[00:02] hey asac [00:03] or is it asac_ [00:03] ? === asac_ is now known as asac [00:10] fta: http://paste.ubuntu.com/236209/ [00:10] LLStarks: its me [00:11] but i have to throw myself in the bed soon ;) [00:11] asac, is autocomplete acting up for you? [00:11] https://bugzilla.mozilla.org/show_bug.cgi?id=507269 [00:11] Mozilla bug 507269 in Location Bar and Autocomplete "Address bar autocomplete doesn't always work" [Normal,Unconfirmed] [00:11] also, when will firefox-3.6 be distro package on launchpad? [00:12] LLStarks: can you give detailed instructions? e.g. how you trigger autocomplete etc. [00:12] asac, l35=l97 [00:12] i have no idea how i trigger it [00:12] fta: oh yeah. the team section at bottom is removed ;) [00:12] thx [00:13] i thought it was better to start with simple stuff like team and toools ;) [00:13] good thing its more or less an open end session ;) [00:14] so if you wake up and feel like jumping in at 9 or 9:30 i might still be acting ;) [00:14] ok with team at bottom removed: http://paste.ubuntu.com/236213/ [00:15] i'm at work at that time [00:15] that makes sense ;) [00:16] * asac is curious who will show up that early ... australians? asians? early-birds? [00:16] no one? [00:16] i just can hope that whatever set it is,its at least with me ;) [00:17] i would hope not :) [00:20] LLStarks: usuually we push things in repo on late alphas or early betas [00:20] for first time [00:24] lzma is good for the size, but the build time jumped to almost 3h [00:24] i dont think thats worth it now [00:24] chromium is alredy heavy enough [00:24] it for now ;) [00:26] heavy? you mean slow to build or heavy debs? [00:28] slow to build [00:28] currenlty PPA cycles seem to be more precious than space ;) [00:28] at least we seem to be ok with the current quota (also not much room of course) [00:29] damn. i uploaded nss jaunty to intrepid in ppa :( [00:29] now the ppa is lost [00:30] lol, with 40GB of quota, that's enough.. even if every few days, i reach 38GB [00:30] still it works for the current approach :) [00:33] have you seen this notify-osd patch from above? [00:34] mark better demand delta debs for lucky liger. [00:34] i cant parse that to be honest ;) [00:34] but its late [00:34] or courgette [00:35] i think the mozillateam is the wrong group to fix debs to use courgette ;) [00:37] asac, http://paste.ubuntu.com/236226/ [00:38] asac, so your patch could be much simpler [00:39] fta: you mean because it unrefs? [00:40] hey guys [00:41] fta: not sure what you mean. the patch definitly could make things much simpler, but i dont think the patch can be made much simpler. i do the check for the reference counting of surfaces because of the data thing that the client has to take care on its own [00:41] asac, no, your if (ptr != NULL) destroy(ptr); [00:41] ok [00:41] just call destroy [00:41] yeah that could be [00:41] it was not specified in api doc [00:41] that's why i fetched the sources [00:42] yeah. but i am not sure one should rely on undocumented api behaviour [00:42] thats why i usually dont look at sources ;) [00:42] but i will think about it... maybe there is some general sectio nthat says that all destroys in cairo are NULL safe [00:42] cairo does ref counting, so that should be pretty robust [00:43] otherwise, we'll see a lot of crashes everwhere in gnome [00:44] fta: you say alot of things in gnome call _destroy with NULL ? [00:44] not sure why refcounting means that something is NULL safe ;) [00:45] anyway. there ar emore memory issues left. at least i am down from 100 Megabyte in 20 minutes using special testcase to 1.4 m ;) [00:45] that's the basis of refcounting, don't crash ;) [00:45] well. not on NULL [00:45] the idea of refconuting is that you can unref it ... but not that you can unref something that doesnt exist [00:46] of course its there to prevent doublt free situations... but you rarely call something with NULL and ge ta double free [00:46] rarely == never ;) [00:46] either its graceful or it crashes [00:47] good. nss bits are in the ppa ... so i can sleep ;) cu 'morrow [00:47] 'night [00:49] night [00:49] I'm going to [00:50] still need to make chromium work on debian [00:50] darn lib32-gtk [00:50] rebuild it [00:50] oh, lol, more ia32 issues [00:51] BUGabundo, i will soon experiment with native x64 [00:51] it's getting closer [00:51] YAY [01:08] asac, http://groups.google.com/group/chromium-dev/browse_thread/thread/3fbd367ac70f6232 [01:38] asac: I would normally show up for that time period as it's 1AM for me :) [01:38] I won't be there tonight though... [01:38] I'll check the logs over the weekend though [01:38] I wish I could make it [08:21] fta_: annoying form thing ... posted two messags. they dont show up [08:21] fta_: if you could point them to my issue, that would be great: http://code.google.com/p/v8/issues/detail?id=411 [08:21] http://groups.google.com/group/chromium-dev/browse_thread/thread/3fbd367ac70f6232 [08:26] pinged evmar directly [09:00] anyone with hardy/intrepid/jaunty here ;)? [09:05] :) [10:29] so i will probably not so another training session at 6 UTC) [10:30] :( i miss the training .. [11:40] asac, which training session? [12:06] asac: did devicekit-power replace devicekit? [12:09] !sound [12:09] If you're having problems with sound, first ensure ALSA is selected, by double clicking on the volume control, then File -> Change Device (ALSA Mixer). If that fails, see https://help.ubuntu.com/community/Sound - https://help.ubuntu.com/community/SoundTroubleshooting - http://alsa.opensrc.org/DmixPlugin - For playing audio files, see !Players and !MP3 [12:13] gnomefreak: dont know. but sounds reasonable [12:13] at least devicekit is dead [12:18] ah [12:18] asac: thanks [13:25] asac: my feelings aside im getting the hint that multisearch shyould be removed. 2 more bugs reported today on it. [13:27] - Drop devicekit dependency. [13:27] - Add Conflicts to devicekit. [13:27] that explains that [13:27] ill be back in a few [13:33] gnomefreak: we will remove multisearch for alpha4 [13:33] that was communicated, wasnt it? [13:33] i will check if we can remove it in a week or so. i guess we have want we wanted by then [13:34] gnomefreak: please use "multisearch" tag for those bugs. thanks! [13:48] asac: they were duplicates of the 2 main bugs we have [13:48] asac: i will check to make sure they are tagged [13:49] asac: yep both are tagged already [13:49] my Lp script isnt working :( [13:52] bdmurray: as i recall you are the author of the LP greasemonkey script (from PPA) if so it seems i lost the reply choices [13:53] firefox-lp-improvements is the name of it [14:01] asac: it doesnt seem you commented about removing it on either bug. Can you please comment about it either on your blog or on the 2 bugs bug 403246 bug 402767. if on blog let me know i will comment ont he bugs linking them to your blog. blog im talking about is at: http://www.asoftsite.org/s9y/archives/162-What-is-this-Multisearch-thing-in-my-Firefox-about.html [14:01] Launchpad bug 403246 in firefox-3.5 "multisearch - please restore old NEWTAB" [Undecided,Confirmed] https://launchpad.net/bugs/403246 [14:01] Launchpad bug 402767 in firefox-3.5 "multisearch add on blocks the functionality of firefox location bar" [High,Triaged] https://launchpad.net/bugs/402767 [14:28] guys...In karmic, why is the right-click google search always searching in the Ubuntu start page? [14:28] I removed ubufox but still it is stuck with the start page :( [14:31] searching from the searchbox searches in the google.com/firefox , but the right click select always uses ubuntu start page [both in 3.0 &3.5] [14:32] fta_: asac gnomefreak : is this a known issue or should i open a separate bug^ ? [14:33] mac_v: yes for now disable ubufox should help it [14:33] gnomefreak: nope , doesnt do it :( , i removed ubufox , but still only start page [14:34] start page == your home page or ubuntu search page [14:34] ubuntu start page [14:34] mac_v: link? [14:35] gnomefreak: ubufox has nothing to do with it [14:35] multisearch ;) [14:35] is it [14:35] gnomefreak: what link? i'm confused [14:35] asac: removing ubufox helps some of them. but yes i know multisearch is the problem [14:35] http://www.google.com/cse?q= [14:35] gnomefreak: please dont tell anyone to disable ubufox [14:36] lol [14:36] gnomefreak: for the search issue. thats a wrong thing and spreads around and everyone drops ubufox ;) [14:36] asac: you started that not me [14:36] i started it? [14:36] i said: disable multsearch ;) [14:36] not ubufox [14:36] guys... gus... [14:36] guys... how do i solve this? ;p [14:36] asac: ah. thats it :( [14:37] mac_v: what is right click google search ... i dont know that [14:37] mac_v: try disabling multisearch :) [14:37] ack ;) [14:37] http://www.google.com/cse?q=fun&ie=utf-8&cx=partner-pub-2070091971271392:hsw1kx-3zxg&sa=Search [14:37] you can also check the button that says web search [14:37] asac: select text and right-click , you have the search option ... i was mentioning that [14:38] at the top of that page but not sure if that is once or once for each search [14:38] gnomefreak: i tried checking the websearch button , no go for me [14:38] mac_v: disable multisearch extension [14:38] asac: there are 4 or 5 arm patches? [14:39] mac_v: yeah disable multisearch then [14:39] there is 18_arm 38_arm bz339782_cvs and bzr 322806_ [14:39] i had 3 of them already [14:40] so i feel im missing one [14:40] hehe... so long i was wondering what was multi-search extension.... just now checked ... that bugger has sneaked in ;p [14:40] mac_v: also file a bug about "right click search going to custome search" ... and tag it multisearch please [14:40] thx [14:40] yeah disabling it solves the problem [14:41] asac: ok... but how did that extension get added? [14:43] hmm that seems those are the only ones. at least the 2 bzr were added [14:44] !multisearch [14:44] Sorry, I don't know anything about multisearch [14:44] !multisearch is http://www.asoftsite.org/s9y/archives/162-What-is-this-Multisearch-thing-in-my-Firefox-about.html [14:44] 15:44 [freenode] [ubottu(n=supybot@ubuntu/bot/ubottu)] Your edit request has been forwarded to #ubuntu-ops. Thank you for your attention to detail [14:44] sigh ... too bad that they dumped me from the ops list [14:44] i got it [14:44] !multisearch is http://www.asoftsite.org/s9y/archives/162-What-is-this-Multisearch-thing-in-my-Firefox-about.html [14:44] I'll remember that, gnomefreak [14:44] :) [14:45] sure. but i am 100% sure it worked for me a while ago ;) [14:45] lol [14:45] asac: you have to be loggede in to bot to add [14:45] @login [14:45] The operation succeeded. [14:45] asac: try that [14:46] @login [14:46] Error: Your hostmask doesn't match or your password is wrong. [14:46] @me [14:46] @whoami [14:46] !whoami [14:46] use @whoami [14:46] you dont rock ubottu [14:47] gnomefreak: we already had that a few days ago. i am not in the list anymore i am sure [14:47] asac: i am working on it [14:47] dont bother to add me again ;) ... just curious when i got removed [14:47] me too [14:47] !snack [14:47] Sorry, I don't know anything about snack [14:47] !treat [14:47] Sorry, I don't know anything about treat [14:47] damnit [14:47] !good [14:47] Usually, there is no single "best" application to perform a given task. It's up to you to choose, depending on your preferences, features you require, and other factors. Do NOT take polls in the channel. If you insist on getting people's opinions, ask BestBot in #ubuntu-bots. [14:47] !snack is not something to drink [14:47] :( [14:47] hehe [14:48] no there is one already [14:51] gedit is crashing damnit [14:51] gedit patches/series [14:51] Segmentation fault (core dumped) [14:51] asac: done ... Bug #406893 [14:51] Launchpad bug 406893 in firefox-3.5 "Right-click search goes to custom search" [Undecided,New] https://launchpad.net/bugs/406893 [15:00] thx [15:10] asac: you wanted a correct version comparison. here you have it: http://paste.ubuntu.com/236641/ [15:10] ok arm patches are added :) [15:10] asac: the only problem is, that you need all packages installed (e.g. firefox-3.0, firefox-3.5, thunderbird, prism, ...) [15:10] bdrung_: please a diff ;) [15:11] bdrung_: yeah. but you need those installed then ... i am not sure if we really want it [15:12] asac: http://paste.ubuntu.com/236644/ [15:12] asac: or would it be enough if you strip the upstream version out of the debian package version? [15:13] bdrung_: we could say that apps should add a custom header to the binary packages control file [15:14] like: Xb-Moz-App-Version: ${xulapp:version} [15:14] however, i am currently not sure if we can have access to that on the builders [15:15] asac: how do access those information in the terminal? [15:16] bdrung_: check apt-cache show mozilla-plugin-gnash | grep Npp- [15:16] those are custom headers that are in control like: Xb-Npp-.... [15:24] asac: i read your classroom session and stumbled upon the xulrunner dependency checks. e.g. "shell pkg-config --exists 'cairo >= 1.5.8'; a=$$?; if test $$a != 1; then echo 1; fi" [15:25] asac: you can simplify this or do you really want to check _exactly_ the result version != 1? [15:27] now it no longer crashes [15:28] asac: "shell pkg-config --exists 'cairo >= 1.5.8' && echo 1" should do the same. the second command is only run, if the first exits with 0. [15:39] bdrung_: yeah i know. [15:39] bdrung_: i am not sure when we did it ;) [15:39] i usually dont care for working code :) [15:39] feel free to suggest improvements [15:40] at best by suggesting merges ... lol... [15:40] yeah this training session was a bit odd. noone really replied so i think they expected something completely different and ran away [15:41] asac: oh, i wanted to only create a patch file :) (so i don't have to write a changelog entry) [15:41] asac: or they had to go to university to write an exam. [15:41] bdrung_: whatever you prefer. but i assure you that you will love bzr if you actually give it a chance ;) [15:41] bdrung_: yeah. actually the time was more for eastern europe/asian/australian folks [15:42] asac: i like bzr, but not love it. :) [15:42] but afaik free software isnt that popular [15:42] bdrung_: that whould be good enough to use it :) [15:42] should [15:42] TomJaeger: hi [15:42] hi [15:42] TomJaeger: so seems xserver 1.7 is at risk [15:43] asac: my preference would be a combinition of git and bzr [15:43] asac, yup [15:43] bdrung_: for our purposed bzr is pretty perfect [15:43] but slow [15:43] bdrung_: nowadays? [15:44] asac: yes [15:44] bdrung_: for me it feels pretty snappy nowadays ;) [15:44] at least similar to hg [15:44] asac: git feels faster. [15:44] and the difference really isnt significant to git for everything but the biggest source tres [15:44] bdrung_: i think your current perception is still biased from past experiences ;) [15:45] asac: gitweb is better. [15:45] thats a fair point [15:45] and i kick folks everytime i see it [15:45] see them [15:45] but the big problem of git is, that is complex and some command are really weird [15:45] but imo git web doesnt outweight the brain-pain that git comes with [15:45] right. i dont think its worth for packaging. really [15:46] its completely over engineered for that purpose [15:46] its for a huge development project with centralize main tree ;) [15:47] :) [15:47] asac: and for ugly packages like eclipse [15:48] packages that need git do something wrong i am sure ;) [15:48] asac: for the speed: bzr pull in mozilla-devscripts need 4 secs (with no changes) [15:48] oh wait. eclipse package is dead [15:48] bdrung_: because you go through ssh? [15:49] kernel uses git IIRC [15:49] try how quick the http: thing is [15:49] bdrung_: if you use lp:~.... it tries ssh first i think [15:49] asac: bzr+ssh [15:49] yes [15:49] asac: but with http i cannot push, right? [15:50] bdrung_: you can have http for pull and ssh for push [15:50] bdrung_: https you can [15:50] its slower by a little bit [15:50] gnomefreak: that's what asac was referring to (its for a huge development project with centralize main tree) [15:50] gnomefreak: no you cannot push [15:51] asac: i thought i did [15:51] bdrung_: for me pull taking 4 seconds is completely acceptable. you dont run pull like twenty times in a row ;) [15:51] anyway ;) [15:51] asac: i have a script which updates all my repos [15:51] apport isnt working in Kamric [15:52] bdrung_: yeaah. then pull through https ... just tried takes about 1-2 seconds [15:52] any idea on how to report a crash using apport? [15:52] bdrung_: bzr pull --remember https://code.launchpad.net/~mozillateam/mozilla-devscripts/mozilla-devscripts [15:52] asac: example i used bzr push sftp://bazaar.launchpad.net/~gnomefreak/firefox/firefox.dev [15:52] gnomefreak: you shouldnt use sftp [15:53] thats rally old and slow [15:53] asac: shouldnt but had to with firefox [15:53] gnomefreak: use lp:~gnomefreak ... [15:53] asac: i couldnt it would stall out every time push or pull using lp~* [15:53] asac: http needs 1.1 secs [15:54] and https 1.1 secs, too. [15:55] thats nearly 4 times faster. [15:56] bdrung_: yeah. thats what i mean [15:56] the other thing is bzr+ssh [15:56] so its all ssh handshake ;) [15:57] so do a bzr pull --remember with https://code.launchpad on all branches and bzr push --remember lp:~... then its fine [15:57] 3 secs for ssh. i currently tested my ssh connection to the university. they need 3 secs, too. [15:58] asac: so i blamed the wrong one. ;) [15:59] ack ;) [16:08] asac: back to xpi.mk: if all apps would provide a custom header Upstream-Version, then this patch should work: http://paste.ubuntu.com/236723/ [16:09] bdrung_: yeah. unfortrunately i found out that we have no access to any apt db during build [16:09] :( [16:09] asac: why not? [16:10] asac: isn't it possible to depend on apt? [16:11] bdrung_: apt yes, problem is you dont have athe apt db locally ... nor do you have access to network on buildd's [16:12] asac: so we need the first approach, that we have the binary packages installed. any other suggestions? [16:15] good question. [16:17] i will think a bit about it [16:17] hi [16:18] we could have two modes: a) accurate versioning (requires build depends) or b) no versioned depends [16:18] hi Willex [16:18] is it possible some update has screwed up my firefox as it struggles to start? [16:18] asac: can you explain b) more? [16:19] hi Willex [16:20] it worked fine in the morning but this evening when I rebooted nothing appears until like 5 min lag... [16:20] I tried reinstalling but still no change [16:21] Willex: did you tried to start ff without plugins? [16:21] no [16:21] bdrung_: basically extension package maintainers could decide. either they opt-into accurate versions ... which would require them to build depend on all applications he wants to have in Depends/Recommends [16:22] or he says: just add packages for the given targetapplication ids ... wihtout considering versions [16:22] Willex: you can try to start ff with "firefox -safe-mode" [16:22] well I already disabled all the addons and it now starts fluidly [16:23] so is it some broken extension update that's screwing it up? [16:23] Willex: so you can enalbe them step by step and you will catch the problematic one [16:23] I don't remember updating them recently... [16:24] or it is a ff problem in combinition with an extension [16:25] well I've had the same ones for quite some time... [16:26] hmm, seems to be Kallout that's the trouble maker here [16:26] waiting... tralalalaaa [16:28] asac: updating the changelog and doing the change in the same commit or in different ones? [16:30] bdrung_: what change? [16:31] the accurate vs. proactive [16:31] err sloppy [16:32] asac: e.g. the one regarding xulrunner (simplifying the shell command) [16:32] bdrung_: rule is: if current changelog is UNRELEASED -> extend the changelog [16:32] if its karmic or so it means that last commit was a release update [16:33] then open new changelog entry with UNRELEASED [16:33] makes sense? [16:33] asac: yes, that was clear [16:33] bdrung_: oh. do it on xulrunner-1.9.2.head branch ... thats where we usually land improvements first [16:33] asac: my question was: a) doing changes, commit, update changelog, commit or b) doing changes, update changelog, commit [16:33] as its trunk [16:33] bdrung_: ah ok. [16:34] bdrung_: doing change, update changelog and use debcommit [16:34] so the changelog entry is used for bzr commit of the change [16:34] ok [16:42] asac: here you are: https://code.launchpad.net/~bdrung/xulrunner/simplify-shell-command/+merge/9469 [16:49] asac: you can find UNRELEASED twice in the xulrunner changelog. something went wrong there... [16:56] let me check [16:56] bdrung_: yeah. thats the 1.9.1 changelog still ;) [16:59] asac: how to proceed with xpi.mk? [17:04] bdrung_: a bit of a difficult question. [17:04] bdrung_: not from xpi.mk side, but from what we require applications to do [17:04] xpi.mk should get a switch like ..._XPI_AUTO_VERSION [17:05] and if its there use whatever mechanism we define for targetapplication packages to expose their version [17:05] also we could consider to define a version mapping for packages so we could guess the right lower/upper version bounds for depends [17:06] for package versions [17:06] let me think a bit ;) [17:06] it's now late enough to think clearly ;) [17:07] asac: my goal is it to simplify the plugin packages. so simplifying depends, but the having to b-d on the ff, tb, etc. would destroy the win === dpm is now known as dpm-afk [17:21] bdrung_: yes. so if we could invent a mozversion -> deb package mapping, we could generate proper version bounds without having them installed [17:22] so we might have impossible combinations like thunderbird (>= 3.0~a1) [17:22] but we could maintain a list of special EOL and SOL versions for certain packages [17:22] ;) [17:22] like thunderbird_sol = 0.1a1pre [17:22] thunderbird_eol = 2.0.0.* [17:23] what do you mean with eol and sol? [17:23] thunderbird-3.0_sol = 3.0a1pre [17:23] bdrung_: eol = end of life ;) [17:23] aha [17:23] sol the opposite ;) ... version of birth like ;) [17:25] not birth date, but fecundation date ;) [17:26] asac: you said that in the building we cannot run apt-cache. how should the mozversion -> deb package mapping work? [17:31] asac: now i understand it. [17:41] asac: here's the implementation of your idea: http://paste.ubuntu.com/236918/ [17:42] afk [17:46] morning asac [18:03] re [18:04] asac: we could set the *_sol to the current upstream version of the packages in unstable/karmic, couldn't we? [18:58] asac: wow, NEW contains nearly 300 packages. === dpm-afk is now known as dpm [19:39] asac: packages should b-d on xulrunner-dev instead of libxul-dev, right? [19:51] asac: do you know how to sru a new upstream version? https://bugs.launchpad.net/ubuntu/+source/adblock-plus/+bug/320762 [19:51] Launchpad bug 320762 in adblock-plus "Please backport adblock-plus 1.1-0ubuntu1" [Wishlist,New] [20:24] bdrung_: yes xulrunner-dev is the right one [20:24] bdrung_: thtas a backport request. not an SRU? [20:25] bdrung_: ok. it basically needs a bug that is severe enough to justify a backport [20:25] asac: the backport team said, that sru fits better than backporting [20:25] asac: maybe we should adjust the severity [20:32] bdrung_: we should make a bug out of it ... like: "adblock-plus filters broken" [20:32] what exactly is broken? [20:33] asac: some filter expressions do not work, because only the newer version supports the syntax. so less spam is filtered. [20:33] asac: so it is not totally broken, but partly. [20:34] bdrung_: i remember that there are different filterlist providers [20:35] are the new filter expressions used by new providers? or by the old ones? [20:35] asac: and depending on how aggressively they use the new syntax... [20:35] asac: by the old ones. [20:35] asac: they are used by (nearly) all providers listet on adblock website [20:38] bdrung_: what i mean is: users that have the old adblock extension ... do they already pull data that they dont properly understand? [20:39] asac: yes. the lists are automatically updated (weakly) [20:47] fta_: i reckon this command works for formatting the commit date like you suggested GIT_DATE = $(shell cd $(TMP_DIR)/src && git rev-list --max-count=1 --date=rfc --format=%ad HEAD | tail -1 | xargs -i date +%Y%m%dt%H%M -d'{}') [20:48] fta_: if you have a better/more compact command... please share [20:49] bdrung_: ok make a bug out of it. change severity and subscribe motu-sru [20:50] asac: other topic: should we set the *_sol to the current upstream version of the packages in unstable/karmic? [20:50] bdrung_: but i think we cannot use the same packaging. so we need to branch off the current jaunty version and maintain a stable branch for jaunty and before [20:51] bdrung_: no. i think _sol is basically first upstream version that matches the branch associated with the name [20:51] fta_: so now the version looks like this... 0.0.3+git20090717t2107x65f7bc7 [20:51] i stuck the hash on the end for good measure [20:51] bdrung_: example: thunderbird-3.0 -> 3.0a1pre [20:51] pace_t_zulu: very long version string [20:52] bdrung_: indeed... that's just for the ppa... for unreleased commits [20:52] bdrung_: firefox-3.5 -> 3.5a1pre ... [20:52] pace_t_zulu: then it's ok. it would be more readable if the date is seperated from the rest [20:53] bdrung_: how do you suggest i do it dfferently [20:54] pace_t_zulu: for git checkouts i used only the date, e.g. 0.0.3+git20090717 and if you need more than one build per day, add the time 0.0.3+git20090717-1704 [20:54] bdrung_: alternatively we could say that it should be the first version that entered the archive [20:55] have to think a bit about it ... but i currently lean to the first upstream version [20:55] asac: and why not the current version in the archive? [20:55] bdrung_: you wouldn't include the hash? [20:56] pace_t_zulu: no, you could mention the hash in the changelog. [20:57] bdrung_: because that would be overly strict imo [20:57] s/you could/i would/ [20:57] asac: but it fits for the normal user. [20:57] bdrung_: think about packages. you wouldnt bump lower version if there is no incompatibility [20:58] bdrung_: why do you think its better to use current version as sol? [20:59] asac: example: we have thunderbird 2.0 in the repo, but the _sol is 0.1a1pre. [20:59] yes [20:59] asac: we have an extension which works with tb till 1.0. [20:59] because what we use _sol for is to figure if it makes sense to have thunderbird as a dependency [21:00] asac: so it would not work with tb and so tb should not be listed as possible dependency. [21:00] bdrung_: no. it should be listed, but with proper bounds [21:00] thunderbird >=0.5 <= 1.0.* [21:01] asac: ok, with bound it would be ok. [21:02] so if we use current ones we would get overly strict depends [21:02] asac: then the xpi.mk need finetuning [21:02] if we use upstream sol /eol we would end up with overly lax depends until we do the next and final step and get perfect bounds [21:03] imo overly lax is better because it does not prevent you install for targets that work [21:03] also its the right way for the final solution imo [21:03] asac, are you available for an upload later? [21:04] not today [21:04] asac, I've decided to take back libagg coz someone wanted to NMU it [21:04] tomorrow might work [21:04] asac, my other packages are all ok so I can have one more package to work on [21:04] asac, I gonna leave for the sea at like 2-3 pm [21:04] bdrung_: the final solution would require a version scheme for packages .. e.g. policy [21:05] asac, can you make it for the morning? [21:05] no [21:05] i cants say when [21:05] damn : / [21:05] i have a haystack of work ;) [21:05] asac: how would the version scheme look like? [21:06] bdrung_: i think its pretty simple (not sure about *) ... 2.0a1pre -> 2.0~a.1~pre [21:07] maybe + instead of . [21:07] ? [21:07] 2.0~a+1~pre [21:07] i am not sure if there can be a perfect mapping ;) [21:07] asac: i would prefer 2.0~a1~pre [21:08] bdrung_: does it always work if we dont split letter to number transitions? [21:09] asac: yes. we only have to split number to letter [21:09] bdrung_: maybe it really works [21:09] asac: because their interpretation is that an empty string is lower than everything else [21:09] dpkg --compare-versions a1 eq a01 && echo asdsa [21:10] a1 is equal to a01 ... so dpkg seems to split those transitions somehow [21:11] * asac didnt know that a01 == a1 ;) [21:11] even a1 > a-2 works [21:11] bdrung_: is it equal in mozilla land too ;)? [21:11] asac: yes, they allow negative numbers [21:11] bdrung_: no i mean is a01 == a1 ;) [21:11] let me test with our script ;) [21:11] asac: yes, it was one of the examples [21:12] works ;) [21:12] ok so lets try to standardize that with debian moilla and debian moz extension maintainers [21:12] i will talk to extension guy ... and ask him to ask instead of me ;) [21:13] asac: mozilla mailing lists i should subscribe? [21:14] bdrung_: no .. debian mozillateam [21:14] mozilla has no stake in deb versions ;) [21:14] asac: link? [21:14] let me get it [21:15] http://lists.alioth.debian.org/mailman/listinfo/pkg-mozilla-maintainers + http://lists.alioth.debian.org/mailman/listinfo/pkg-mozext-maintainers [21:15] bdrung_: ^ [21:15] asac: and ubuntu lists? [21:16] bdrung_: we only have our mailing list (/topic) [21:16] hmm ... there is no url ;) [21:16] https://lists.ubuntu.com/mailman/listinfo/ubuntu-mozillateam [21:17] bdrung_: ^ [21:18] asac: done. :) the ubuntu list was the fastest (although i subscribes as last) [21:23] asac, k [21:23] asac, gonna prepare everything for you for tomorrow [21:23] asac, it's just a bug fix anyway on amd64 [21:25] asac: pkg-mozilla-maintainers needs a spam filter [21:26] yeah [21:26] debian doesnt want spam filter ;) [21:26] why? [21:28] bdrung_: i wasnt accurate. debian has tradition to not require a subscription [21:28] and since alioth has no spam filtering (compared to debian-devel) its a mess [21:28] but even debian-devel gets loads of spam through afaik [21:28] but they have pretty good mailadmins that tweak spamfilter quite well [21:30] asac: where to report the spam filter request for alioth? [21:33] let me check if i can enable spam filtefring for that list [21:34] hmm ... i am not admin i think [21:34] of the list [21:34] just the project [21:35] i dont know where to file something. sorry. [21:39] asac: what was the conclusion of the _sol version? using the version of the first package release? [21:44] hey . gonna get my self a Android G2 tonight! [21:47] asac, does it matter if I gonna provide you orig diff e dsc files instead of branches? [21:56] pace_t_zulu: for getting the latest git commit date, i used: "git log --pretty=format:%h -1" [21:56] bdrung_: i will test that now [21:57] pace_t_zulu: your command looked much more complex [21:57] bdrung_: indeed... [21:58] pace_t_zulu: ups [21:58] pace_t_zulu: that command was only for the hash. for the date i used the current one (date +%Y%m%d) [21:59] ? [21:59] pace_t_zulu: and i used "${VERSION}+git$(date +%Y%m%d).${GIT_HASH}" as version string. so i did include the hash === fta_ is now known as fta [21:59] hi [21:59] bdrung_: i think so. yes. [21:59] (booo, localtime, baad) [22:00] asac, so, how was your session? [22:00] bdrung_: that date command will only get the current date... not the commit date [22:00] asac, does it matter if I gonna provide you orig diff e dsc files instead of branches? [22:00] pace_t_zulu: yes, i know. i thought i improved that. [22:01] bdrung_: the current date isn't as useful as the commit date [22:01] andv: i think only you can answer that [22:01] asac, lol [22:01] asac, nvm then :S [22:01] * :D [22:01] a) i have no clue what package you are talking about ... b) i have no idea if there were any branches used before/in future [22:02] etc. [22:02] asac, libagg, there are no branches yet [22:02] asac, there will be some branches in the future, but I gonna setup them after the package is uploaded (fixes a grave bug) [22:03] asac, so on monday when I get back, I'll make some working branches [22:03] k [22:03] asac, gonna provide you a dgettable url [22:03] asac, don't worry [22:05] andv: all ubuntu packages are now in bzr branches. so you could easily branch the karmic version. [22:06] bdrung_, are you talking about my package? [22:06] (libagg?) [22:06] andv: https://code.launchpad.net/ubuntu/+source/agg [22:06] andv: yes. [22:07] yeah. should be easy to downstream maintain packages in debian from ubuntu [22:07] bdrung_, have to upload a package as soon as possible, so I won't have time to do all revisions in bzr [22:07] bdrung_, on monday I'll set them up [22:08] bdrung_, for now I gonna work with orig diff dsc file [22:08] andv: it's only a hint for the future work. [22:08] bdrung_, appreciated ty [22:08] bdrung_, I'll have to merge the package anyway [22:08] I see two ubuntu patches there [22:08] for fixing something in jaunty [22:09] andv: you're welcome [22:09] didnt know all packages are branched now [22:10] it's for a short time now [22:10] bdrung_, how does it work? [22:10] bdrung_, they get synced on bzr automatically? [22:11] andv: yes. every time a new package is uploaded a new commit is generated. [22:11] cool [22:12] andv: i don't know how fast it is. it's probably a cron job which runs every hour or every day or something in between [22:12] yeah [22:12] looks like a good idea [22:14] asac, http://dev.chromium.org/user-experience/omnibox may give you idea for the addon ;) [22:14] +s [22:23] fta: idea for which addon? [22:24] multisearch? [22:26] yep [22:26] roh, i'm kidding [22:28] does chromium not have a search field by default? [22:31] no, it's omnibox [22:33] k [22:34] fta: we have to decide what to do with dailies [22:34] err .head for xul + ffox [22:34] i will transition stuff like yesterday ;) [22:34] at firs tadding xulrunner-dev to 1.9.1 [22:35] then transitioning rdepends and then moving firefox to 3.5, enablling branding and asking profile migration question [22:37] so i see two questions ew have to answer with yes to just continue: [22:37] a) is it ok to migrate daily users to xulrunner-dev 1.9.1 by default [22:37] b) is it ok to migrate daily users to firefox 3.5 by default [22:37] i dont think its a big issue ... and i think thats aktually what most probably want, but still it hsould be an explicit decision that you also agree with ;) [22:37] fta: ? [22:38] rather than me injecting that in .head and then it happens everywhere [22:40] i guess it's fine [22:40] yeah only thing i am a bit unsure about is xulrunner-dev [22:40] maybe they want to respin a hardy package and then it breaks [22:40] but lets assume that those that need this can use pinning [22:41] and the rest doesnt care most likely [22:43] they should respin in chroot / pbuilder, not on their own box [22:47] asac, a linux magazie in italy posted the link to chromium daily ppa [22:48] asac, on this month [22:48] * magazine [22:49] bdrung_, do ppa builds for sid too? [22:50] no [22:50] andv: no [22:50] k ty [22:50] andv: and only x86 and amd64 [22:50] lpia [22:50] I gonna change the changelog to karmic then [22:50] (due to virtualisation) [22:51] asac: lpia counts as x86 for me [22:51] still they build separtely [22:51] andv: remember to keep version in ppa lower than the actual upload [22:51] ~andv1 [22:51] ok [22:52] i wonder if it's possible to hack apport to send crash reports to upstream somehow [22:53] google would like that [22:54] fta: can google deal with our crash files? [22:54] the blocker usually is that they need our symbols [22:54] no, they use breakpad [22:54] right so its the same thing as with mozilla [22:54] yep [22:54] breakpad is actually google, not mozilla ;) [22:54] fta: the blocker is not really the sending of the crash report [22:54] i knew [22:55] its the "how to get the symbols uploaded and keep them versioned in their db so they can find the right ones etc." [22:55] i would be happy if we solve this with google directly ;) [22:56] i often resolve my crash files locally [22:56] you do. but most dont have the huge -dbg packages availabl [22:56] and it shouldnt be required [22:58] that's something lp should do [23:00] fta: what should launchpad do? [23:00] fta: google has to be ready to accept uploads of symbols from distros [23:00] lp has all the debs, fresh and not so fresh, it should be able to resolve all the crashes that are interesting to have and feed the results to a tracker specified in the package [23:00] and later be able to assocaited those symbols [23:00] fta: you say forward the bugs. ok [23:01] thats a different thing [23:01] no [23:01] than uploading the crashes to breakpad [23:01] just an apport service [23:02] we discussed forwarding retraced crashes to their db [23:02] but there were problems too [23:02] i would like to add some apport hooks to my packages to teach apport what it should do with resolved crashes, instead of posting them in a lp bug, or reject it because "it's not a genuine ubuntu package" [23:03] fta: unlikely that apport will get a feature to run code etc on the retracers machine [23:03] fta: i understand. but where should it post the results to? [23:03] fta: you said know if i asked if those should be auto fild upstream [23:03] a remote bug tracker [23:03] err you said no ;) [23:04] not to lp, then forwarded, but directly upstream [23:04] ah ok [23:04] but still in launchpad ? [23:04] nm [23:05] apport will run on the ubuntu side, as all the debs are there [23:06] so that would involve: a) [23:07] make an apport webservice that is the primary upload service, rather than launchpad bugs [23:07] yes [23:07] something like that [23:07] specify a contract that upload plugins can implement [23:07] and use that after retracing to do the upload [23:08] main problem is really that the retracers machine are deep in a datacenter ... not sure if someone would wnat to give them net access [23:08] but one has to find out [23:08] something we should suggest to pitti [23:08] could be proxified, like pushed in a queue and processed by another machine [23:09] yeah. the other thing is that we are talking about privacy related data [23:09] so if users upload to ubuntu ... they dont expect google to get the private data [23:10] not sure if that is relevant [23:10] but users even complained that we violate privacy law in many states by putting multisearch in [23:10] even though we just get the amount of clicks done where ;) [23:11] and all the requests still go to google [23:12] well, it's no different that going through breakpad [23:12] fta: its different [23:12] well of course it applies to the idea of us uploading to breakpad [23:12] as well [23:12] maybe i should just find a way to bypass apport and use breakpad then [23:12] i meant its different if users directly upload to breakpad [23:12] and we provide te symbols [23:13] fta: as i told you thats the best way. but the problem is somewhere else [23:13] its to get the symbols uploaded ... and afaik breakpad is not ready out of the box to do that [23:13] if we can solve that i would jump up and down for 10 minutes ;) [23:21] not enough ;) [23:21] https://code.edge.launchpad.net/~mozillateam/ [23:22] fta: ok one time a week 10 minutes additional push ups for a month;) [23:22] with each month it takes till this happens i remove 1 minute ;) [23:22] bug? they all have the same graph [23:22] !date [23:22] Sorry, I don't know anything about date [23:23] fta: yeah [23:23] thats what i wanted to say ;) and forgot to write it after the link [23:23] not sure yet which one they take [23:23] i had a different one before i reloaded [23:32] anyone using chromium and having crashes on shutdown? [23:33] http://code.google.com/p/chromium/issues/detail?id=18112 [23:35] fta: no [23:35] works [23:35] since when does ithappen? [23:35] with --enable-plugins? [23:35] i noticed a few days ago [23:35] fta: i use the dailies [23:35] i am upgrading to latest now [23:35] but i think i upgraded yesterday [23:36] try with --enable-plugins [23:36] on the cmd line or in /etc/chromium-browser/default [23:37] upstream noticed some crashes when totem plugin is installed, but that's supposedly fixed now [23:37] fta: do i need to use something that trigers plugins? [23:37] or jus topen and cloe? [23:37] no [23:37] argh [23:37] typing spastics [23:37] --enable-plugins works too [23:37] i am on 32bit [23:38] hmmm ... even works with nspluginwrapper ;)? [23:38] nice ;) [23:39] hmm. seems they dont resolve links [23:39] so i have multiple entries now [23:44] asac, sent to build on ppa [23:45] asac, if it builds fine on amd64, problem is fixed [23:48] asac, it crashed if i have acid3 in a tab [23:49] -d+s [23:54] no need to report every progress. just the big steps :) [23:54] asac_, k