ftaasac, http://code.google.com/p/chromium/issues/detail?id=2644800:05
asacfta: whats that?00:32
asaci mea n... in hardy we have a recent enough nss right?00:32
asacdoes this really happen in our built?00:32
asacwe generate the lower version through symbols00:33
asacif new symbols are not used, then yeah00:33
ftai'm not sure, people with vanilla hardy? or broken symbols?00:33
asacyes. but vanilla hardy is not supported imo00:35
ftai've just been Cced to the bug, i don't any additional information00:36
asacyeah. but you said you will bump requirements ;)00:37
asacwhich i think is not needed. users that install vanilla hardy ... dont get fixes/updates ... so chrome is broken :)00:37
asacthats how it is ;)00:37
asacthe fix for that is available in -security00:37
asacand -updates00:37
asacand all real life hardy users should have that00:37
asaceven downstreams should at least track -security00:38
ftaplease add that to the bug00:53
ubottuHere I am, brain the size of a planet and you expect me to respond to a ping? How depressing.00:54
ftaasac, o3d is now native x6401:00
ftaand is now very similar to chromium.. gyp, scons, etc.. but it builds its own flow of libs01:01
ftaincl v8 & skia01:01
LlamaZorzHello all, I am woundering if some of you could please enlighten me on when firefox forks and gets control of the flash player?01:05
micahgLlamaZorz: what do you mean?01:07
LlamaZorz When you are in firefox and go to a page that utilizes flash.  Does firefox create a new instantiation of flash or does it use one sole instance with a lock, im trying to figure out how that works01:09
=== asac_ is now known as asac
=== jonathan__ is now known as eagles0513875
=== dpm_ is now known as dpm
=== jonathan__ is now known as eagles0513875
asacfta: i think the tools/ directory should be in build-system?10:25
asactried to build fennec with 1.9.2 build-system ... choked because build/Makefile.in wants to do something with tools/rb...10:26
* eagles0513875 waves to asac :)10:29
asachi eagles051387510:29
eagles0513875asac: question bout lucid dev do you recommend upgrading or using a chroot environment to work on extensions10:30
eagles0513875also how then does one go about testing them in the version of firefox in karmic10:30
asaceagles0513875: i would suggest to setup chroot ...10:32
ubottuchroot is used to make programs believe that the directory they are running in is really the root directory. It can be used to stop programs accessing files outside of that directory, or for compiling 32bit applications in a 64bit environment (https://wiki.ubuntu.com/DebootstrapChroot)10:32
asacfor extensions you might even not need to have lucid at all10:32
asacexcept maybe for a final build tests etc. you should be able to do that in karmic10:32
eagles0513875for now testing and working on karmic should be fine10:33
ftaasac, yes, probably11:00
stsquadIs there any reason why the backported xulrunner-1.9.1-dev has an unsolvable mozilla-nspr>4.8.2 prerequesite in its libxul.pc?12:12
stsquadCould this be due to building on a later version of Ubuntu?12:24
asacstsquad: its replaced during build time ... NSPR_VERSION=$(shell $(NSPR_CONFIG) --version)12:31
asacnot sure how you backport it ... but you seem to built it somewhere where there is nspr 4.8.212:32
asacoh yeah12:32
stsquadI'm just checking now, the current xulrunner-1.9.1-dev is from the PPA (I'm on Hardy).12:32
asacand if you dont use native nspr during build you get NSPR_VERSION=$(shell $(DEPTH)/nsprpub/config/nspr-config --version)12:32
asacso the reason is that we fallback to in-source nspr if the system version is too low12:33
asacstsquad: you have mozilla-nspr.pc too?12:33
asacright ... so there should be a mozilla-nspr.pc ... dont we ship that?12:34
asacyou can surely workaround by backporting karmic nspr first ... so we use system-nspr12:35
asacor did you already backport nspr 4.8 from karmic?12:35
* stsquad looks12:36
stsquadThe only nspr I have is libnspr4-dev from Hardy's repos (which I assume must be the right one as FF3.5 is running)12:37
* stsquad checks ldd12:37
asacstsquad: no12:38
asacthe one in hardy is too old12:38
asacit will fallback to in-source nspr12:38
asacbackport the karmic nspr + nss12:38
* stsquad is confused12:38
asacthen build the xulrunner 1.9.112:39
stsquadxulruner packages an libnspr4.so12:39
asacstsquad: dpkg -L xulrunner-1.9.1 | grep nspr12:39
asacthats the in-source one12:39
asacso if system-nspr is too low our build system reverts to its own nspr12:39
asachowever, building against that is a bit busted (upstream)12:39
asacwould probably required a few fixes12:39
asacstsquad: do you need libxul.pc? maybe check if libxul-embedding.pc is what you want (for standalone applications its the right thing to do and libxul.pc is just for plugins and components that get loaded in a running xulrunner)12:40
stsquadSo the in source nspr is only for FF itself, it can't be used for embedding?12:40
stsquadWith the system FF (3.0) I was using pkg-config --exists xulrunner-gtkmozembed12:41
stsquadI'm now using  pkg-config --exists mozilla-gtkmozembed which pulls in libxul.pc12:41
asacstsquad: thats a red herring ... if you embed gecko you must not use that12:41
asacsome redhat guy pushed for that ... and hte name is confusing12:41
stsquadSo the correct pkgconfig for embedding gecko is?12:42
asacyou need to use the standalone glue ... whihc is libxul-embedding-unstable12:42
asacor libxul-embedding12:42
asacalso check this: https://wiki.ubuntu.com/XulrunnerGecko12:43
asacwhich gives info what needs to be done to port stuff from 1.8 (which stil used mozilla-gtkmozembd) to 1.9+12:43
asacmight be a bit outdated here and then but in general should be right12:43
asacsearch for #ifdef XPCOM_GLUE12:44
asacon that wiki page12:44
asacthats basically the code you need to properly use the xpcom glue12:44
stsquadAhh so the calls in the app to gtk_moz_embed_new etc need to be changed to use xpcom  glue instead12:44
stsquadasac: thanks for the pointers, I think I know what I need to fix now12:49
asacstsquad: more or less yes. there are a few implications though: e.g. you must not use internal symbols etc. just xpcom interfaces12:59
asacbut you shouldnt use those internal symbols anyway13:00
=== debfx_ is now known as debfx
fta~5 new dupes a day for the scim crash bug :(13:14
mac_vfta: i was also subscribed to that bug , finally i got fed up and unsubscribed ;)13:31
mac_vasac: why does updating firefox always reinstall the search engines? I remove them and they keep getting installed with every update :(13:32
ftamac_v, yep, i just unsubscribed too13:33
asacmac_v: thats a bug13:34
mac_vasac: oh , good...  i was worried it was intentional ...  \o/13:35
* mac_v now searches for bug :)13:35
asacits probably the same as the "keywords wiped" things13:36
asacin launchpad13:36
asachavent found time to investigate13:36
* mac_v wonders if anyone still uses yahoo search engine o.013:37
mac_vah found it.. Bug #42830613:39
ubottuLaunchpad bug 428306 in firefox-3.5 "default search engines are removed and readded (keywords wiped) with upgrade" [Low,Confirmed] https://launchpad.net/bugs/42830613:39
asacdarn ... this about: stuff kills me ;)15:12
asacyes its a mess ... supposed to fix ubufox with about:ubuntuhome about:offlinehome ...15:35
asacfta: is anything needed besides chromium-browser to build chromium?15:49
asaclike is there a build depends ppa still?15:49
=== pace_t_zulu_ is now known as pace_t_zulu
ftaasac, no, the build-dep ppa is not ready, i had issues with cdbs backports17:18
ftaasac, http://codereview.chromium.org/174162 (last comment), what do you think?17:20
=== aakashd_ is now known as aakashd
=== micahg1 is now known as micahg
asacfta: not sure. i think its ok to keep inspector if thats the user experience chrome folks want to ship18:15
* asac spins chromium on porter box18:15
ftachromium-browser                 19439   1.48%      3512    8849    7074       418:21
ftachromium-browser-inspector         964   0.07%        70     170     724       018:21
ftaasac, the initial idea was that if less than 5% of the users really need that, it's not worth making it a hard dep18:22
ftawe are at 4.96% now18:22
asacfta: right. but if upstream thinks they want such a rarely used thing by default then there is no point fighting it ... also atm folks using chromium are of the "lets try it" kind ... so they install -inspector with higher likelyhood than normal users.18:32
asacfta: how can i continue a build when i fixed something without making full clean?18:33
asacfta: in common.gypi there is 'armv7%': 0, ... is that something we have to set to 1? or are those variables detected somewhere18:37
asace.g. do i need to put that in GYP_DEFINES or somehting? i would hope that its auto detected (which as it seems it isnt)18:39
ftait's not18:42
ftawell, supposedly, there's target_arch='<!(uname -m | sed -e "s/i.86/ia32/;s/x86_64/x64/;s/arm.*/arm/")'18:44
asacfta: right. target probably is right18:45
ftatry to add armv7=118:45
asacbut in common.gypi there also is armv7% : 018:45
asacfta: is there a way to incremental build?18:45
asachmm its not a patch ... let me just respin18:45
asaclets see18:46
asacname 'armv7' is not defined while evaluating condition 'armv7==1 and _toolset=="target"' in18:47
asacthat feels a bit odd though18:47
asacit should at least be 018:47
asacwhat does that % mean?18:47
asacis that gypi syntax?18:47
asacfta: https://wiki.mozilla.org/Releases/Thunderbird_3.0rc1#Build_Revisions18:49
asaccan we produce a tarball based on such revisions?18:49
asaci assume we cannot specify two/three different revisions for comm-central/mozilla-central18:50
ftai know, they made THUNDERBIRD_3_0rc1_RELEASE earlier today18:50
asacoh there i THUNDERBIRD_3_0rc1_BUILD118:50
ftait's not needed18:50
asacin comm-central there is no such tag as it seems18:50
asacat least i dont see it in hg.mozilla.org/comm-central18:51
ftajust grab tb3 with that tag, it will fetch the right wul18:51
asacoh wait ... maybe its a branch18:51
ftait's in a branch18:51
ftathat's why i've made a tb3.1 package18:51
ftai should add it to the bot18:52
ftabut with that silly restriction..18:52
ftasame for lucid18:52
asacno such tag for comm-central :(18:52
asacoh in releases18:52
* asac checks18:52
asacgot it i think18:53
ftamozclient should do that for you18:53
asaci just didnt find it at all18:53
asacnow i have it18:53
ftajust get-orig-source as usual18:54
asacfta: i looked in comm-ncetral ... not comm-1.9.118:54
ftai use the RSS feeds18:54
asacgood ... i think chromium beuilt got past that gypi thing18:54
asacnow waiting18:55
asachmm the comm-central revisions http://hg.mozilla.org/releases/comm-1.9.1/rev/6ec88ca5bf7a seems to not have that BUILD1 tag18:56
ftai don't see a BUILD tag, just RELEASE19:01
fta[reed], can you steal the fix from the tracemonkey branch?19:08
[reed]fta: yeah, should be able to19:09
[reed]let me check when TM is going to be merged next19:09
[reed]if it's going to be a while, I'll just push it to m-c19:10
asacfta: Linux jocote 2.6.28-14-lange51 #46arm1 Thu Aug 6 23:33:27 CST 2009 armv7l GNU/Linux19:18
asacwhats the best way to make that "dynamic"?19:18
ftawhat do you mean?19:19
asace.g. in upstream build system19:19
ftawhat is uname -m ?19:19
asacfta: make armv7 : 1 in common.gypi if its armv7* i guess19:19
asacis uname -m19:20
ftayes then19:20
ftacheck uname -m for armv7 in the armel test19:22
ftawhy did you add a test using DEB_BUILD_ARCH instead of DEB_BUILD_GNU_CPU?19:22
ftaasac, ^^19:22
asacfta: isnt BUILD_ARCH right? e.g. in case someone wants to cross-compile?19:39
ftaboth should work, but GNU_CPU is less ubuntu specific, it matches hw better19:46
ftaand i wanted a generic x64 instead of amd6419:47
ftaer, x86_6419:47
ftaasac, ^^, could plz pastebin dpkg-architecture ?19:48
BUGabundosick man entering the room19:52
ftaasac, so DEB_BUILD_GNU_CPU=arm is nice, use that to set both target_arch=arm & armv7=119:57
asacfta: are you sure that GNU_CPU has the host arch?19:59
asaclike for cross-compiling and stuff?19:59
ftait's BUILD.*CPU, not HOST.*CPU20:01
ftacdbs knows about both20:01
ftaok, nice, i build xul in less than 12 minutes incl the pbuilder setup20:07
micahg1fta: how?20:07
=== micahg1 is now known as micahg
ftaby using all cpu/cores available20:08
ftacommitted in the 1.9.3 branch20:08
ftathey don't want us to take a lot of builder time, so we'll take power more for each builder20:09
fta(more misplaced)20:09
ftamicahg, it won't help if you just have 1 cpu20:13
micahgthat's good20:13
ftabut then, locally, you can use distcc, or ccache20:15
ftatoo bad the builders are just dual core, not quad or more20:15
ftaasac, http://code.google.com/p/chromium/wiki/LinuxChromiumArm20:16
BUGabundo85% [145 xulrunner-1.9.3-dbg 22299555/64.0MB 34%]                                                                         69.1kB/s 10min 3s20:19
BUGabundocan't you guys make this any smaller??20:19
BUGabundoits taking FOR EVER20:19
ftaFetched 262MB in 5min 14s (833kB/s)20:20
ftathat's still slow though20:20
BUGabundonow imagine ALL LUCID updates20:21
BUGabundonot done in a week20:21
BUGabundoNeed to get 293MB of archives. After unpacking 229MB will be used.20:21
BUGabundothe mere package list is HUGE20:22
BUGabundoFetched 18.4MB in 1min 36s (191kB/s)20:22
ftai can use lzma, but asac said it's not acceptable, for some reason20:23
ftai use lzma for chromium only20:24
ftaas it's not in the repo20:24
ftato lzma just the -dbg, it will mean patching cdbs20:24
BUGabundoI'm purging -dbg packags20:28
BUGabundo285 MBs freed :)20:29
hackelIs there any ppa that just carries Mozilla's "releases" (e.g. FF 3.6 beta 2) instead of updating every day?20:33
BUGabundomozilla security ppa20:33
BUGabundo!g mozilla security ppa20:34
ubottuError: I am only a bot, please don't think I'm intelligent :)20:34
hackelBUGabundo:  Ahh, but it doesn't have 3.6 yet, I checked.20:34
BUGabundoit doesn't?20:34
BUGabundofta care to help ?20:34
micahghackel: I built it, but I have to fix the branding still20:35
micahgBUGabundo: mozilla security is security releases :)20:35
micahgof stable versions20:35
hackelmicahg:  thanks I'll give that a try, I'm not fussy about branding. :)20:37
micahghackel: yeah, but we're not supposed to have official branding for pre-releases20:37
hackelOh, you mean it actually says Firefox instead of Namoroka or whatever?20:38
hackelI won't tell anyone. ;)20:38
BUGabundofta is there a "open in new tab from memory" shortcut on chromium?20:40
ftaasac, micahg1, BUGabundo: http://paste.ubuntu.com/317269/21:28
bdrungasac: around?21:28
* BUGabundo checks fta virus21:29
ftamicahg1, didn't you bump cairo in xul 1.9.2?21:35
=== jtv1 is now known as jtv
micahg1I've been bumping as the warnings happen21:37
micahg1nope, not yet21:37
=== micahg1 is now known as micahg
micahgdid it on 1.9.1 and ff3.521:37
micahgand TB321:38
* BUGabundo is seeing doubles21:39
micahgfta, you're starting lucid builds?21:39
ftaall ppa will explode21:39
micahgfta, I'll bump xul192 and ff36 tonight21:40
ftai need to add tb3.1 too21:40
ftaprobably need fixes too21:40
micahgfta: songbird patches, do I need to be in the outside build-tree to use them?21:40
micahgfor xul and ff, I go into build-tree/mozilla and work there21:41
ftayou need to be where the .pc folder is21:41
micahgso the .pc folder is created during debuild?21:42
ftait's created by quilt, so during the patch step, before configure21:42
micahgok, got it, I'll try to get the patches up to date over the weekend then21:43
ftadamn, i ordered two things from amazon us, they ungrouped them21:44
ftalol "estimated delivery date: December 22, 2009"21:46
ftacan anyone play this smoothly? http://vimeo.com/668676821:49
ftai can't, neither with ff, nor with chromium21:50
bdrungasac: please release m-d as version 0.17.1 (i want this rc bug fixed)21:51
micahgfta, seems mostly smooth with the occasional audio hiccup21:54
ftawith 3.7, the sound is jerky, like every 3 secs, so the video sometimes stops21:55
ftathe browser is at 100% cpu, so the plugin probably lacks resources to play smoothly21:55
ftait's a dual core 2.3GHz21:56
ftadtchen_, ^^21:56
ftamicahg, asac: http://paste.ubuntu.com/317291/21:58
* micahg will be waiting for it to break tomorrow :)21:58
av`bdrung, which RC bug?21:58
av`bdrung, you can upload it yourself21:58
bdrungav`: i can upload it, but not sponsor it21:59
hackelfta:  plays perfectly with mplayer.  Flash has never played video well for me.  Great video, by the way...21:59
av`bdrung, ?21:59
bdrungav`: https://code.launchpad.net/~mozillateam/mozilla-devscripts/mozilla-devscripts22:00
bdrungcommit 27422:00
av`bdrung, re-link it again please, lost the window22:01
ftahackel, yep, i love timelapses, i'm trying to make my own videos, but linux mostly sucks for photography :(22:01
bdrungav`: ?22:01
av`bdrung, re-link me the branch22:02
bdrungav`: can't you scoll up to this post? https://code.launchpad.net/~mozillateam/mozilla-devscripts/mozilla-devscripts22:02
ftaav`, it's a native package, don't release stuff without going to asac or myself22:03
av`bdrung, yes, works now22:03
ftawe need to make an "upstream" release22:03
av`fta, actually bdrung is one of the authors22:03
av`fta, plus m-ds has an R-C bug that makes it unusable atm22:04
av`so I don't know if we can wait long to have it uploaded22:04
ftastill not a reason to do a wild upload22:04
av`I don't think bdrung would do a wild upload22:05
ftaiirc, asac said he'll do it22:05
micahgav`: there are versions in the u-m-d ppa22:05
bdrungfta: are you a DD?22:06
av`ok, great, I saw bdrung was asking several times about it so I didnt understand why he couldnt upload it himself22:06
ftano, but i'm upstream for this package, i've created it22:06
bdrungfta: it's a bug fix and a point release (0.17 -> 0.17.1) would fit, wouldn't it?22:06
av`micahg, it's not a local fix, buildds doesnt work with the current release22:07
av`micahg, so it needs to be pushed to the archive and not locally, since the buildd gets the latest package built22:07
micahgah, you're saying nothing will compile in the builders22:07
ftainitially, i just wanted X.Y (a float), but well, i don't mind22:07
micahggot it22:07
av`micahg, exactly, it fails to build cause of that RC bug :)22:08
bdrungjazzva is not on the uploaders list. so sponsoring would not work. then i can upload it, too.22:08
av`fta, jazzva used 0.18 not 0.17.1 from what I saw22:08
av`bdrung, no, you can't22:08
bdrungav`: ?22:09
av`bdrung, the changelog's entry should be owned by you22:09
av`bdrung, the changelog's entry should be owned by you22:09
bdrungav`: yes, i have to change the ownership22:09
av`= your name must appear into the changelog latest entry22:09
av`yes, exactly22:09
bdrungav`: the changelog owner needs to be in the uploaders list22:10
av`but you must be the changelog owner to have the upload accepted22:10
bdrungit is required to change it anyway22:10
av`so uploaders list + ownership of changelog's entry22:11
bdrungav`: yes, i will change this if i upload the package22:11
av`k, perfect22:11
av`but anyway ask to fta or asac, I don't wanna set up a war for a bug fix :)22:11
bdrungfta: am i allowed to upload it to debian?22:12
bdrungfta: do you prefer 0.17.1 or 0.18?22:13
bdrungasac: ^22:13
ftaupload yes, but only once it has been closed. and as i'm not in sync myself, i want asac to close it22:14
ftathat's what i meant by upstream22:14
ftasorry, i'm tired, tough to write tonight22:14
bdrungfta: what do you mean with closed?22:15
ftathe UNRELEASE to xx commit22:15
bdrungfta: with uploading i meant changing UNRELEASE to xx, commit and tag it, and then upload it22:16
ftathen we need to agree if there's enough stuff in to make the release, or if we have other stuff in the queue to stick in22:18
ftaas i said, i lack context, so i won't do it myself, sorry22:18
bdrungfta: i don't know any stuff.22:19
bdrungfta: so it's better to release this one commit bug fix.22:19
bdrungfta: release often, release early - especially bug fixes22:20
=== cyphermo1 is now known as cyphermox
=== BUGabundo1 is now known as BUGabundo
dtchen_fta: if a PA client requests low latency, PA is happy to grant that at the expense of high cpu utilization23:44
dtchen_fta: this is a consequence of the mainloop23:44
ftadtchen_, is there a way to mitigate that? i mean, slightly higher latency in exchange for less cpu?23:46
ftaso playback becomes smooth again23:46
dtchen_well, yes, the PA app needs to be changed23:48
dtchen_I'll try to push a PPA version soon23:48
dtchen_time. I never have enough.23:48
ftadtchen_, would it help to have daily p-a PPA?23:54
dtchen_fta: no23:54
dtchen_I discussed this with PA upstream at the beginning of the Karmic cycle, and he said it wouldn't be particularly useful, because changes aren't committed to the public VCS daily23:55
dtchen_note that I wouldn't mind having one, but he seems indifferent to it, which is somewhat baffling23:56

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