/srv/irclogs.ubuntu.com/2008/01/03/#ubuntu-mozillateam.txt

=== \sh is now known as \sh_away
=== Ubulette_ is now known as Ubulette
=== asac_ is now known as asac
asac[reed]: there?10:14
asac[reed]: what do i need to request in order to get things in atm? blocking1.9? or just approval1.9 on the patch?10:16
asac(or both) ?10:16
[reed]asac: is it a blocker?10:16
asacdepends for whom :)10:16
[reed]is the patch reviewed yet?10:17
asacmozilla bug 408062 ... build system improvement, but important for distros10:17
ubotuMozilla bug 408062 in Build Config "Provide short-hand includetype=unstable .pc files for libxul and libxul-embedding" [Enhancement,Assigned] http://bugzilla.mozilla.org/show_bug.cgi?id=40806210:17
[reed]you can't request approval until you have review10:17
[reed]once you have review, then you should request approval10:17
asacwell its good, i just have to update according to comment10:17
[reed]well, he gave you r-10:17
[reed]means you need to get r?bsmedberg again10:17
asacnow i wonder if i need to rerequest review10:17
[reed]yes, you do10:17
[reed]since he gave you r-10:17
asacsure ... but didn't you say that one doesn't need to rerequest if its just a tiny modification that is addressed?10:18
asachttps://bugzilla.mozilla.org/show_bug.cgi?id=408062#c310:18
ubotuMozilla bug 408062 in Build Config "Provide short-hand includetype=unstable .pc files for libxul and libxul-embedding" [Enhancement,Assigned]10:18
[reed]correct, but that's only when the reviewer gives you r(eview)+10:18
[reed]he gave you r-, so he wants you to re-request review10:18
asacok thanks.10:18
[reed]once he gives you r+, then request approval1.9 and give a short comment on what the patch does and risk/reward10:23
asacso blocking1.9 would mean that i want to nominate a blocker ... which only makes sense if there isn't a patch? so if there is a patch one needs approval?10:23
[reed]no, you can nominate something for blocking at any poinmt10:23
[reed]point10:23
[reed]but you better make a good case as to why it should block 1.9 release10:23
asacyes thats clear10:23
[reed]basically, the requirements for landing a patch are either that the bug be blocking1.9+/blocking-firefox3+ or that the patch have approval1.9+10:24
[reed](and all appropriate reviews, of course)10:24
[reed]so, you only need one or the other, really10:25
[reed]bah10:25
[reed]I screwed up10:25
asacagain?10:25
[reed]I changed the assignee on you10:25
[reed]:)10:25
* [reed] fixed10:25
[reed]I moved it to XULRunner10:26
asache? qua contact you mean?10:26
[reed]well, the component10:26
[reed]it was in Toolkit :: Build Config10:26
asacwhich did i use before?10:26
[reed]I moved it to Toolkit :: XULRunner10:26
[reed]since it's XULRunner only thing10:26
asachmm its still build config here :/10:26
asacwell at least the drop down box shows it as selected10:27
[reed]shift-reload?10:27
asachaha10:27
asacstrange is that just reload updated the QA contact but not the combo box :)10:27
[reed]well, it makes sense, in some strange way10:27
asaci can imagine ;)10:27
[reed]in case you've made changes but want to reload the page for updates10:28
[reed]or something10:28
[reed]dunno, ask bz!10:28
asacright10:28
[reed]:P10:28
asac[reed]: now i have fixes for some other issues that overlap with the code o the patch above ... should i include the full diff when opening those bugs or an incremental one with Depends on: on that bug?10:31
[reed]incremental with depends on10:31
asacgood ... i have them because i use a patch system, but how would you do that with CVS ?10:32
asacmost likely you won't :)10:32
[reed]a lot of our devs use hg or git now for their personal repos10:32
[reed]they import CVS10:32
[reed]and diff against it10:32
asac[reed]: mozilla bug 37391810:33
ubotuMozilla bug 373918 in Embedding: GTK Widget "gtk_moz_embed_get_nsIWebBrowser not in the glue" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=37391810:33
asaci wonder if that is the right patch as its clearly a follow up now for the microb backout10:33
asachaven't found the right bug though for adding the missing symbols :(10:34
[reed]dunno... I know nothing about gtkmozembed :)10:34
[reed]file a bug if you don't see a dupe10:34
asacits a mess10:37
asacmozilla bug 40823810:37
asac:)10:37
ubotuMozilla bug 408238 in Embedding: GTK Widget "back out unreviewed microb changes." [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=40823810:38
[reed]well, that's what happens when stuff doesn't get proper review10:38
asacits completely unclear to me how things will proceed ;)10:38
[reed]:)10:38
[reed]well, maybe you should comment in mozilla.dev.platforms.linux ?10:38
asacthats true ... i wonder why it was checked-in ... e.g. did they just send him the cvs account without educating about checkin rules?10:38
[reed]who landed most of the stuff?10:38
asacno idea :) ... i think some nokia guy10:38
[reed]timeless or some other people?10:38
[reed]oh10:38
[reed]probably10:39
[reed]timeless was supposed to be watching him10:39
[reed]timeless failed10:39
asacah ;)10:39
asacwell if he just checks in there is nothing timeless could do :)10:39
[reed]http://bonsai.mozilla.org/cvsquery.cgi?branch=HEAD&dir=mozilla%2Fembedding%2Fbrowser%2Fgtk&date=explicit&mindate=2006-07-01&maxdate=2007-04-0110:40
[reed]looks like timeless landed most of the stuff from that branch he was using!10:41
[reed]so, we can all just blame timeless :)10:41
asache should have known better :)10:41
[reed]yes, he should have10:41
[reed]he knows we have high expectations for mozilla code10:42
asacmozilla bug 38430410:45
ubotuMozilla bug 384304 in Startup and Profile System "/usr/bin/firefox is not able to handle symlinks due to error in script" [Major,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=38430410:45
[reed]yes?10:46
asacreed i think that hasn't landed on 1.910:46
asacbut has approval10:46
asacat least we stiill have that patch in build-tree10:46
[reed]comment #710:46
[reed]landed10:46
asacwell ... the patch still applies :)10:47
[reed]http://bonsai.mozilla.org/cvslog.cgi?file=/mozilla/browser/app/mozilla.in&rev=HEAD&mark=1.1610:47
asacotherwise i would have got build failures i guess10:47
asacstrange10:47
asacoh damn10:47
asacits against mozilla/xulrunner/app/mozilla.in10:48
asacnow browser10:48
[reed]haha10:48
asacso how many mozilla.in files do exist?10:48
* [reed] looks10:48
asaci have a tree with xulrunner,browser,mail i guess ... if there is more i need to do a new checkout10:49
[reed]http://mxr.mozilla.org/mozilla/find?string=mozilla.in10:49
asacok calendar as well10:49
[reed]file a new bug that deals with the others?10:49
[reed]or reopen that bug10:49
[reed]and attach a new patch10:49
[reed]that fixes the others10:49
asaci reopen the bug and attach a patch for the rest10:49
[reed]k10:50
* asac checks out ,+calendar10:50
[reed]I just have a checkout of all of HEAD10:50
[reed]:)10:50
[reed]that I use for committing10:50
asacmozilla bug 34481810:51
asacall?10:51
ubotuMozilla bug 344818 in Build Config "Linking - missing library deps" [Normal,Resolved: wontfix] http://bugzilla.mozilla.org/show_bug.cgi?id=34481810:51
[reed]yes10:51
[reed]all :)10:51
[reed]looks like vlad didn't understand10:51
[reed]reopen the bug10:51
[reed]say which patch is against 1.910:51
asacyeah10:51
[reed]re-request review10:51
asaci'll do10:52
asaclet me check if the patch is complete10:52
[reed]reed@jarodplus:~/mozilla/HEAD/mozilla$ du -hs .10:52
[reed]719M    .10:52
[reed]repo isn't too big10:52
[reed]:)10:52
[reed]reopen the bug10:53
[reed]and request review from vlad specifically10:53
[reed]not the wind10:53
asacyes done10:53
[reed]k10:53
asacmozilla bug 38661010:54
ubotuMozilla bug 386610 in XPCOM "pyxpcom fails to build against python 2.5 (api changed)" [Normal,Assigned] http://bugzilla.mozilla.org/show_bug.cgi?id=38661010:54
asacdo i need approval for that even though its in a n unsupported component?10:55
[reed]no10:55
[reed]if it's not part of the normal Firefox build10:55
[reed]you don't need approval10:55
asacso how do i ask for landing ... keyword need-checkin ?10:55
[reed]checkin-needed10:55
[reed]keyword10:55
asacthanks10:56
asacgood10:56
asacbug 40463410:56
asacmozilla bug 40463410:57
ubotuMozilla bug 404634 in Embedding: GRE Core "http urls don't render in TestGtkEmbed" [Critical,Assigned] http://bugzilla.mozilla.org/show_bug.cgi?id=40463410:57
asacok10:57
[reed]that's blocking1.9+10:58
[reed]so, once it gets review10:58
[reed]you can just add the checkin-needed keyword10:58
[reed]no approval needed10:58
[reed]got it? :)10:58
asacyeah ... i am learning :)10:59
asachow the hell can i file a bug directly against xulrunner installer10:59
asaci cannot find it in the new bug wizard :(11:00
[reed]Toolkit :: XULRunner ?11:00
asachttps://bugzilla.mozilla.org/enter_bug.cgi ?11:00
asacwhere from there?11:00
[reed]what do you see on that page11:00
asacah11:00
asac:)11:00
asacENOTENOUGHCOFFEE11:00
[reed]lol11:00
[reed]for 404634, you might want to request superreview from bz11:01
[reed]just to speed stuff up11:01
[reed]I don't know if bsmedberg is a valid docshell reviewer11:01
* [reed] looks11:01
[reed]oh11:02
[reed]he is11:02
[reed]he owns it11:02
[reed]n/m!11:02
armin76!11:02
[reed]http://www.mozilla.org/owners.html#docshell11:02
[reed]:)11:02
armin76asac: did you read what i told you yesterday?11:03
asac [reed] so i need superreview?11:03
[reed]not really... bsmedberg is a super-reviewer11:04
asacarmin76: system hunspell?11:04
armin76yeah11:04
[reed]I'll take his r+ as enough11:04
[reed]:)11:04
asac[reed]: so is it a bad idea to ask him directly for review or what?11:04
[reed]no11:04
[reed]you did the right thing11:04
asaci usually end up asking him and not his peers11:04
[reed]well, you basically want to pick who-can-review-it-the-quickest11:04
[reed]but-still-give-me-a-good-review11:05
[reed];)11:05
asacwhich is hard to say :) ... he has been more or less responsive for me11:05
[reed]yeah, he's a good choice11:05
[reed]he's pretty quick11:05
[reed]but it's been the holidays11:05
[reed]so, he probably hasn't looked at patches lately11:05
[reed]you might could ping him in #developers on moznet tomorrow11:05
asacyeah ... i am not that good on following up either ... so i have no point in blaming him for slowness anyways :)11:06
[reed]hehe11:06
[reed]true11:06
asacall this porting of embedders sucks up a lot of time11:06
asacwe are the first that do all this ... fedora just did some basic porting ... so it ends up all here ;)11:06
[reed]hehe11:07
[reed]well, go ubuntu11:07
[reed]I'll land the python xpcom stuff later today11:07
[reed]it's 5:08am now11:07
asacthanks11:07
[reed]I should sleep :)11:07
asacoh ... you are on the other side ... right :)11:07
asacgood night!11:07
[reed]hehe11:07
[reed]yeah, I'm in the USA11:08
[reed]but I'm a nightowl11:08
[reed]stay up late11:08
[reed]hehe11:08
asacmy days have shifted to the dark side as well11:08
asac:)11:08
asacarmin76: yes, that might be true11:09
asacarmin76: well it is true ;)11:09
asacdid you run into issues with that?11:09
asacarmin76: ?11:21
armin76errr...not sure11:23
armin76i'm using external hunspell atm using a patch, but doesn't seem to pick up the dicts11:23
armin76it doesn't work if i put the dicts in ${ffdir}/dictionaries11:24
armin76using internal works fine, though11:24
asachmm11:31
asachave that patch?11:31
armin76http://overlays.gentoo.org/proj/mozilla/browser/www-client/mozilla-firefox/files/100_system_myspell-v2.patch11:38
armin76asac: ^11:39
* asac looking11:39
asacwhich hunspell version?11:41
asac(do you ship in system)11:41
armin761.1.911:42
asactrunk has: "Update Hunspell to version 1.1.12, patch by Ryan VanderMeulen, r+sr=mscott, a=dsicore "11:43
asacso maybe its too old?11:43
asacoh right11:44
asachmm11:44
asachttp://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&subdir=mozilla/extensions/spellcheck/hunspell/src&command=DIFF_FRAMESET&file=mozHunspell.cpp&rev1=1.3&rev2=1.4&root=/cvsroot11:44
asacdo you already have that patch in your trunk sources?11:44
asacotherwise it would be loaded from /usr/lib/xulrunner-1.9/dictionaries ... but since it works for you with in-source hunspell i guess that its the version mismatch11:45
armin76nope11:45
armin76probably11:45
=== \sh_away is now known as \sh
armin76will try11:59
=== \sh is now known as \sh_away
armin76asac: this is stuppid14:58
armin76stupid14:58
armin76wtf ff3 builds with MOZ_XUL_APP=1 by default?14:59
armin76unless i force it in a patch...15:00
armin76nope15:00
asache?15:00
asacarmin76: why is that wrong in your opinion?15:01
armin76dunno *g*15:02
armin76asac: thing is, MOZ_XUL_APP is set 1 when using --with-libxul15:03
armin76however it's already the default15:03
armin76isn't MOZ_XUL_APP when ff is built against xul?15:04
asacno ... it just means that its a XUL_APP15:05
asacthat applies even if its in-source15:05
asacxul15:05
asacdo you run into issues because of that?15:06
asacarmin76: ?15:06
armin76http://mxr.mozilla.org/mozilla/source/extensions/spellcheck/hunspell/src/Makefile.in15:06
armin76look at line 6615:06
armin76thing is, with the patch i told you(which is not mine), it'll only load the myspell dicts if MOZ_XUL_APP is not set15:08
asacis the spellchecker an external component or is it in libxul?15:09
armin76no idea :/15:09
asacoh maybe DirProvider is buggy?15:09
armin76if MOZ_XUL_APP is not set, it uses mozHunspell.cpp, but if it's set it uses mozHunspellDirProvider.cpp15:10
armin76however in the patch, the modified thing to load dicts from /usr/share/myspell is on mozHunspell.cpp15:10
asacarmin76: it looks in a different directory15:11
asactry to link the dictionaries directory in $XUL_DIR/extensions/15:13
asace.g. $XUL_DIR/extensions/dictionaries/15:13
asaci still don't get whyt that behaviour would change if you use system hunspell15:17
armin76nope, nothing15:18
asacarmin76: strace -f -eopen ... then grep for dictionaries :)15:35
asacit should try some dirs15:35
armin76it doesn't15:38
armin76at least with 'dic' 'spell'15:38
asache?15:38
armin76it doesn't try to open anything with 'spell' or 'dic' in the name15:38
asacso there is no entry with dictionaries in the path?15:39
armin76nope15:41
Some_PersonIs there a package that makes the Totem plugin work with Firefox 3 Beta 2?15:51
asacnope ... i could provide it though. lets see.16:05
Some_Personthat would be awesome16:06
asacSome_Person: ok in an hour or so you can get the new totem packageset from mt ppa16:16
Some_Personthank you very much16:16
asachttps://launchpad.net/~mozillateam/+archive16:16
asac(hardy)16:16
Some_Personoh, its for hardy?16:17
asacyou need gutsy?16:17
asacSome_Person: ?16:19
Some_Personyeah, i use gutsy16:19
=== \sh_away is now known as \sh
asacSome_Person: then please test the firefox-3.0 and xulrunner-1.9 in the mt archive16:19
asacthe apt lines you can get by using the combo box16:19
Some_Personthat will work with totem? because right now im using the build from mozilla16:20
asaci will try to upload the same package to gutsy ppa ... lets see16:20
Some_Personworth a shot16:20
asachmm16:20
asacyeah?16:20
asacgood16:20
asacplease install it16:20
asacand tell if it works ... its beta216:20
asacif it doesn't work it doesn't make sense to push up a totem try16:21
Some_Personok, i'll try it16:21
Some_PersonAre your packages supposed to be "not authenticated"?16:24
asacyes ppas unfortunately are not authed yet16:24
Some_Personthats ok, as long as it works, i dont care16:25
Some_Personok, i installed the PPA firefox 3 package16:28
Some_Personbut no plugins are installed16:28
Some_Personah, i see, your link to the plugins folder points to the wrong place16:31
asache?16:32
asacwhat do you mean?16:32
Some_Personthe link to the plugins folder in your package points to '/usr/lib/firefox-addons/plugins'16:32
asacwelll linking to the old plugins folder is wrong because some might crash new ffox16:32
asacthat is a link to xulrunner-addons, right?16:33
Some_Personit links to /usr/lib/firefox-addons/plugins16:33
asacbut that is a link as well?16:33
asacor just a dir?16:33
Some_Personbut shouldn't it link to the old one?16:33
asacno16:33
Some_Personas thats where ubuntu stores its plugins16:33
asacread above16:33
asacplease tell me if /usr/lib/firefox-addons/plugins is a link on its own16:33
asacso i can prepare the proper totem16:34
Some_Personits a folder16:34
asacok thanks16:34
asacthe other plugins just need to be linked into that new dir (if you want them)16:34
Some_Personyeah, and i just did that16:34
asactotem unfortunately makes use of xul symbols and thus needs to be rebuilt16:35
Some_Personso it uses symbols that are incompatible with ff3 so it crashes, correct?16:36
asacyes16:36
asacother plugins should work fine16:36
Some_Personthey do, i just checked, flash and java are working16:37
asacwill take a while ... my gutsy chroot is completely outdated ... takes ages to upgrade16:52
Some_Personthats ok, thank you very much for doing this by the way16:54
Some_Personi think you'll please a lot of people using FF3 in gutsy16:55
asaclet me know if you have issues with the ppa package16:55
asacthat package is a test before upgrading the one in gutsy16:55
Some_Personok16:56
Some_Personthe ff3 package seems fine16:57
asacSome_Person: uploaded a totem package ... one hour or so and you should be able to test17:59
Some_Personok17:59
asacok off19:08
asacSome_Person: the package is built19:10
Some_Persongreat19:10
* armin76 builds asac19:11
asacarmin76: no! ... i just wanted to bail out ;)19:11
asac:-P19:11
asacok reall out for a while19:11
Some_Personok, im downloading it19:12
Some_Personit works!19:14
Some_Personthank you thank you very very much19:16
armin76asac: got it19:52
armin76http://bonsai.mozilla.org/cvslog.cgi?file=/mozilla/extensions/spellcheck/hunspell/src/mozHunspell.cpp&rev=HEAD&mark=1.419:52
armin76the latest change is what broke the patch19:53
asacarmin76: that code looks good to me19:58
armin76yeah, but maybe with the patch for external myspell it changes something19:58
armin76let me show you...19:58
asacNS_XPCOM_CURRENT_PROCESS_DIR == "/usr/lib/firefox-3.0/" .... NS_GRE_DIR=/usr/lib/xulrunner-1.919:58
armin76http://rafb.net/p/oLRNOH45.html20:05
armin76that's the current code + external patch20:05
=== \sh is now known as \sh_away
asac[reed]: any idea why mozilla bug 408791 has no switch to request blocking1.9?20:11
ubotuMozilla bug 408791 in Password Manager "nsLoginManagerPrompter.js does not work with TestGtkEmbed prompt." [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=40879120:11
asac[reed]: maybre its the wrong component? e.g. use Toolkit -> Security instead?20:13
Ubulette /usr/lib/firefox-3.0-3.0b3pre/firefox-3.0: relocation error: /usr/lib/xulrunner-1.9b3pre/libxul.so: symbol SEC_StringToOID, version NSS_3.12 not defined in file libnss3.so.1d with link time reference20:21
armin76asac: thing is it never loads the dicts because the loaddicts function is inside the else20:23
armin76before that change, it was outside20:23
armin76so a small hack would be load the dicts inside the if20:26
asacarmin76: have you tried to link the dictionaries dir in the xul dir?20:32
asacarmin76: look at the code path ... if anything is successful it would load the dictionaries20:35
asacarmin76: maybe the install now has DICTIONARY_SEARCH_DIRECTORY ... and thus the GRE + APP dirs are not tried?20:35
asacarmin76: you should remove the whole else block ... e.g. the gre+app dir code should be always run .20:53
asacarmin76: and finally i understand your problem ... if you use system-xul the spellchecker will be built without the dirprovider ... if you use in-source xul it will be built with the dirprovider ... all this can't be right20:54
armin76yeah, i dropped the else block20:54
asacdon't you use system-xul?20:55
asacits strange anyway ... from what i understand Get(DICTIONARY_SEARCH_DIRECTORY ... should always fail20:56
asacat least in mozHunspellDirProvider.cpp there is no match for that key20:56
asacGetFile -> always NS_ERROR_FAILURE20:56
armin76nope i don't, i have to look at it20:56
asacGetFiles -> always NS_ERROR_FAILURE except for DICTIONARY_SEARCH_DIRECTORY_LIST20:57
armin76this patch is derived from debian's myspell one for ff20:57
armin76220:57
asacarmin76: ok so you should probably just remove the ifdef MOZ_XUL_APP blocks from the myspell patch you have20:59
[reed]asac: that's because it's under Firefox21:06
[reed]so, it's blocking-firefox321:06
[reed]instead21:06
asac[reed]: ok, thats what i thought ... now i just wonder if it really belongs there ... its toolkit/.*/passwordmgr code21:12
[reed]yes21:14
[reed]bugzilla is lame, so it's a huge pain to move components21:14
asacok21:14
[reed]yay for asac requesting approval21:36
[reed]go asac21:36
[reed]go asac21:36
[reed]:)21:36
armin76go me21:43
asacbug 17236821:49
ubotuLaunchpad bug 172368 in firefox "Downloaded the update from "Main Server" for Firefox. Now some sites (eg. vodacom4me.co.za) dont display as before. Cannot log in to this site now." [Undecided,New] https://launchpad.net/bugs/17236821:49
asac[reed]: yay ... you can even checkin another now: mozilla bug 40463421:50
ubotuMozilla bug 404634 in Embedding: GRE Core "http urls don't render in TestGtkEmbed" [Critical,Assigned] http://bugzilla.mozilla.org/show_bug.cgi?id=40463421:50
[reed]yep21:51
asacpainful as it was ;)21:51
[reed]as long as you add checkin-needed keyword, I'll get to them soon21:51
asacsure21:51
armin76asac: http://rafb.net/p/iE0yx866.html22:09
armin76that's the definitive patch22:09
armin76wfm, feel free to fix :)22:09
asacarmin76: so thats broken?22:14
asacarmin76: the code in line 123-126 looks wrong22:16
asacat least for us its /usr/share/myspell/dicts :)22:16
asacarmin76: thats broken ... i mean you try to open myspell directory ... and then you don't try anything else22:17
asacbecause that dir exists22:17
asac(well except DICTIONARY_SEARCH_DIRECTORY_LIST which is probably empty)22:17
asacarmin76: anyway ... imo that hunk is not needed anymore22:18
asacyou just install a proper link to the right directory: done!22:18
[reed]asac: mozilla bug 409803, comment #922:36
ubotuMozilla bug 409803 in Build Config "gcc zealously avoids inlining at -Os" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=40980322:36
[reed]please comment?22:36
armin76asac: in gentoo they get installed in /usr/share/myspell23:35
armin76feel free to modify, it works fine for me23:35
armin76asac: http://dev.gentoo.org/~armin76/spell.png23:38

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!