asacMook_sb: yeah. thats kind of tough ... but interesting00:02
asacand yeah. seems like they want to shake off all reuses of their platform00:03
asaci am not sure they want. but this way will probably mean there is no chance for embedders or even other xulapps to keep up to speed00:04
asacbut its also consequent ... and i personally accepted that this will happen00:05
micahgthey're having trouble with their own apps keeping up SM/TB00:05
micahgI guess that's why everyone's migrating away from GRE00:06
asaci guess TB will survive a bit longer00:07
asacthan SM00:07
micahgwell, the SM team is hyped up about the new release00:08
micahgthey plan on launching soon00:08
asachyped up?00:08
asac ok00:08
micahgvery excited00:08
asacsure. the problem is that last release was from same branch as ffox 2.000:09
asacso it took till now to get the new release done00:09
asacgranted, there was quite some work done too (not just porting)00:09
asacbut now they hear that they need to get things ready for 1.9.2 ... :)00:09
micahgyeah, well, we're kinda in the same boat as they are in terms of migration (granted our migration is a little easier)00:10
asaci have  need help porting hardy to 3.600:12
micahgare we going to try to do that?00:13
asacand yes. we sit in the same boat ... but more on the outside ;)00:13
mconnorTb3 is pretty awesome, fwiw00:13
asacyes it is00:13
asaconly thing is mem consumption00:13
asacbut the new search rocks ;)00:13
asacthe UI is a bit inconsistent ... i would think in the long run the main mailvinwdow will move more in a similar direction00:14
asacmicahg: we kind of have to00:14
asacreinvent ourselve. tracking full upstream releases00:14
micahgbecause it's LTS?00:15
asacsomehow ... dont ask me how now00:15
asacmicahg: yes. 1.5 years to go when 3.0 goes eol00:15
asacbut same for jaunty00:16
asacbut i think hardy is a bit tougher00:16
micahgyes, but jaunty has 3.5 already00:16
asacmicahg: jaunty doesnt have 3.500:16
micahgyes it does00:16
asacit has a preview ...00:16
micahgin universe00:16
asacyes. its not a so big issue to move the app00:16
micahgit's the same app with the branding turned off00:16
asacbut we also have other embedders that use the same rendering engine00:17
asacwe probably have to port all apps that have remote content exposure00:17
asacthats probably the minimum set we can argue. and all apps that are in main of course00:17
micahgthat means backporting what we jsut did for karmic to jaunty00:17
micahgand hardy00:17
asacmicahg: only that in hardy there are more apps ;) ... asa webkit is almost non-existing ;)00:17
asacand in jaunty too00:18
asaci think hardy has the most00:18
asacwith intrepid having no real difference00:18
asacand jaunty slightly less because webkit started to grow00:18
asacbut epiphany in main is a mess00:18
asacthe 1.9.1 port will be not so nice ;)00:18
asacand epiphany-webkit is not feature complete enough to just drop that in as a forced replacement00:19
micahgwhat is debian doing with this stuff?00:19
asacthey probably will surrender. usually i gave them all the backports i made00:19
asacand even did the uploads00:19
asacnow they ride the upstream minor updates00:19
asacbut major updates will be a tough thing for them00:20
asacif not impossible00:20
micahgasac: http://browsing.justdiscourse.com/2009/10/08/debug-version-of-prism-1-0b2-for-ubuntu/00:20
asacso debian will be back where they were when i startred to do security support for mozilla like a bunch of years ago ... no security support00:20
asacmaybe we can help them with our patches00:20
micahgwhat will they do with epiphany htough00:21
asacbut its even tougher for them00:21
asacthey also need to get all archs properly working00:21
asacmicahg: epiphany is gecko ... so they dont have suport. in future its webkit so thats ok00:21
micahgno, I meant epiphany in debian00:21
asaci think they will not release mozilla at all00:21
asacor go the same road as we do00:22
asacno r-depends00:22
asacand major updates00:22
asacmicahg: yes. in debian its also gecko ... and if they cannot support xulrunner its automatically not supported00:22
asacbut debian does only support "1 year" after relase of stable00:22
asacso they will be off the hook sooner than we will ;)00:23
asacone thing is clear. in lucid we need to have everything in place so we can ride the usptream releases for 3 years ;)00:23
micahgwell, easiest is to go back to just "firefox" and push every update00:24
asacyes. but we need to get all the xulrunner consumers with remote access out of the archive too00:24
micahgbut then you can't package extensions00:24
asacjust firefox isnt good enough00:24
asaci am not cared about firefox00:24
asacthat will sort out on its own00:24
asacjust the rest00:24
asacbut i guess we just need to be harder about killing apps that use gecko00:24
micahgI was wondering, why are you trying to get more extensions in the archive anyway?00:25
asace.g. removing from archive00:25
micahgif we can't reliably update them during the release cycle00:25
asacwe could  ;)00:25
asaca matter of resources00:25
asacbut basically the idea is: if they work at release time00:25
asacthey shouldnt stop working while release is maintained00:26
asacso they deliver a "stable" experience00:26
asacstable in the sense that you know the bugs and you dont get new bugs tomorrow00:26
micahgyes, but stuff like noscript where they find new vulnerabilities all the time00:27
micahgdon't get updated00:27
asacthats a mistake00:27
asacand needs fixing00:27
micahgversion in archive latest:
micahgcan you automate extension building from amo?00:28
micahglike we do the dailies00:28
micahgbut maybe weekly instead of daily00:28
asacin general i agree that we should balance work spend on maintenance vs. new extensions00:28
asacand stop doing new if the maintenance is too much00:28
asacmicahg: we started on having auto syncs ... yes.00:29
asacmicahg: we have a set of scripts that basically already does that00:29
asacit just needs to be plumbered together00:29
dtchenfta: yeah, the SDL bit is known. I'm going to look at it this evening (in an hourish).00:29
asacthat was one of the plans that ended up short before the end ;)00:29
asacmicahg: https://code.edge.launchpad.net/~mozillateam/firefox-extensions/med-auto-scripts00:30
asacmicahg: what that does is it syncs the registered extension ids and detects if there are newer xpis than what is currently in branch00:30
asacor something00:30
asaci think whats missing is the branch updating00:30
asacbut maybe thats even there too00:30
* micahg wants to know how to recruit more triagers for bugs...00:31
asacthats a communication thing.00:31
asacmore blocking00:31
asacreporting status00:31
micahgoh, also, should update my 3.6b1 test build to not use system sqlite?00:32
asacmicahg: if it still works it measn that your version is high enough00:32
asacyou should just have the right minversion in debian/rules00:32
asacat least atm00:32
micahgno, because of the compile flags that reed pointed out00:33
asacif we go back to all in-source, w can probably do that00:33
asacah. maybe. i will upload the update tomorrow00:33
asacto security ppa and then to -proposed for some baking for hardy-jaunty00:33
asacmicahg: i fixed tb3 head ... i think ... did you see what was wrong with 1.9.3 patches?00:58
micahgasac: I thought it was cairo that was the issue with 1.9.300:59
asacthen all is fixed01:00
asacthe patches applied and its building here01:00
micahgasac: https://edge.launchpad.net/builders/bohrium01:11
micahgI'm trying to figure out who to contact01:31
micahgit's that build machine01:31
micahgthat's why we didn't have a build yesterday though01:32
asacmicahg: in #launchpad01:38
asacask there01:38
asacmaybe there is someone on duty01:38
asacotherwise i can ask around tomorrow01:38
asacnow dropping out01:38
micahgalready did01:38
asacoh. no answer?01:39
micahgnot yet01:39
asacok if noone ansewrs just drop a line ... enjoy!01:42
jcastromicahg: when you post that a bug is reported upstream in mozilla please do an "also affects project" and paste the bug in the url field03:35
micahgjcastro: what are you referring to?03:36
micahgIf I have the bug I add it03:36
ubottuLaunchpad bug 448686 in firefox-3.5 "CSS table borders do not display" [Medium,Triaged]03:36
micahgI usally do03:37
jcastroI just linked it, I have a report that goes through bugs with comments with links in comments that aren't link, to catch them. :D03:37
jcastroit's ok it's my job to catch the ones that get away03:37
micahgwas 1AM03:37
jcastrobut I find that some people never know about the feature so I always follow up.03:37
* micahg knows03:37
micahgI also add them for people if I catch it03:37
jcastroI am currently working with the lp guys to make it not so annoying to do though03:37
jcastro<3 awesome03:37
micahgwell, you don't want to entirely automate it03:38
jcastrooh no03:38
jcastroit won't be automated for sure03:38
jcastrobut something inlineable so you don't have to do a page load I think would be much easier03:38
micahgperhaps a flag next to a bugwatch if the upstream project is labeled03:42
jcastroand I think "also affects project" is confusing03:43
jcastrothere should be a specific "link to an upstream bug project" or something03:43
micahgjcastro: well, there's 2 functions for it03:44
micahg1 is to mark upstream03:44
jcastroyeah, I think that's the problem03:44
micahg2 is to mark another project affected03:44
jcastrobut we'll see what UI people say03:44
micahgLP UI team is reluctant to add more buttons03:44
* jcastro nods03:44
jcastroI'm going to have a session about this at UDS, so if you have any feedback or concerns feel free to just mail me03:45
jcastroI am going to video some people upstreaming bugs to get more data, etc.03:45
* micahg wishes he could attend UDS03:46
markeyhey all05:54
markeylately there have been no Chromium nightly updates05:55
markeyanyone happen to know why?05:55
=== micahg1 is now known as micahg
micahgI thought I got an update05:57
micahgmaybe not :)05:58
micahgmarkey: daily is broke :)05:58
micahgwanna fix it?05:58
markeynot really, I'm not much of a packager :)05:59
markeywhat's borked with it?05:59
nillebhello everybody08:09
nillebi'm here again about the problem I discuted yesterday w/ asac08:09
nillebI think I got it08:09
nillebthe apparmor profile for FF is wrong in karmic.08:10
nillebthis statement is true for users on an active directory domain, that has been joined, for example, with likewise-open08:10
nillebfor these users, the home directory looks like /home/DOMAIN/user08:11
nilleband this, IMHO, prevents FF from setting a lock on the .parentlock file.08:11
nillebwell, now the evidences of this misbehavior:08:12
nillebOct 22 11:13:52 ly-qa-bellinux kernel: [ 8549.052195] type=1503 audit(1256202832.201:81): operation="file_lock" pid=6085 parent=5539 profile="/usr/lib/firefox-3.5.*/firefox" requested_mask="wk::" denied_mask="k::" fsuid=1893211406 ouid=1893211406 name="/home/ESKERCORP/bellin-salarin/.cache/event-sound-cache.tdb.f1b0a25408a48159af64e77c486e058d.i486-pc-linux-gnu"08:12
nillebI think I'm going to open a log, I will file it against mozilla-firefox08:12
nilleband apparmor08:12
nilleb(log=bug report)08:12
asacjdstrand: ^^^09:48
asacjdstrand: ^^09:49
asacapparmore ;)09:49
jdstrandasac: already responded to his bug09:49
jdstrandbug #45884609:49
ubottuLaunchpad bug 458846 in apparmor "apparmor prevents firefox from starting for an user who belongs to an Active Directory domain (dup-of: 447292)" [Undecided,Won't fix] https://launchpad.net/bugs/45884609:49
ubottuLaunchpad bug 447292 in apparmor "AppArmor does not allow access when @{HOME} is not /home" [Low,Won't fix] https://launchpad.net/bugs/44729209:49
asacah ;)09:50
jdstrandhe simply needs to read https://wiki.ubuntu.com/DebuggingApparmor#Adjusting%20Tunables09:50
asacjdstrand: isnt it way too early?09:50
asaci mean ... i just got up :-P09:50
jdstrandasac: yeah-- dog woke me up and then I couldn't get back to sleep09:50
asac(a tad late i agree)09:50
asacok ... feels like i should get a dog soonish ;)09:50
jdstrandjust don't get a male09:51
asacyeah. my girl wants a girl ;)09:51
jdstrandI'm sure some males are fine. ours is stubborn and a digger (wakes us up scratching his bed)09:51
micahgasac: I don't understand the difference between LIBXUL_SDK and SKIP_COPY_XULRUNNER09:56
asacmicahg: SKIP_COPY_XULRUNNER?09:57
asacisnt that a patch by us?09:57
asacmicahg: so bascially if you build a libxul-sdk based aplication and want to ship it09:58
asacyou need to ship it side by side with xulrunner ...09:58
asacunless you are a distributor that knows that there is xulrunner on system09:58
asacthats when you SKIP_COPY_XULRUNNER09:58
asaci guess LIBXUL_SDK Is clear?09:59
micahgi would think09:59
asacso if you build mozilla apps with --with-libxul-sdk09:59
asacit means it will just use the system xulrunner09:59
asacand not build the full xulrunner on its own09:59
asacthats why SKIP_COPY_XULRUNNER comes into play10:00
micahgwhy do you need SKIP_COPY_XULRUNNER if you're building with LIBXUL?10:00
asace.g. if you dont build xulrunner you need to copy it if you want to ship a bundle with everything10:00
micahgwhy 2 variables?10:01
asac10:58 < asac> micahg: so bascially if you build a libxul-sdk based aplication and want to ship it10:01
micahgthat's what I don't get10:01
asac10:58 < asac> you need to ship it side by side with xulrunner ...10:01
asac10:58 < asac> unless you are a distributor that knows that there is xulrunner on system10:01
asac10:58 < asac> thats when you SKIP_COPY_XULRUNNER10:01
asacthats what i tried to explain i nthose lines10:01
micahgi saw10:01
micahgbut I don't understand10:01
micahgI thought building with libxul-sdk implies system xul10:01
asacmicahg: xulrunner is not only made for distros10:01
asacso thin kabout windows10:02
asacif you build a xulapp .... with libxul-sdk10:02
micahgI think I get it10:02
asacyou want to ship something that works10:02
micahgone is for building and one is for running10:02
asacso you need the full xulrunner copy10:02
asacthats what it does by default10:02
asacand thats why one needs to explicitly SKIP that on linux distro packages10:02
asacmicahg: yes. if you do --wit-libxul-sdk it will build against that ... and in the end copy the full xulrunner10:03
asacso you have it in the same tree and can ship a complete solution that works independently of the system10:03
asacyes. what you said is about right10:03
micahgso I messed up the patches...I'm trying again now10:07
micahgok, so there's a ifdef SYSTEM_LIBXUL in there now10:17
micahgwhere we had LIBXUL_SDK10:18
micahgshould I update our patches with the upstream CONSTANT?10:18
asacmicahg: if that constant is now used ... why not10:20
asacmicahg: so SKIP_COPY_XULRUNNER imo was also defined elsewhere10:21
asacbut its there for ages, so i am not sure10:21
asacto give you a 100% sure answer i would have to read the code. but if you think the upstream constant is a better way, then propose it and i will check10:21
micahgsomething's not right10:22
micahgLIBXUL_SDK is still used10:23
micahgI think this is too much for me at this hour10:25
micahgI'll try again when I get up10:26
asacmicahg: yes. try to take a step back ... and then lets seee10:31
asacmicahg: what are you doing actually?10:31
micahgfixing the patches as they no longer cleanly apply10:32
asacoh ok ... ffox ... right10:33
asacLIBXUL_SDK will be used for sure10:33
asaci dont think this will ever get removed10:33
asacthought the SKIP_COPY_XULRUNNER was removed by something else10:33
asacmicahg: oh yeah.10:34
asacmicahg: lets check that when you wake up10:34
asaci am then in weekend mode and can probably give better input ;)10:34
asacbdmurray: hey ;)14:52
asacbdmurray: can you give cyphermox the intro wiki page for bugcontrol and add him to that team once he confirms he understands? he is NM team member and needs to be able to work on bugs ;)14:53
ftaasac, how do i add a tab in a bzr branch?16:23
asacfta: for package releases?16:25
asacfta: basically same approach as we do ... release commit ... just use decommit -r -e16:25
asacthat will do the right thing and allows you to be more concrete16:25
asaci think we should adjust debcommit code to be more of the format that we use in mozillateam16:25
asacit just says: "relesaing x.1.2-0ubuntu2"16:25
asacwhile it should say "releaseing x.y.a.s. to karmic" ... at least16:26
asacfta: otherwise its just bzr tag 1.2; bzr push16:26
asacif you do an upstream release16:26
ftacould it be done afterwards? i mean, adding tags to older revisions?16:26
asacfta: sure16:28
asaci think you can do bzr tag -r xxx TAGNAME16:28
asacyou can do that accoreding to bzr help tag16:29
asacyou can see all tags with bzr tags16:29
asacor delete tag with bzr tag --delete TAGNAME16:29
ftawould be nice to be able to commit with a user specified date too. so i can re-create a branch for a 15y+ project i'm still working on16:30
asacfta: is that in svn or cvs?16:31
asacthey are easy to import16:31
asacand dates etc. shouuld be fine then16:31
asaci think16:31
asacotherwise set date before commit ;)16:32
ftano, nothing, no vcs at all, i just have my old tarballs and changelogs16:32
asacyeah. sounds like a plugin16:32
asace.g. tarball-series import16:32
asacnot existing. just saying that that should probably be done as s plugin16:33
asacfta: bzrtools ... bzr help import16:37
asacPurpose: Import sources from a directory, tarball or zip file16:38
asacUsage:   bzr import SOURCE [TREE]16:38
asac  This command will import a directory, tarball or zip file into a bzr16:38
asac  tree, replacing any versioned files already present.  If a directory is16:38
asacmight be worth a try16:38
asacthough not with date tweakage16:38
asacnot sure how smart it is16:38
asacbut maybe give the tarball an old timestamp and see if it uess that ;)16:38
asacimport does not commit on its own though16:40
asacfta: i think should be easy to do16:49
asaclet me see16:50
* asac pulls bzr bzr16:50
asacgrr ... branch format outdated16:50
ftabug 44302817:43
ubottuLaunchpad bug 443028 in evolution "evolution hangs when trying to open/save attachments in an email with 2 ppt attachments" [Medium,Incomplete] https://launchpad.net/bugs/44302817:43
ftacool, i'm not alone17:43
=== dpm is now known as dpm-afk
asacfta: great. thats a start18:02
asacfta: i will write the timestamp patch for bzr ;)18:02
asacbut it took 1h or so to convert my bzr branch to a format so i can pull the latest ;)18:02
ftanot sure it's the same freeze, i don't have logs for mine18:02
asacso if you wait till tomorrow you can use bzr import + bzr commit18:02
ftathey happen only on my x64 box18:03
asacfta: can you get a threaddump if it freezes?18:03
ftai should try, but i'm at home (32bit)18:03
ftaseems virtualbox can't do d3d10, too bad18:07
ftano aero with W718:07
ftafunny, d3d is in fact using the wine driver18:19
ftaRejected: File gwibber_2.0.0~bzr476.orig.tar.gz already exists in Primary Archive for Ubuntu, but uploaded version has different contents.18:24
ftai wonder if the bot should be smarter and 1st look at the archives before calling get-orig-sources18:24
ftaor if it's a marginal case for dailies18:25
fta(not worth the efforts)18:25
asacfta: http://pastebin.com/f72bb14ce18:35
asacfta: i am not sure. i had the feeling that we get a bit often rejections because of bad version in some way18:35
ftayes, but here it's different18:36
asacisnt it the "if on top there is released, it should bump version rather than lowering"?18:36
asacfta: the patch above works great for me ... e.g.18:36
asac/home/asac/local_bzr/bin/bzr commit --commit-time='2009-10-16 08:59:00 +0100'18:36
ftaa --commit-from-reference=some_file would be nice too18:37
ftaer, --commit-time-from-reference18:37
ftasimilar to 'touch -r'18:37
asacyeah well ;)18:38
asacyou can wrote a wrapper for now18:38
asaclike ls -l /path/to/tarball -> make date18:38
asacuse that to commit after bzr import /path/to/tarball18:38
asacfta: i think that should rather in a "import-and-commit" ;)18:38
asaci will think about it ... now that i am the bzr master ;)18:39
asacfta: can you open a wishlist but for the --commit-date thing against bzr upstream so i can request a merge?18:39
ftanot sure i can change upstream bug status18:46
ftaor is wishlist a tag?18:47
ftanope, it's "importance", i can't change that18:48
ftaasac, ^^ bug 45927618:50
ubottuLaunchpad bug 459276 in bzr "user specified commit date" [Undecided,New] https://launchpad.net/bugs/45927618:50
asacheh yeah18:54
ftatb3/chromium/o3d red19:11
ftasongbird too19:11
ftagwibber out19:11
ftanot a good day for dailies :P19:12
gnomefreakanyone else notice that the folders with mail in them in tbird3 are not bold anymore?19:28
gnomefreakstarted a week or 2 ago i think19:28
gnomefreakbug 33914919:30
ubottuLaunchpad bug 339149 in thunderbird "Thunderbird add on lightning 0.9. Cannot add task " [Undecided,Invalid] https://launchpad.net/bugs/33914919:30
asachavent opened tbird for a fe wdays19:31
asacnot sure if its more than a week19:31
gnomefreaki dont recall date or timeframe but it has been a little while19:32
asacnot sure. maybe you didnt subscribe those folders for auto updates or someting?19:32
gnomefreakwhat do you mean? they get new mails in them19:33
gnomefreakif i leave mail in the folders without reading them they turn bold/black when i reopen tbird3 but i have to have looked in folder before closing19:34
asacmy folders are bold19:36
asacyou can subscribe explicitly to certain folders by right clicking etc.19:36
thunderstruckasac: default theme?19:36
=== thunderstruck is now known as gnomefreak
gnomefreakwhat do i choose to subscribe after right clicking on folder?19:37
asacfta: tb still red19:37
asacdid i forget to push my fix ;)?19:37
ftaasac, probably19:47
ftagood, next chromium should be green, upstream fixed the bug for me ;) good upstream.. compared to others19:47
fta(no name)19:48
gnomefreakseems that FF3.5 3.6 and 3.7 can use search again19:48
mconnorfta: harsh :P19:48
gnomefreaki hope green == build succeded19:49
ftamconnor, let's face the truth, some upstream are more cooperative than others and are more willing to keep downstream build green19:50
mconnorfta: if your builds reported to our systems, I bet it'd be easier19:51
ftagnomefreak, lol, yeah19:51
ftamconnor, I don't want to fight on this but experience proved that each time we had a build failure, we had to write the patch ourselves, submit it, and wait weeks if not months (not to say forever) to get it in, and we had to refresh it many times in the meantime.19:55
mconnorfta: bug #s?19:56
fta(but i wasn't targeting moz in my initial peak)19:56
mconnorI don't want to fight on it, I want to figure out how real any of this is19:56
ftamconnor, no bug in particular but i've built moz stuff hundreds of times since 200719:57
mconnorit's hard to change stuff when I can't see how we're failing :-/19:58
ftai also agree we're not very good at upstreaming our downstream patches19:58
=== asac_ is now known as asac
mconnorprocess issues are just like bugs, saying they're there without STR means I can't fix it!19:59
micahgasac: I'm going to try prism again in about 20 minutes20:03
Mook_sbI guess songbird at least knows where we're failing? :( (--enable-system-sqlite, in the most recent case)20:21
[reed]fta: you're horrible at upstreaming your patches20:33
[reed]end of story ;)20:34
[reed]iirc, my stats show: Debian is first. Red Hat is next. openSUSE is third.20:34
fta[reed], you're so horrible at accepting patches or even critics that i no longer bother even filing bugs20:36
ftaend of story20:36
micahgcan't we all just get along ;)20:37
micahgasac: can I drop the use 1.9 only patch in prism?20:42
[reed]mconnor: 513067, 471359, 466250, 478871, 463872, 423334, 417345 all have various issues where upstream (Mozilla) didn't take the time to help, took too long, didn't care, or some other issue (this is from a quick read, so I may have missed something).20:45
mconnorfta: feel free to cc me on anything where there's inaction, I can still hit people with a hammer ;)20:48
micahgasac: are you therE?21:05
micahgfta, can you help me with an error?21:10
micahgfta: asac did commit the patch for TB3 yesterday21:19
micahgaccording to the logs21:19
ftamicahg, checking21:22
ftatoday's error is "checking for libnotify >= 0.4... Package libnotify was not found", fix is easy21:23
micahgok, I'm still wresling with prism21:23
micahgfta: can you help me with this: http://pastebin.com/f127c38a221:27
ftafixed tb321:29
micahgoh, you already fixed tb321:29
micahgI was oging to trade you :)21:29
ftamicahg, what did you change compared to my branch? (prism)21:29
micahgadded libasound build dep, change to xul-1.9.1, drop 1.9 only patch21:30
micahgdaily build is failed as well, but I think that was a patch21:30
ftaprism daily failed?21:32
asacmicahg: yeah. we need kind of a system-libnotify patch21:32
asacerr... magic for debian/rules i mean21:32
ftafor prism?21:33
asacfor tb321:33
micahgI thought prism daily's been failed21:33
ftaer? we just need a build-dep21:33
micahgasac: fta fixed tb321:33
ftafor prism, i don't know, i need to see your changes21:33
micahgok, I'll push a branch up21:34
ftapastebin a bzr diff21:34
ftamicahg1, hm, at 1st glance, it should work21:46
micahgfta: any ideas, I've got about another hour before I have to quit21:48
BUGabundoboas noites o/22:10
ftamicahg, for some reason, the build system is broken, configure correctly creates prism/installer/Makefile from the .in, but not prism/Makefile22:21
ftaBUGabundo, ola22:21
micahgok, so what do I do22:22
BUGabundoolá fta22:22
BUGabundodid you manage to use shared folders on VB?22:23
ftamicahg, a trick could be to patch prism/makefiles.sh but there's probably a better way22:23
micahgI don't know too much about the make files22:23
micahgI guess I can try to patch it up sat night22:23
ftasomething changed in the build system.22:24
ftaasac, ^^, any idea?22:24
BUGabundodtchen: FYI my sound seems really nice now! even keeps volume settings post reboot :)22:26
ftaBUGabundo, for me, it regressed in the last few days. it used to be fine post reboot, now it's ever mute or very low22:27
ftaBUGabundo, do you have h/w acceleration in vbox? (dx9 or better)22:34
BUGabundono idea22:34
BUGabundohaven't tested it22:34
ftayou have no windows guest?22:35
BUGabundoat work22:35
asachave to check prism22:55
asactb3 failed again?22:56
ftayep, but fixed, again22:56
ftafor prism, allmakefiles.sh changed a lot22:57
ftait's from xul22:57
ftaseems best to just patch prism/makefiles.sh22:57
ftabut keep use_xul_1.9_only.patch (just updated to 1.9.1)22:58
asaclets first see whats up ;)22:59
asacall makefiles.sh in prism/ is done by us?23:02
asacseems to be just buggy23:02
asacgot it23:02
asachmm micahg is not there23:02
asacdont really want to take all work away from him ;)23:02
asacfta: isnt the prism/makefile.sh from the prism svn?23:02
ftait is23:05
asacah but the makefile isnt?23:05
asacshould be in there23:05
asacthen makefile.sh is indeed just buggy ;)23:06
asacodd ... not sure how allmakefiles.sh finds all the other Makefiles23:06
asacjust the top level one is needed23:06
asacok seems to go down through SUBDIRS or something23:07
asacand prism/installer isnt there23:07
asacyeah ;)23:07
asacso after that mozilla-devscripts xpi.mk needs another change and then all should be fixed ;)23:08
dtchenfta: interesting -- does clearing out ~/.pulse* help?23:11
BUGabundodtchen: do you have the time to help joaopinto with audio probs with games?23:14
ftaasac, lol, 3rd failure: checking for cairo >= 1.8.8 freetype2 fontconfig... Requested 'cairo >= 1.8.8' but version of cairo is 1.8.623:14
asacoh no ;)23:14
asachmm... i thought i bumped cairo23:14
asacoh wait23:14
asaci didnt bump it there because i didnt see the failure23:14
asacin tb23:15
dtchenBUGabundo: if they're SDL-based games, I have a review of the latest upstream changes to the ALSA backend queued for this weekend.23:15
asacand the other failure was on 1.9.323:15
asacmaybe means we will see the same cairo bump23:15
asacin 1.9.2 tomorrow23:15
asacmost likely23:15
ftadtchen, could you put the sdl fixes in the ppa?23:15
dtchenBUGabundo: but I really, really need to focus on some PA bugs first.23:15
dtchenfta: yes, that is the plan23:15
ftadtchen, also, if you have a fix for choppy sound in virtualbox23:15
BUGabundothanks dtchen23:16
BUGabundoI'll let him know23:16
dtchenfta: do you have a reliable test case for vbox?23:16
dtchenI can't reproduce it locally using vbox 3.0.8, but...23:16
ftavbox karmic, W7 as guest23:17
dtchenok, I'm not sure how I would reproduce that unless there's a free version of W723:18
dtchenanyhow, head-down PA hacking for some hours now.23:18
BUGabundodtchen: there's a RC _free_23:21
BUGabundofor 6 months or so23:21
micahgfta, so I should modify the use_xul_1.9_only to 1.9.1?23:23
ftabecause otherwise, it will take a random xul, it should use the one it has been built with23:24
micahgI'll update that tomorrow night23:25
micahgwhat do I have to fix with prism/makefiles.sh23:25
asacmicahg: back23:25
asacadd prism/Makefile23:25
asacthats a plain bug in prism build system23:25
asacand i remember now that i already complained to plasticmillion about it when i last looked23:25
asache promissed to fix it ;)23:25
micahgok, so I make a patch to do that, right?23:26
asacyes. then next there is a problem of how mozilla-devscripts xpi.mk is used23:26
asacyou have to fix it by not using a refractor.xpi:: rule approach23:26
asacbut rather copying the .xpi to top level dir after the cdbs build23:26
asacso i would say in binary-post-build:: or something like that23:26
fta4th try for tb3 :P23:27
asacmicahg: so you see the refractor.xpi:: thing?23:28
asacremove that and use common-post-build-arch::23:29
asacjust put the copy of the .xpi there23:29
asacthat should work23:29
micahgin rules?23:29
micahgasac: there are 3 lines there?23:29
micahgi remove them all?23:30
asacmicahg: its just the wrong rule ... the action is right23:30
asacso the :: line replace that as i said23:30
ftatb3 probably shifted to a fresher xul23:31
asacrefractor.xpi:: build/prism zip -d dist/xpi-stage/refractor.xpi prism/\* cp -f dist/xpi-stage/refractor.xpi .23:31
asacnot sure why there is a zip -d23:31
asaci thought that that was already done by the prism build system23:31
asacmaybe drop that line too23:31
asacand only cp -f23:31
asacfta: sure that 1.9.2 didnt just land that cairo bump in between?23:32
asacwould be naturally ... first land on trunk23:32
asacthen to 1.9.223:32
fta5 new reqs since yesterday23:32
asacthen on 1.9.1 :(23:32
asaci thought just libindicate > 0.423:33
asacand this cairo thing as well as the other23:33
asacso 3 ;)23:33
micahgasac, I'm adding a patch to prism called add_makefil23:33
ftaasac, libindicate, libiw-dev, cairo, mesa-common-dev23:33
asacmicahg: fix_prism_makefile_sh ;)23:33
asacuse what you want23:33
asacjust document it and send it upstream23:34
asacso he doesnt forget to appl that23:34
ftathe later is for WebGL, which is only in 1.9.3 (iirc)23:34
asacfta: libiw-dev?23:34
asacdo they ship a libiw now in-source?23:34
asacgod help me :)23:34
asacbut libiw-dev was a requirement for quiet some time23:35
asacfor geolocation23:35
asacoh it was optional23:35
asacbut we had it i thought23:35
asacif not it was my fault to sync  it to the other branches23:35
asacmesa-common-dev sounds a bit scary too23:35
asacbut could be something freaky ;)23:35
asacbut common-dev sounds a bit coards23:35
fta$ grep -c libiw-dev */configure.in23:35
asacnot sure why libiw-dev would be in there23:37
asacthats a package name23:37
asacnot a lib23:37
asacor header or .pc file name23:37
asacthats the check23:38
asacthe libiw-dev is just a error message23:38
micahgasac: does libasound2 need to be an install-dep as well as a build dep?23:38
asacmicahg: why?23:38
asacnot sure why that would be23:38
ftaasac, that's what we have in xul23:38
asacwe have that23:38
asacbut why are you saying they bumped the lower verison23:38
ftano, just cairo23:39
ftai said new reqs23:39
asac00:33 < fta> asac, libindicate, libiw-dev, cairo, mesa-common-dev23:39
ftanot reqs for new versions23:39
asacbut we had libiw-dev for 1.9.1 branch23:39
asaccheckout the control23:39
asaci forgot to add it then23:39
asacif its not on the other branches23:39
asacmicahg: why do you think it needs libasound2?23:40
asacthats what looked odd when i looked at your branch23:40
fta$ grep libiw-dev xul*head/debian/control23:40
ftaxulrunner-1.9.1.head/debian/control:    libiw-dev, libiw-dev (>= 29-2ubuntu6) [sparc powerpc]23:40
ftaxulrunner-1.9.2.head/debian/control:    libiw-dev,23:40
ftaxulrunner-1.9.3.head/debian/control:    libiw-dev,23:40
micahgasac: will you be on sometime between UTC 00:00 and UTC 12:00 on Sunday?23:41
asacthats true23:41
asacthe sparc/powerpc is needed everywhere23:42
asacmicahg: i can also finish the prism stuff if you don thave time23:42
asacno problem at all23:42
asacjust push your branch and i can add this on top23:42
asaci will even keep you as changelog owner ;)23:42
micahgasac: it quit on it when I was trying to build23:43
micahgwell, that's why I"m wondering if you'll be on sometime sunday23:43
micahgI can finish it sat night23:43
asacseems its needed23:43
asacyou need to add it as build dep23:43
asacnot as dep23:43
micahgok, I have build-dep on libasound2-dev23:44
asacit will automatically be detected and expanded in ${shlibs:Depends}23:44
asacif that name xists23:44
asacmaybe its just libasoúnd-dev23:44
asaci would hope23:44
asacat least if the binary is libasound2 ...23:44
micahgthere is no ${shlibs:Depends}23:44
micahgok asac, so ping me when you get on Sunday morning23:45
micahghopefully I'll be done :)23:45
asacmicahg: we need a shlibs:Depends23:46
micahgok, I'll add it23:47
asacfor the Depends: of the package that shipps the .sp files23:47
asac.so files23:47
micahgjust for the prism main package, right?23:47
asaci would think so23:47
asacwhatever file ships the .so files ... (binary componanets)23:47
asacneeds that23:47
micahggot it23:47
micahgand I have plenty of time sat night to fix this23:47
asacactually i think all need ${shlibs:Depends} and ${misc:Depends}23:47
micahgit's a lot harder than I thought, but good practice23:47
asacfor binary all misc:Depends alone should be enough23:47
asacprism is a mess23:48
micahgfor what songbird will need :)23:48
asacwell not a mess23:48
asacjust broken23:48
asacand mozilla buildsystem is kind of not obvious23:48
asacif you have not worked with it23:48
asacusualyl things are easy ... like rebase etc.23:48
asacthen they become a bit harder ...and go back to easy.... but then there are weeks where there are tough things only ;)23:49
micahgtty on sunday23:49
ftaasac, i still need to move tb3 out of umd...23:57
ftadebian 29696323:59
ubottuDebian bug 296963 in kernel-image-2.6.10-1-k7 "USB mouse continuously disconnecting" [Important,Closed] http://bugs.debian.org/29696323:59

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