:40] I mean, the title [15:40] Jazzva: if you manually modify it it will stick [15:40] maybe dump it and create a new fresh one [15:40] right... that's probably why it didn't change [15:40] by dragging from menu or something [15:40] yeah.. I'll do that, when I go back to firefox, to see if it will change when I switch to a-browser [15:42] good thing about the name is that it will appear on top of the menu ;) [15:42] err at the top ;) [15:42] yep :) [15:44] Jazzva: http://paste.ubuntu.com/45821/ [15:45] switching to firefox, just a sec ;) [15:45] i wonder if korean folks will feel offended ;) [15:46] oh, yay... my gnome-panel is not reloading :) [15:48] Jazzva: for me the first time it took a long time [15:48] but then became snappy ;) [15:48] asac, I don't really know korean [15:48] why would they be offended? :) [15:49] becuase i just added a "A " in front of whatever i think means "Web Browser" [15:49] aha :) [15:49] noticed [15:51] * asac goes to New bugs processing for NM :( [15:51] asac, what's the current version of abrowser? [15:51] 0ubuntu2~mt1? [15:52] nope, ~mt2 [15:54] Jazzva: mt2 [15:54] yes [15:57] oh, great. now I'm left without the terminal and gnome panel... [15:59] ok, panel is back :) [16:07] * armin76 waves to panel [16:07] armin76: killall gnome-panel === asac_ is now known as asac [17:26] asac, how about using strict versioning in debian/control for abrowser and abrowser-3.0-branding [17:26] ? [17:27] I just upgraded abrowser, but it didn't upgrade abrowser-3.0-branding [17:33] hmm... it seems it doesn't update the quick start icon [17:33] I deleted the old one, and put a new one, that I dragged from the menu [17:33] bug #260055 [17:33] Bug 260055 on http://launchpad.net/bugs/260055 is private [17:33] hmm [17:34] Jazzva: huh? [17:34] installed abrowser after that, killed gnome panel, restarted it, the icon changed, but it still says Firefox Web Browser [17:34] Jazzva: why didnt it upgrade -branding? [17:34] dunno... [17:34] anyway .. its uploaded ;) [17:34] I did apt-get install abrowser, and it downloaded it [17:34] and after that a-g install abrowser-3.0-branding [17:34] let me check i the binaries are built [17:34] (they are NEW) [17:35] that's ~mt2? [17:36] Jazzva: no its bumped [17:36] ok [17:39] Jazzva: ok. pitti will proces them tomorrow [17:39] ok [17:39] Jazzva: you sure -branding doesnt upgrade to mt2 for you? [17:43] asac, I did it manually [17:44] asac, http://paste.ubuntu.com/45847/ [17:44] that's the output [18:25] asac: hi again, so the FF3/NFS issue is back on the table. please advise [19:16] Jazzva: ok. thats ok. i thought it didnt upgrade when running "upgrade" [19:16] pmatulis: are you in the state where you cannot connect to the places.sqlite DB? [19:19] asac, oh... ok [19:20] asac, wouldn't it be safer with strict versioning :)? [19:20] Jazzva: there is no risk ;) [19:20] asac, ok [19:20] so it cant be safer ;) [19:20] mbrowser is just a meta package with nothing in it [19:20] abrowser ;) [19:20] ah [19:23] asac: from a working system (as well as a broken one) the .tables command shows 'database is locked'. do you have a specific test for me? [19:23] pmatulis: so it isnt different on a broken and working system? [19:24] asac: that's right [19:24] (both when browser is closed) [19:24] asac: one sec [19:24] pmatulis: it should be reasonable that the database is locked when browser is running [19:24] what would be a problem is when database is locked while browser isnt running [19:24] asac: right [19:24] which i understood is the case for nfs cold boot [19:24] is that really the case? [19:25] asac: locked on broken [19:25] asac: how goes it? [19:25] pmatulis: locked=logged? [19:25] Lns: if someone asks like this i always have the feeling that it wont get better ;) [19:25] asac: on working it does not connect ('encrypted or not a database') [19:26] pmatulis: ok. thats when browser is not running? [19:26] asac: on broken, i connect but the .tables command gives 'database is locked' [19:26] asac: yes, not running [19:26] asac: lol... just checking in with you regarding what we discussed last week regarding Firefox 3.01 in Ubuntu/LTSP, and the slowness/lockup issues I've (and others) have been experiencing.. https://bugzilla.mozilla.org/show_bug.cgi?id=453704 [19:26] Mozilla bug 453704 in General "Extreme slowness, "Firefox is already running" error for >3 users launching Firefox in LTSP environment" [Critical,Unconfirmed] [19:26] pmatulis: do you use sqlite3? i think i saw that "encrypted" error when trying sqlite (aka v2) [19:27] asac: i'm on hardy, it's sqlite3 [19:27] asac: but my broken system is intrepid [19:27] pmatulis: hardy also has sqlite (aka 2) [19:28] asac: yes, just saw that [19:28] pmatulis: $ sqlite places.sqlite [19:28] Unable to open database "places.sqlite": file is encrypted or is not a database [19:28] asac@tinya:~/.mozilla/firefox/8bil8cfx.default$ sqlite3 places.sqlite [19:28] SQLite version 3.4.2 [19:28] pmatulis: so please try sqlite3 ;) [19:28] asac: ah [19:28] Enter ".help" for instructions [19:28] sqlite> asac@tinya:~/.mozilla/firefox/8bil8cfx.default$ [19:29] pmatulis: ok. so i assume that the problem really is that its locked [19:29] on broken system ... but not on working system (when ffox is not running for both) [19:29] asac: now with sqlite3 on both systems i get same output [19:29] asac: no difference [19:29] pmatulis: what exactly? [19:30] "Datbase locked" ? [19:30] asac: .tables command gives 10 tables (moz_anno_attributes, moz.....) [19:31] pmatulis: and are you positive that firefox is still broken now? [19:32] asac: on broken, i lost all toolbar bookmarks and back/forward buttons [19:32] pmatulis: ok. but when stopping firefox, moving places.sqlite3 away and then back to that location and starting firefox again cures you? [19:33] asac: stand by [19:33] asac: yes [19:33] pmatulis: really? [19:33] asac: yes [19:34] pmatulis: hmm. what i would find interesting is whether the size of that file changed [19:34] from before and after [19:34] but you probably cannot test that quickly? [19:34] asac: stand by for cold boot and cure? [19:36] pmatulis: let me test something [19:38] hmm i managed to get the db into a "locked" state [19:39] but then i wanted to see if that reproduces it [19:39] but now i cant get that into that state anymore [19:42] asac: nope, no change in file at all [19:42] asac: but cure works again [19:43] pmatulis: good. [19:43] pmatulis: i am still thinking that its a lock causing this [19:43] which is done on a filesystem basis and moving back and forth resolves that [19:43] now i am trying to get my database in such a state ;) [19:44] asac: pop up an nfs share! [19:45] asac: i think my previous results became inconsistent b/c my nfs share wasn't set up properly (uid mapping problem) [19:45] pmatulis: well. for you the db isnt even locked? [19:45] (when using sqlite3)? [19:45] asac: wasn't before, trying again [19:46] asac: note that i still have to go through the profile locking barrier [19:46] asac: by removing .parentlock and .lock [19:46] asac: not even looking at that (not sure if it's related) [19:48] * asac currently reading http://developer.mozilla.org/en/Storage#SQLite_Locking [19:50] ah ok [19:50] so begin EXCLUSIVE; [19:50] in sqlite3 will put the db into a locked state [19:51] asac: this time the DB is locked (?) [19:53] pmatulis: ok. but ffox isnt running? [19:56] asac: correct [19:57] pmatulis: ok. now what we would need to know if there is a way to forcefully release a flock [19:57] which i think is on that file [19:57] (if you dont get "database locked" firefox probably adoesnt have an exclusive lock, but a shared one) [20:06] asac: you want me to try something else? [20:42] pmatulis: i am out for 1h or so cu later [20:43] asac: do you want me to open a bug for this or not? [22:43] asac: around? [22:45] bdmurray: one sec [22:46] bdmurray: so... go ahead ;) [22:47] bug 216354 looks like a dup of bug 204631 - I was planning on merging them [22:47] Launchpad bug 216354 in firefox-3.0 "firefox crashed with SIGSEGV in __kernel_vsyscall()" [Medium,New] https://launchpad.net/bugs/216354 [22:47] Bug 204631 on http://launchpad.net/bugs/204631 is private [22:48] and making that one not private [22:49] bdmurray: yeah. also maybe remove the [] brackets around master [22:49] launchpad doesnt find MASTER if they are in brackets (anymore) [22:49] that's really neat [22:58] asac: where did you notice that search feature? [22:58] bdmurray: hmm. works better now [22:58] strange [22:59] hmm, and I already changed it - oh well [23:01] bdmurray: not a problem :) ... as long as it shows up when searching for MASTER its fine imo [23:03] asac: bug 216496 and bug 263749 also seem to be dups should the newer one be used as the master? [23:03] Bug 216496 on http://launchpad.net/bugs/216496 is private [23:03] Bug 263749 on http://launchpad.net/bugs/263749 is private [23:04] bdmurray: imo. there is no need to re-master bug-clusters :) [23:04] so keep the original main bug [23:04] if there is valuable info it might be worth pasting that into the main bug [23:04] neither has any dups and the newer one is about the 8.10 package [23:05] whoops, I guess the older one has one dupe [23:06] yeah. whatever bug has less dups becomes a child bug imo [23:07] right, I'll just add the version from the 8.10 bug into the parent [23:09] bdmurray: is there any hint about how to reproduce them? [23:10] bdmurray: hmm [23:10] 263749 is a reall old bug [23:10] i mean a dupe of a old bug [23:10] look for gtk_style_realize [23:11] in firefox maybe and not firefox-3.0 maybe? [23:11] bdmurray: yeah in firefox [23:12] oh dear [23:12] launchpad search again :) [23:13] bdmurray: 88011 [23:13] 88011 [23:13] 91054 [23:14] 114943 [23:14] 70875 [23:14] 74252 [23:14] hmm ... i am sure we had a master with 200+ dupes [23:14] where is it? [23:14] maybe 72018? [23:15] bug 72018? [23:15] bug 72018 [23:15] Launchpad bug 72018 in xulrunner-1.9 "MASTER Firefox Crash [@gtk_style_realize] [@nsFilePicker::Show]" [High,Fix released] https://launchpad.net/bugs/72018 [23:15] ah yeah [23:15] bdmurray: ok (variant) [23:15] was the left over bug [23:15] which i set to "wont fix" [23:15] because i didnt get more dupes [23:15] bdmurray: so if the traces are similar we should reopen that one [23:18] They look similar to me. [23:18] bdmurray: i dont mind [23:18] we can also make a new master for ffox 3 [23:19] the dups talk about gtk_style_realize, gtk_style_attach and gtk_widget_set_style_internal which all match [23:20] tse [23:20] bdmurray: so which ffox 3 bugs are candidates for dupes? [23:20] oh got you [23:21] bug 216496 and its dups [23:21] Launchpad bug 216496 in firefox-3.0 "firefox crashed with SIGSEGV in __kernel_vsyscall()" [Medium,New] https://launchpad.net/bugs/216496 [23:22] bdmurray: ok. i dont mind what you do with them [23:22] they are certainly confirmed [23:22] but i get the feeling that this is a whole class of crashes [23:22] and most likely its different to the one we had in ffox 2 [23:24] too bad that we have no clue how to repro [23:25] iirc, this crash class happens when there is a dangling window for an already freed widget [23:25] or was it the other way around [23:25] hmm [23:25] asac: This isn't a dupe, but I was wondering if you had time to work on this bug with me: https://bugzilla.mozilla.org/show_bug.cgi?id=453704 [23:25] Mozilla bug 453704 in General "Extreme slowness, "Firefox is already running" error for >3 users launching Firefox in LTSP environment" [Critical,Unconfirmed] [23:25] Lns: did you get feedback from ogra? [23:26] asac: no i didn't, was he notified of it? [23:26] Lns: didnt i send you to him? [23:26] asac: nope..you told me to catch up with you the next day re: it [23:27] yeah. ok [23:27] and my wife had our baby that day...so...i wasn't too available until now [23:27] you should remember me earlier [23:27] he is again gone for today [23:27] ;) [23:27] Lns: ah ok [23:27] heh.. ok =) so did you talk with him at all about it? [23:27] no [23:27] Lns: maybe ill remember when i see him chatting tomorrow [23:27] Lns: its easier to open a bug in launchpad [23:27] and subscribe him there i think [23:28] I'll do that [23:28] (open a bug) [23:28] do you have a bug? [23:28] just in bugzilla [23:28] yes. subscribe me too [23:28] asac [23:28] and ogra [23:28] ok, np [23:28] * Lns didn't know you could subscribe other people [23:28] that mail will end up in a high prio mailbox then [23:28] you can on the side there is an action [23:28] "subscribe someone else" [23:28] nice. ok i'm on it. thanks =) [23:29] welcome [23:30] bdmurray: where is the latest bug graph for ffox 3? [23:30] asac: i did sub to this bug, thinking it might be related..any thoughts? https://bugs.launchpad.net/firefox/+bug/229745 [23:30] Launchpad bug 229745 in xulrunner-1.9 "after fix for #215728 - Committing to urlclassifier3.sqlite still causes excessive CPU usage and disk I/O (the 2nd)" [High,In progress] [23:32] Lns: well. thats still a problem and it causes IO [23:32] Lns: if you dont have a high IO load [23:32] then its not the problem most likely [23:32] or is that IO load you are seeing? [23:33] asac: well i haven't checked IO, but 35 people jumping on FF at the same time would definitely cause blockage... [23:33] well, the CPU definitely pegs out, even with just a few users jumping on [23:33] Lns: that problem doesnt start right on startup from what i know [23:33] sitting idle will still cause single procs to consume 30-60% cpu [23:33] its something that gets worth over time [23:33] ok [23:33] maybe 2 separate things [23:33] worse ;) [23:34] Lns: for you disabling that malware thing helps? [23:35] asac: I think it did help, but there are still issues. I did get some reports that after "a couple of minutes" firefox will launch - some after getting the "Firefox is already running" error and trying again [23:35] It seems to have just made it a bit more bearable (but still not acceptable to wait 2-4 minutes for it to launch) [23:36] the firefox already running error sounds wierd [23:36] but i think i already said that [23:36] sounds more like a wierd X11 issue [23:36] at least if those are different unix-users with different home directories [23:38] Lns: anyway. i would suggest to open a new bug anyway. if its a dupe we can still merge it into another bug [23:40] asac: ok, i'll do that. Yeah, I really thought it was weird too (the error). It really shouldn't be happening, as they are all different users w/different homedirs. [23:41] Lns: firefox looks for locks in profile dir [23:41] Lns: it also tries to contact a running X window [23:41] Lns: so if X desktops are somehow shared this might cause confusion [23:41] asac: yeah, someone mentioned that in an LTSP list - check this out [23:43] asac: http://pastebin.com/m7386054d [23:46] Lns: by coincident i looked at that locking code today [23:46] ooo, fresh on the noggin ;) [23:47] any ideas? [23:47] I find it weird that it uses 127.0.0.1, which may or may not conflict with it looking at the server vs. thin-client [23:47] depending on the code I guess [23:47] Lns: that symbolic link just allows users of the same profile to obtain a lock [23:48] on filesystems that dont have native locking support [23:48] Lns: can you look if .parentlock is also a symlink? [23:48] I can see if i can find any on my own system, although i'm not experiencing this problem [23:49] nope - -rw-r--r-- 1 jerickson jerickson 0 2008-09-10 14:28 .parentlock [23:50] lrwxrwxrwx 1 jerickson jerickson 15 2008-09-10 14:28 lock -> 127.0.1.1:+3887 [23:50] jerickson@Fibonacci:~/.mozilla/firefox/5rwscmny.default$ file lock [23:50] lock: broken symbolic link to `127.0.1.1:+3887' [23:50]