[reed]ah, I'm feeling much better04:08
[reed]my flight has been fixed04:09
jtvasac: ping07:40
=== asa2 is now known as asac__
asac__fta: i can't reproduce the crash with a normal build tree using -libxul-sdk for whatever reason. will see if its because xul is build with debug10:03
asac__illl do another build with optimize and not debug10:05
asac__asac: are you really still here? you should be dead for some time now10:08
DarkMageZasac__, it doesn't reply to version requests. just ghost it with nickserv10:14
asac__DarkMageZ: huh?10:15
asac__the point is that my gateway is dead right now10:15
asac__but still this nick running on it is here :)10:15
asac__and the nick disappeared from other nets i am on a few minutes ago10:16
DarkMageZfreenode has a long timeout period.10:17
asac__ok lets wait a bit longer10:17
asac__ok at least that works :)10:21
asac__fta: i cannot reproduce crash ... maybe its your nss/nspr things that break this?10:40
asac__ok trying to build package now to see if that makes a difference10:41
asacok i am back in this thing11:03
asaclets hope my provider and power provider stay up for the next 10 hours or so11:04
asac(both went down)11:04
asac__logging this sidetrack out11:05
asaceverything lost in irssi backlog ... if you asked soemthign in the last 12 hours repost11:06
armin76asac: <jtv> asac: ping <- at 07:01 GMT11:27
armin7607:40, not 01 :P11:27
asacjtv: yes?11:28
jtvasac: hi11:32
jtvasac: I was just wondering about the clashing message keys we discussed11:32
jtvasac: right now afaics, if a key occurs in two files, you still see only one file listed.11:32
jtvasac: I could, in principle, change that so that messages with the same key and the same English text were shown once, with multiple occurrances.11:33
jtvasac: but that might screw up your parser.11:33
asacfta: what i find strange is that i could build latest cvs trunk checkout with current nss, but now can't when trying to build the xul 1.9 package from bzr11:42
asacjtv: hmm. why is it shown only once?11:43
asacjtv: i thought we add a msgctxt now11:43
asacso they should appear with same #: comment, but with different msgctxt11:43
asacor did i get that wrong?11:43
asacor do you mean for cases hat are not because of win/mac arch (which is a special case here)11:44
asacaeh ... hard to parse i guess11:44
asacwhat i said11:44
asacjtv: now i am a bit confused. what was the status quo of our discussion again?11:46
asacit was to add a special case for platforms, and include a msgctxt to differentiate those, right?11:47
jtvasac: right11:47
jtvasac: I didn't mean to confuse you...  For every message in the exported file, you get exactly one comment showing the file name that the message came from, right?11:48
jtvasac: (BTW I think I lost the link to that sample template you got from Carlos... where is it again?)11:48
asacjtv: i received it by mail11:49
jtvasac: Well, since we can have the same message in multiple files, I could in principle try to unify all versions of that message that have the same key and the same English text.11:49
jtvasac: in principle (though I'm not sure how far this would go in practice with current code) we could present the truly identical versions of the same message key as a single translatable string.  It'd save some effort.11:50
jtvasac: but for now I think I won't work on that.11:50
jtvasac: for now I'm concentrating on providing disambiguating context.11:51
asacjtv: ok. i don't think its required now11:53
asacjtv: lets keep things simple so we can get something at least11:53
jtvasac: that's what I'm doing.  :)11:53
jtvasac: the idea came up because I was still exploring what changes would be needed.11:54
asacjtv: well. in cases there is a id clash that is dealt with by contexting this would be useful12:02
asacbut i hope its not required :)12:03
* asac lunch time12:28
fta2asac, what ? how does it fail ?12:39
armin76asac: fta2: you aren't using hunspell :P12:45
fta2eh ?12:48
fta2sure i am12:48
fta2i have  --enable-system-hunspell12:48
fta2and i'm using it12:49
armin76on firefox you aren't12:49
armin76--enable-system-myspell \12:50
fta2it's not needed12:50
fta2fta@cube:~ $ lsof -p 10826 | grep huns12:50
fta2firefox 10826  fta  mem    REG        8,1   204416   541758 /usr/lib/libhunspell-1.1.so.0.0.012:50
armin76then may want to remove that line :P12:51
fta2I don't have it12:51
ubotuMozilla bug 423334 in XPCOM "crash at startup in [@ NS_CompareVersions] when using --with-libxul-sdk" [Critical,New]12:51
fta2that's firefox-3.0.dev, i'm talking about *.head12:52
armin76oh :)12:53
armin76its still there :P12:53
armin76line 11612:53
fta2oh myspell12:54
fta2but anyway, i have no problem to build trunk or the .head branches12:54
armin76i know12:55
fta2asac, so what's your problem ?13:01
asacfta2: my problem is that xulrunner head doesn't build against current nss, while my trunk checkout does :)14:02
asacnevermind :)14:03
asacfta2: ok with the leaking parser i don't see a crash14:47
asacfta2: i get the feeling that your system is special :)14:51
asac(which isn't bad in this case)14:51
asacfta2: how do you built your local packages?15:00
asac(you see the crash there as well, right?15:00
asacfta2: the interesting thing is that you build with symbolic-functions15:04
asaci don't do that and don't see any crash15:04
asac(at least with the leak)15:04
asacnow i wonder how it happens that you pull in those default LDFLAGS15:04
asacwhat kind of builder setup are you running?15:04
asac(all assuming that you use those options on your local build as well ... i only can see the logs from ppa)15:05
fta2check your config.status after the build, it should match mine15:23
fta2cdbs adds some parameters by itself15:23
asacits dpkg-buildpackage apparently15:24
asacbut bzr bd doesn't use that - reason unknown15:24
asacdebuild does neither ... so i suspect that its debuilds fault in the end15:24
fta2i build with bzr bd15:25
asacdo you see the same flags locally?15:26
fta2fta2 crashes, it's the last one from yesterday15:30
asacyes. i think its that flag15:30
asacfor me its not used and i don't see the crash. that matches the fact that redhat has no problems. they are unlikely using this.15:30
asacfta2: ok its obvious why i don't see it ... i use debuild with bzr most of the time, because i usually pass: --builder='debuild -b'15:32
fta2i don't15:34
fta2bdm = bd --merge --dont-purge15:34
fta2bdn = bd --native --dont-purge15:34
fta2ppa = bd --merge --build-dir=../ppa --builder='dpkg-buildpackage -rfakeroot -S -sa -kB6EE20E8'15:34
fta2ppa2 = bd --merge --build-dir=../ppa --builder='dpkg-buildpackage -rfakeroot -S -sd -kB6EE20E8'15:34
fta2ppan = bd --native --build-dir=../ppa --builder='dpkg-buildpackage -rfakeroot -S -sa -kB6EE20E8'15:34
asactry debuild15:35
asaci try dpkg-buildpackage after the rebuild without any patch is finished15:35
asacto see that it really crashes15:35
fta2no, i prefer to use what the ppa builders use15:35
asacthis is to test if your crash goes away15:36
fta2otherwise it will work for me but not for the others15:36
asacnot to test that you change your default setup15:36
asacif it goes away we can fix it15:36
fta2I can just set LDFLAGS15:36
asacyou can also wait till my dpkg-buildpackage build is done (which i just started)15:37
asaci think 28 min :)15:39
asac \o/ :)15:39
asacthe crash is triaged :)15:39
* asac happy before anything is done ;)15:39
armin76did you guys saw that?15:50
ubotuOpenOffice.org bug 86670 in tools "system-mozilla: libxul and new pkgconfig files" [Task,Resolved: fixed]15:50
asacarmin76: yes we have something similar15:55
armin76on openoffice or xul?15:56
armin76because oo is using libdir15:56
armin76with that patch, that is15:57
armin76MOZ_LIB=`$PKG_CONFIG --variable=libdir libxul`15:58
armin76i don't have libdir on my pkgconfig files...15:58
fta2asac, I have the commit ready and i'm building but you will be done before me15:58
asacu never know16:01
asacarmin76: then they did wrong16:02
armin76thats why i'm telling you :)16:02
asacarmin76: for 1.9 you must not need rpath16:02
asacarmin76: but maybe it gracefully does what we want if libdir is empty?16:02
asacwhat is done with MOZ_LIB16:03
asacyou see that?16:03
armin76don't think so16:03
armin76a guy just asked me to add libdir16:03
asacits obviously not in the patch16:03
asacno thats not needed16:03
armin76i know16:04
asacits most likely that the yuse rpath16:04
asacbut libxul is dependent and thus loaded in an already bootstrapped env16:04
asacarmin76: look in the ubuntu ooo packaage ... there hsould be a patch16:04
armin76link? :)16:07
asacfta2: just xulrunner didn't cause the crash16:11
asacnow trying ffox16:11
asacon top16:11
asacarmin76: i am not sure ther eis a link :)16:12
asacthe smallest you can get is the diff.gzu and apply the debian/ part against an empty directory16:12
fta2xulrunner alone never crashed16:13
asacarmin76: https://edge.launchpad.net/ubuntu/hardy/+source/openoffice.org/1:2.4.0~rc2-1ubuntu3/+files/openoffice.org_2.4.0~rc2-1ubuntu3.diff.gz16:14
asacfta2: xulrunner alone with symbolic-functions i mean16:14
asacusing the old ffox16:14
asacstill no crash. i am out of ideas right now16:15
asacjtv: i found one glitch that might be good to get fixed on lp side16:33
asacjtv: http://paste.ubuntu.com/5939/16:33
jtvasac: looking...  btw just implemented the context trick.16:34
asacthanks ... that is a pattern that is regularly used to include another .dtd16:34
jtvasac: you had that in a message you showed me the other day, IIRC.  Is it something we generate?  Something we should handle?16:34
jtvAh ok16:34
asacnow i am thinking about how we can do that without too much work16:35
jtvasac: If the other file has a name ending in .dtd, it will be imported anyway and no special handling is in order.16:35
asacjtv: well... the point is that we don't really have that information :)16:35
asacatm we only get en-US.xpi exported16:36
asacfurther it would be pretty hard to get the info if we get de.xpi alongside for example16:36
asacjtv: can this be handle as a translation?16:36
asace.g. for the special key: "% realBundDTD SYSTEM" ?16:37
asacthe translation parser could see that and produce the right pattern from it16:37
jtvasac: I'm afraid I don't follow16:37
jtvHow does this syntax work exactly?16:38
asacok ... the normal pattern for translatable entities looks like:16:38
asac<!ENTITY size.label             "Size">16:38
asacthats mapped to the .po file like16:38
asac#: .... (size.label)16:39
asacmsgid "Size"16:39
asac... and so on16:39
asaclet me think before i go on16:39
asacyes right16:39
asac<!ENTITY % something SYSTEM "chrome://...">16:40
asacwould be mapped to16:40
asac#: .... (% something SYSTEM)16:40
asacmsgid "chrome://...."16:40
asacjtv: understood?16:40
jtvBut normally it includes another file?16:41
asacif our parser sees a entity starting with "% " it could create the &something;16:41
asacjtv: right, but you cannot find that without parsing chrome.manifest16:41
jtvSo you want us to spit it out as a translation message, so your parser can handle it.16:42
jtvOr a special comment maybe?16:42
asacevery file we cannot produce from .po will be copied from en-US.xpi16:42
jtvAnd you want to be able to reconstitute the reference.  Gotcha.16:43
jtvThis could take me some time, because I'd have to figure out a lot more about how the DTD parser returns its data to us.  :/16:43
asachmm ... you use a real dtd parser that auto resolves entitites?16:44
asaci assumed you had something hand written :)16:44
jtvWe are not the knights who say NIH16:46
asacjtv: thats good ... but how do you resolve that entity then?16:48
jtvasac: that's the problem...  I don't know.  It's something the standard dtd parser does, and I have zero experience with DTD.16:49
asacor do you ignore failed entity resolves?16:49
asacjtv: yep ;)16:49
jtvSo you may have to grep for this pattern...16:50
asacfta2: no crash for me16:55
asacfta2: do you see that on i386?17:00
asaci really don't see that17:02
asacfta2:  you have an strace ?17:02
asacfta2: ok stupid me again ;)17:05
asacits all fine17:06
asaci ran my non symbol-functions dev-tree :)17:06
asacwith symbolic-functions it crashes17:07
fta2seems ok here too17:08
* asac happy that he saw symbolic-functions in the log :)17:11
asaci always knew that we would get struck by that at some point17:12
fta2the question is now, is the leak patch needed ?17:12
asaci don't think that kind of hack is needed, no17:13
fta2i've updated the case17:18
fta2giving you credits17:18
fta2mozilla bug 42246317:18
ubotuMozilla bug 422463 in Build Config "configure fails on x86_64 if pyxpcom is enabled" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=42246317:18
asacfta2: please let me know if they complain on the bug or something17:25
asacill try to push for something like timeless suggested i guess.17:27
armin76trunk now needs sqlite-3.5.717:28
armin76you guys have it?17:31
armin76!status sqlite17:31
ubotuSorry, I don't know anything about status sqlite - try searching on http://ubotu.ubuntu-nl.org/factoids.cgi17:31
armin76or whatever it was the command :P17:31
armin76!info hardy sqlite317:31
ubotuPackage hardy does not exist in gutsy17:31
armin76!info sqlite3 hardy17:31
ubotusqlite3 (source: sqlite3): A command line interface for SQLite 3. In component universe, is optional. Version 3.4.2-2 (hardy), package size 19 kB, installed size 84 kB17:32
asacyes sqlite is prone to break things on major version upgrades ;)17:33
xhakerarmin76: i think sqlite is static linked into firefox317:38
xhakeroops, libnss-1d depends on the version in the archives17:39
[reed]fta2: stop reseting the blocking flag!17:53
[reed]yes, again17:53
asacfta2: everytime we have to get blocking flag back, someone quite busy has to be come around :) ... please remember that and be careful :)17:54
asacbut i assume it a mouse wheel accident17:55
fta2[reed], how is that possible? it's not a mid-air thing this time.17:57
fta2I've added a comment to a freshly loaded page17:57
[reed]mouse wheel, probably17:57
fta2no, i never clicked on the form in this session17:58
fta2and I don't get notification for my own changes so I don't know17:59
fta2well, i could stop updating this bug if you prefer17:59
[reed]no, but something isn't right on your side18:02
armin76fta2: you can setup the notifications stuff18:47
armin76uh, did they move back the home button?19:01
armin76mozilla bug 42242019:02
ubotuMozilla bug 422420 in General "Revert home button move and related migration code" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=42242019:02
[reed]well, didn't move it back, but the move code got reverted19:02
[reed]if your profile already had it moved, you'll have to move back yourself19:02
armin76you can't set a preferred homepage now on the default settings?19:04
armin76i'll try with ubuntu :P19:05
fta[reed], it works without the leak patch too...19:50
ftai now wonder why -Wl,-Bsymbolic-functions is used by default in ubuntu19:52
asacfta: it gives performance improvements and is considered safe for most cases.20:07
asacunless you do something like this :)20:08
ftaso what now. moving jemalloc into a .a was also for a performance improvement reason20:10
ftawe trade an improvement for another20:10
asacwell. i will comment on the bug (unless it already evolved)20:10
asacmoz should link jemalloc into the binaries - as timeless suggested20:11
asacthats sound20:11
ftajust let me add that it work without the patch too as i said i will, then go on20:11
asacsure. thought that was already said20:12
asacthat it works without the leak patch ;)20:12
ftareally ? I didn't20:12
ftachecking the flags before i hit submit...20:18
ftadouble checking20:19
ftadid I drop something ?20:19
asacwhat do you mean? uncommit?20:21
ftano, the flags on bonsai20:25
asacbugzilla you mean?20:40
asaclaunchpad codebrowse is unbearable slow, if not completely locked up20:40
ftaworks for me21:04
ftai find the ppa browsing much slower21:05
asacmozilla bug 40462721:44
ubotuMozilla bug 404627 in DOM "[FIX]XPinstall whitelist bypass using refresh after fix for bug 402649" [Normal,Verified: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=40462721:44
asacmozilla bug 37207521:45
ubotuMozilla bug 372075 in DOM "javascript: URI evaluation should use sandboxed context for toString, etc" [Major,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=37207521:45
asacmozilla bug 28266021:45
ubotuMozilla bug 282660 in JavaScript Debugger "Crash [@ jsds_NotifyPendingDeadScripts] ds->script is null" [Critical,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=28266021:45
asacmozilla bug 39929821:45
ubotuMozilla bug 399298 in Security "Bypassing XPCNativeWrapper by redefining XPCNativeWrapper" [Normal,Verified: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=39929821:45
asacmozilla bug 39742721:45
ubotuMozilla bug 397427 in Style System (CSS) "[FIX]Stylesheet href property shows redirected URL unlike other browsers" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=39742721:45
asacmozilla  bug 41935021:45
ubotuMozilla bug 419350 in XPCOM "[ia64] build failure using gcc 4.3" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=41935021:45
armin76guys, any thoughts why http://dev.gentoo.org/~armin76/ff3srio.png looks so ugly? browse with your ff and you'll see if cool22:17
asacarmin76: looks ugly yes22:20
asacbad webdesigner22:20
armin76see it in your ff322:20
ftaarmin76, yours is ugly indeed22:27
* fta blames armin76 22:28
Jazzvaasac: I came back :)...23:05
asacJazzva: hey23:06
JazzvaIt feels good when exams are over :)23:07
JazzvaSo, I ran the script that checks for new extensions that are not added to gnome-app-install... I noticed some new packages there. Should they be added?23:08
asacJazzva: yes. that would be a good thing23:10
asachowver, i would first like to add a few more extensions23:10
asacthere are already a bunch new extensions in the pipeline23:10
asacin code.launchpad.net/firefox-extensions23:11
asaci did a session about packaging extensions and wrote a simple helper .mk file to make writing debian/rules trivial for most extensions23:11
asacthat can be found at https://wiki.ubuntu.com/MozillaTeam/Firefox3Extensions/Packaging23:11
asacthere is a page where we add extensions we are working on .. or suggest for addition :)23:12
JazzvaOk, I'll check them out now :)23:13
asacif you have more xtensions in mind :) ... fill them in23:13
asacanother i pointed that has proper licensing is all-in-one-sidebar23:17
asacits compatible with ffox 3 afaik23:17
asacbut there should be plenty more23:18
JazzvaWell, I checked the extensions I use and I have found three that have appropriate licenses and work with FF 3.0b5pre (Greasemonkey, Better Gmail 2, Bettersearch). I suppose that Better Gmail is also licensed under GPL. I'll add them to the list now. Pmog and Stumbleupon works with FF, but I don't think they're licensed under any of mentioned licenses... I can't find any info so far.23:34
Jazzvaasac ^23:35
asacJazzva: greasemonkey has a package in the archive that needs to be upgraded23:35
asacat best its redone using the new xpi.mk23:36
asacbut is a bit trickier as you need to process .idl files during build.23:36
JazzvaWhat are .idl files?23:37
asacxpcom interface definitions23:37
asacxpcom component are cross language. e.g. you can call them from javascript and C++.23:38
asacso you need a language independent description of such components. thats basically what idl is :)23:38
Jazzvaaham... I see. So, could I find how to process them in debian/rules in the old package :)?23:39
JazzvaGreat :)23:39
asaci think you have to call all using the xpidl compiler in the xul dev package23:39
JazzvaOk, I'll just go through the rules file and see if I can see what it's doing.23:41
asacJazzva: the right path to run this is:23:41
asac`pkg-config --variable=sdkdir libxul`/bin/xpidl23:41
asace.g. its not in /usr/bin/ anymore since xulrunner-1.923:41
asacwell ... its just to get the xulrunner-1.9 sdk dir23:42
asacdo echo `pkg-config --variable=sdkdir libxul`/bin/xpidl23:42
asacto see what i mean :)23:42
Jazzvai see... it's in /bin :)...23:43
asacyeah ... but the version is in the path so you cannot use a fixed path23:43
Jazzvamhm... Ok23:44
JazzvaSo, is it ok to add better gmail (2) to the list, as they're not in the archives?23:45
ftais there a gnome applet to see the cpu temp ?23:45
Jazzvai think there is23:45
Jazzvai saw it on some screenshots23:46
ftado you know the name ?23:46
Jazzvajust a sec...23:46
JazzvaYou should install lm_sensors and gnome-applet-sensors in order to do that, afaics...23:47
Jazzvai'm not sure if those are the right names23:48
Jazzvabut they should be similar to that...23:48
Jazzvalm-sensors and sensor-applet23:49
Jazzvanope, better gmail is compatible only up to 3.0b223:53

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