/srv/irclogs.ubuntu.com/2009/05/15/#ubuntu-devel.txt

asacpace_t_zulu: stripping down the third_party libs/source that are required in the source tree is a continuous task i would say00:00
asacpace_t_zulu: there might be some third party sources that could be packaged (e.g. skia for instance)00:01
asacfta: do you know how stable skia is abi/api wise?00:01
ftai don't think it is, it's very chromium oriented00:02
pace_t_zuluasac: http://skia.googlecode.com ... make a skia.deb or libskia.deb ?00:02
asachmm. then i dont see why they dont fix the skia font api to allow proper on-demand fallback through fontconfig00:02
ftaand chromium is now pulling skia from trunk00:02
ftasame for v8 and webkit00:02
asacfrom how the bug read they seemed to rather not change skia00:02
asacfta: yeah. but trunk doesnt necessarily mean that the ABI/API is unstable00:03
asacbut i see what you mean.00:03
ftaasac, they are evolving in parallel. so we'll have to bump them at the same pace. like ff and xul00:04
pace_t_zuluasac: i suppose their modifications to webkit make it unreasonable to depend on libwebkit00:04
pace_t_zulu?00:04
asacfta: yeah. but skia is also used by gears00:04
asacso i thought there was potential00:04
ftagears is small in comparison00:04
asacyes, still needs reduction :)00:04
ftapace_t_zulu, upstream said chromium will never work with libwebkit00:05
ion_Why is that?00:05
pace_t_zulufta: *never*?00:05
ftatricky. their copy of webkit is built with chromium specific flags, and it needs the chromium sources to build00:07
pace_t_zulufta: i know they wanted to push their webkit modifications upstream...00:07
ftathey did, for the most part00:07
pace_t_zulufta: i understand why it wouldn't right now... but *never* seems quite absolute00:07
ftato be more accurate, when i asked how far it was from building with a system webkit, they said "pretty far. We'll never work with something like webkitgtk."00:09
ftaif you look at the code, it's easy to understand why00:09
pace_t_zulufta: fair enough00:10
ftai want as many system libs as possible since day 1, but it's not the priority upstream, they want feature parity with windows 1st00:12
pace_t_zulufta: understandable... i'm with you on system libs00:12
ftai worked last year to move about 12 libs to system libs, but it was creating so many regressions in the testsuite that i decided to fallback00:14
ftamost of our libs are sort of "not good enough"00:14
ftaour libjpeg for example, far too old00:15
ftawe lack about 10 years of patches and optimizations00:15
pace_t_zulufta: why would we have such an old libjpeg?00:17
ftai introduced the -testsuite package to easily spot regressions when moving libs to system libs, but the testsuite needs some love00:17
ftapace_t_zulu, because upstream is dead00:18
asacfta: do we know where they too the libjpeg from?00:18
ftayes, from mozilla00:18
asacfta: maybe they copied firefox libjpeg which seems to be a fork00:18
asacyeah00:18
ftamozilla should be the new upstream, the jpeg consortium died 10y+ ago00:18
asacright00:18
pace_t_zuluwhat if we package a 'libjpeg-mozilla' ?00:18
asacso we might wnat to consider that00:18
pace_t_zuluor something like that00:19
ftaasac, i asked mconnor, he liked the idea00:19
pace_t_zuluhow does wegkitgtk get away with using libjpeg?00:19
asacwell. for that i would at least like to know someone who can explain to me what changes were done :)00:19
asacbut even in mozilla tree most changes are so long ago that its hard to find out00:20
pace_t_zulufta and asac: did you guys see my blueprint for giving chromium a "Human" theme?00:21
ftai tried to split the patches between moz trunk and the last known upstream jpeg, got like 110 commits since the "free the lizard" day (when the netscape code was freed), plus a giga commit before that00:21
ftachromium is not yet "themable"00:21
asacright. we looked at that at some point. i stopped looking at patches when there were commits without even a bug referenced00:21
pace_t_zuludoes mozilla have a libjpeg or has it become too much embedded in the mozilla codebase?00:22
ftathey have their own version, but it's easy to split00:23
ftai wanted a clean patch stack on top of our lib, it turned out to be almost impossible00:24
pace_t_zulufta: what if we were to try to split it and create a libjpeg-mozilla ? whatever the name may be00:24
ftait only makes sense if it replaces our lib00:24
pace_t_zuluor is that undesirable... i suppose patching our libjpeg is what you are looking for00:25
pace_t_zulufta: right...00:25
pace_t_zulufta: i assume our libjpeg comes from Debian... correct?00:26
ftayes00:26
pace_t_zuluasac: i know you are on the firefox team... would it be useful to build firefox against a patched libjpeg?00:27
ftaone experiment could be to package from the mozilla sources and see if it creates regressions in gnome00:27
ftawe used to build firefox with system jpeg, until it started to crash when printing00:28
ftaand it was also much slower as we don't have the mmx/sse/sse2 optimization mozilla have00:28
ftaso i guess the whole desktop could benefit from that00:29
pace_t_zulufta is this something that is doable before karmic is released?00:31
asaci think its worth to look at that during UDS00:31
lifelessTheMuso: btw I'm done on the isw bug00:31
lifelessTheMuso: nothing further to do at this point pending that kernel patch and upstream waking up00:31
ftapace_t_zulu, everything is doable, depending on how much time you can spend on this ;)00:31
pace_t_zuluasac how long would it take to pull libjpeg from the mozilla codebase?00:32
pace_t_zulufta, like i said... i am very interested in contributing...00:32
pace_t_zulufta, i am also a fast learner00:33
ftato pull libjpeg from the mozilla, hm, 1 sec ?00:33
ftawell, with hg, more, you have to clone the full trunk00:33
pace_t_zulu!hg00:34
ubottuSorry, I don't know anything about hg00:34
asaci think its technically easy00:34
asacthe problem is more of figuring what we really want00:34
ftahttp://hg.mozilla.org/mozilla-central/file/38d50cc03a72/jpeg/MOZCHANGES :(00:35
TheMusolifeless: right00:35
fta"????/??/??  -- Lots of undocumented changes. :("00:35
ftaasac, eh.. a fast libjpeg creating no regression? :)00:38
pace_t_zulufta ftw00:40
pace_t_zuluthe discussion has fallen silent00:47
asaci think there is a consent that this is a good idea and that we will look at UDS at this00:49
ftaasac, is mconnor coming?00:50
asacnot that i know ;)00:50
pace_t_zului would love to go to UDS00:51
pace_t_zulumaybe next one00:52
pace_t_zuluasac: http://pastebin.ubuntu.com/172669/00:52
asacyou need to install a package called bzr-builddeb00:52
pace_t_zuluasac, thank you00:53
pace_t_zulushouldn't that be in ubuntu-dev-tools?00:53
pace_t_zuluhmmm....00:54
pace_t_zuluthe builddeb is checking out a new copy of chromium00:55
pace_t_zuluseems as if fewer commands are needed00:56
pace_t_zului'm learning here... i'll figure out why it is going through these steps00:56
asacpace_t_zulu: i think we should move that discussion to #ubuntu-motu ;)00:57
pace_t_zuluasac and ftw: thank you for being patient w/ me00:57
asacpace_t_zulu: you didnt place the tarball in the right dir00:57
lamonthrm.. someone correct me if I'm wrong... when I say 'extract' to sound juicer, it's not supposed to immediately terminate, right??? :-(01:20
lamontin jaunty01:20
lamontwhere does it drop logging/watever?01:20
lifelessvalgrind sound-juicer01:20
lifeless:)01:20
lamontstrace comes to mind as more appropriate... this is totally immediate01:21
lifelessyeah01:21
lamont--- SIGSEGV (Segmentation fault) @ 0 (0) ---01:21
lifelessyou're thinking 'deliberately quit01:21
lifelessI'm thinking 'fail01:21
lifelessand.. I win :)01:21
lamontno, I wasn't thinking deliberately quit01:22
lamontseb128 wouldn't do that to me01:22
pace_t_zuluso asac... i'm not seeing a libjpeg in the firefox source01:33
pace_t_zuluasac: would libjpeg need to be checked out directly from mozilla?01:33
pace_t_zuluasac: i suppose i need to learn mercurial now :)01:34
lamontlifeless: never try to extrack with a blank title01:44
lifelesslamont: fun01:51
lifelesslamont: This is why writing a desktop environment in C is a good idea01:51
lamontwell the alternative is to have it perform like it's interpreted. :-p01:52
lamontbrb01:52
Sarvattis the gcc atom patch going to be permanently dropped now then? :(04:32
quentusrexI think I've found a memory issue with Xorg04:51
quentusrexXorg is currently taking up 1.8GB or residential ram on my box that has 4 GB of ram.04:52
Chipzzquentusrex: 1) this is not the right channel to report bugs; 2) there is #ubuntu-x (or ubuntux, dunnow off-hand) and 3) are you sure it actually takes up that much? what top reports is not the amount of memory it actually uses05:05
Chipzzthat being said, please move to #ubuntu-x05:06
dholbachgood morning06:28
lifelesshmmm, must fix dmraid udev rules06:39
lifeless'dmraid -ay' in console on every boot isn't pretty06:40
TheMusolifeless: I'd be interested in why either dmraid-activate is not being run, or a log of dmraid-activate with -x itself.06:43
lifelessTheMuso: have you seen other like issues before?06:44
lifelessTheMuso: if so point me at bugs and I'll just fix this06:44
lifelessdmraid-activate is in the initramfs06:44
TheMusolifeless: yes, and no I haven't seen issues like this so far as I have determined.06:46
TheMusolifeless: no hurry, just whenever you can get a log of dmraid-activate running in the initramfs06:47
TheMuso./c06:47
lifelessok06:49
lifeless358255 may be related06:49
lifelessthough I'm not using lvm its possible the root cause isn't an interaction06:50
lifeless(filing a bug to track my case)06:50
lifelessTheMuso: Is there anything more I need to do, or you'll land it in karmic?06:55
TheMusolifeless: to fix that problem? I will land it in karmic first, and if needed, can SRU for jaunty.06:55
lifelessas long as dmraid isn't altered in jaunty I don't care if you do/don't06:56
lifelessbut if it gets changed, I'd really like the patch in, so that it doesn't regress :)06:56
TheMusoOk.06:56
lifelesshow do you log output from a initramfs ?06:57
TheMusoWhat I'd do, is modify dmraid-activate to log a file output to /dev/.initramfs/filename and set -x06:57
lifelessthanks06:58
pittiGood morning06:58
lifelesshi pitti06:58
pitti05/15/09 02:40:02: checking #376771 for duplicate07:01
pittiDEBUG: blacklisted from auto-duplication 26000107:01
pitticjwatson: ^ seems to work07:01
dholbachcan someone please get ibus-table-extraphrase out of binary new?07:22
saravananhi07:26
dholbachara: you rock! :)07:38
aradholbach: :)07:39
c_korn'%WN?hzJ9".pzG3MRPpkD08:08
c_kornoops, this was my passphrase :P08:09
dholbachc_korn: it very much looked like one :)08:09
pittic_korn: very good one, though; too bad it got busted now08:11
pittimust have taken a week to learn it08:11
dholbachcan an ubuntu-archive member please get ibus-table-extraphrase out of binary new? :-)08:11
pittidholbach: done08:13
* dholbach hugs pitti08:13
dholbachthat'll make reviewing, building and testing the other ibus-* packages on REVU a bit easier :)08:13
=== tkamppeter_ is now known as tkamppeter
tkamppeterpitti, hi08:39
tkamppeterpitti, I have uploaded the SRU for bug 376732 now.08:39
ubottuLaunchpad bug 376732 in foomatic-filters "Hardy foomatic-rip PageSetup is broken, fails LSB tests" [Undecided,Fix released] https://launchpad.net/bugs/37673208:39
pittitkamppeter: I had some questions in the bug report08:43
tkamppeterpitti: I have answered in the bug report. It is really a regression against Hardy and the patch to fix it is MUCH smaller than the diff attached to the original LSB bug report.08:49
pittibryce: if you have scripts which evaluate HalComputerInfo from apport-generated bugs, you need to adapt them; I ripped out hal from apport now and replaced it with udev db/log. Please let me know if you need any help with that09:19
brycepitti: I don't actually interpret HalComputerInfo in particular, but that's good to know09:22
pittibryce: ok; I checked the source_xorg.py hook, and it shouldl still be okay; I'll run a few tests, just to be sure09:22
brycepitti: I've been thinking it would be nifty to have some stock python code that reads/writes a data "structure" from the bug description that provides basic hardware/software version information09:22
pittibryce: you know that there's code to convert a LP bug report back into an apport.Report dictionary?09:23
pittibryce: would that help?09:23
brycesomething like,  "foobar: 1.2.3\nbarfoo: 3.2.1" etc.09:23
brycemaybe; I'll have to take a look09:23
pittiphone, brb09:23
pittibryce: so you would say "report = download(12345)"09:24
pittibryce: and then could access report['Package'], report['Uname'], report['UdevDb'] etc.09:25
pittibryce: the download call is about two or three lines of code09:25
brycepitti: I think what I'm thinking of is a touch orthogonal09:30
brycepitti: I'm thinking more along the lines of a plain and simple set of "key: value" pairs in the bug description09:31
brycethat can be loaded, modified, and saved back into the bug09:31
pittibryce: don't we have that already?09:31
pittinot the 'save back' of course09:31
brycepitti: well that's the magic; it builds on what is already there09:32
brycejusts adds a more read/write aspect to it09:32
brycewhich allows us to "accumulate knowledge" as the bug progresses09:32
pittiright now, crashdb.update() for launchpad only attaches new comments09:33
pittibut it should be possible to make it change the description09:33
brycewhat we don't have right now is an API for doing this09:33
brycebut it ought to be pretty trivial.09:34
pittibryce: I think I have a rough idea about it; let's talk about it in Barcelona, shall we?09:40
brycepitti: sounds good09:41
brycepitti: won't have to wait long :-)09:41
directhexi think i've lost my euros again. i should put them somewhere safe09:42
=== dpm_ is now known as dpm
loolpitti: Can apport be reenabled now?  ISTR some work is ongoing for soyuz support of ddebs09:48
pittilool: that's pretty independent of soyuz09:49
seb128please don't reopen apport now, we will be flooded by crash bugs, it's early in the cycle, after uds rather09:50
pittiit's primarily a matter when we start wanting to open the floodgates09:50
loolWaiting is fine for me; I'll just turn it on where I need it09:50
seb128pitti: the gvfs gdu monitor crash every time I start gedit on my desktop btw, did you get the same issue?09:50
seb128I noticed yesterday because I enabled apport to take screenshots09:51
pittiseb128: I have apport enabled, too09:52
pittigedit starts fine here09:52
seb128pitti: gedit starts fine but I get a gdu crash after that09:52
pittihm, I don't09:53
seb128pitti: https://bugs.edge.launchpad.net/ubuntu/+source/gvfs/+bug/37614509:54
ubottuUbuntu bug 376145 in gvfs "gvfs-gdu-volume-monitor crashed with SIGSEGV in gdu_pool_get_presentables()" [Medium,Triaged]09:54
loolCan I get apport to run with a SIGABRT?09:54
seb128no09:55
seb128there is a spec about that I think09:55
* pitti is on the phone09:55
loolAh I can't run strace under valgrind09:56
loolI guess both use ptrace09:57
seb128lool: https://blueprints.edge.launchpad.net/ubuntu/+spec/security-karmic-apport-abort09:58
cjwatsonI wouldn't have thought valgrind used ptrace - it's more like an emulator09:59
loolseb128: thanks10:00
loolcjwatson: Me neither; it's the only reason I could think of for its non-transparency10:00
=== fta_ is now known as fta
loolActually "valgrind strace ls" works fine10:03
loolSo I get permission denied under valgrind + strace and not under strace10:06
tjaaltonpitti: hal ships 10-x11-input.fdi, but so does evdev in debian. maybe we should take the one from evdev instead?10:07
pittitjaalton: is that in upstream evdev? (since it's in upstream hal)10:08
pittiI wouldn't mind removing it from hal upstream10:08
pittitjaalton: OTOH, xorg's -evdev should better be converted to not use hal at all any more :)10:08
pittiKeybuk: apart from /var/log/udev and an udev db dump, can you think of anything else an udev-extra apport hook should collect?10:11
Keybukpitti: could we get a udev hook to do that at the same time? :p10:11
pittiKeybuk: hm, perhaps we should add that hook to udev itself, and  udev-extra just ships a symlink to it10:11
Keybukdmesg/kern.log is often helpful10:11
pittiKeybuk: snap :)10:11
pittiKeybuk: attach_hardware() does that, too10:12
pittiKeybuk: in fact, attach_hardware() already has all of that10:12
pitticpuinfo, cmdline, lsusb, lspci, DMI, udev db, udev log, dmesg10:12
Keybukyeah, that bunch10:13
pittiKeybuk: will do that then10:13
tjaaltonpitti: actually not from upstream, debian added it. well, I don't mind keeping status quo for now10:14
pittitjaalton: since I hope all of this will disappear very soon, I don't particularly mind either :)10:14
tjaaltonpitti: you can hope, but upstream has no plans beyond hal :/10:15
pittitjaalton: do you happen to know if there's some upstream movement to move away from hal towards asking udev for input devices?10:15
pittihm, is that a lot of code?10:15
tjaaltonpitti: the server already supports dbus10:15
pittiI can't imagine xorg would have been infested by lots of hal specific code10:15
Keybukpitti: the cabal certainly spoke to keith at it in SF10:15
Keybukhe was keen10:16
Keybukespecially since the libudev socket filtering stuff allows it to only receive input subsystem events10:16
seb128lool: btw where have you seen ongoing soyuz ddeb work?10:19
pittiKeybuk: attaching the file names in /etc/udev/rules.d/ might also be interesting?10:25
Keybukpitti: possibly, though in practice everything in that directory is user-generated10:26
pittiKeybuk: exactly10:27
Keybukin general, it's better to get udevadm test results as they're more revealing10:27
pittiwe might check for weird local customizations10:27
Keybukbut sure10:27
pittiKeybuk: that needs a devpath, though?10:27
tjaaltonpitti: it's 660 lines, so not that much ;)10:28
lesshastehi all10:28
pittiKeybuk: is there some general command we could put into the hook for that?10:28
cjwatsonseb128,lool: https://bugs.launchpad.net/soyuz/+bug/28520510:30
ubottuUbuntu bug 285205 in soyuz "Soyuz needs to be able to process and publish ddebs" [Medium,Fix committed]10:31
pitti!10:31
pitti*bounce*10:31
cjwatsonI was about to ask whether anyone had been coordinating with pitti - guess not ;-)10:31
pitticjwatson: well, we discussed how the system needs to look like about a year ago10:31
cjwatsonI meant for deployment10:32
pittinot so far10:32
seb128excellent10:32
seb128I didn't know they were actively working on it now10:32
Keybukpitti: only if you have a devpath that's broken10:34
lesshasteis it well known that if you install adobe-flashplugin with gnash already installed mozilla just silently ignores it?10:34
lesshastes/mozilla/firefox10:35
lesshasteit's quite an annoying feature10:35
seb128firefox let you choose which one to use10:35
lesshasteseb128:  no it completely ignores it.. i.e. it doesn't show up in about:plugins10:36
seb128what ubuntu version do you use?10:36
lesshastehardy10:36
seb128ok, that's a new feature, it was not in hardy10:37
seb128it's there in jaunty10:37
lesshasteah I see10:37
lesshastemaybe this counts as a bug?10:37
lesshasteit certainly confused me for about an hour10:37
seb128it does count as a bug but that's not likely to change in hardy10:37
lesshastecough.. lts.. cough :)10:38
pittiKeybuk: check out https://bugs.staging.launchpad.net/ubuntu/+source/udev/+bug/37624810:38
ubottuUbuntu bug 376248 in teeworlds "Ubuntu 9.04 has an obsolete version of teeworlds" [Undecided,New]10:38
lesshasteshall I report it?10:38
pittiKeybuk: (ubuntu-bug udev)10:38
seb128lesshaste: no10:39
seb128lesshaste: lts means security support, not that every new feature will be added to the lts10:40
pittiKeybuk: CustomUdevRuleFiles will only appear if there are actually files besides 70-persistent-* and README10:40
seb128lesshaste: the firefox code to let you choose what plugin to use is probably non trivial and will not be considered as a stable update candidate10:40
seb128lesshaste: and it's easy to workaround, uninstall the one you don't want to get used10:41
Keybukpitti: perfect10:41
Keybukpitti: odd errors in CurrentDmesg.txt10:41
pittiKeybuk: whoopsie, good catch10:42
pittiKeybuk: ugh, I have some serious trouble building an udev source package10:49
pitticomplains all the way about test/sys/10:49
tjaaltonpitti: you wanted to know if the hotkeys ceased working? The battery/suspend keys don't work on my thinkpad X61, although it could be that what should be listening to those is busted10:49
pittiKeybuk: I pushed my apport hook, but i got to leave for ~ 1 hour; I'll wrestle with it later, unless you beat me to it10:51
pittitjaalton: ok, later, need to leave, sorry10:51
tjaaltonpitti: heh, ok10:51
Keybukpitti: -i'(^|/)(\.bzr|\.gitignore|test)($|/)' :p10:52
jazzdoghi10:52
jazzdogi have fixed a lot of warnings in the sniffit package. this should fix this bug: https://bugs.launchpad.net/ubuntu/+source/sniffit/+bug/10718010:53
ubottuUbuntu bug 107180 in sniffit "Segmentation Fault" [Undecided,Invalid]10:53
jazzdoghow can i help the community with this? :)10:53
Keybukjazzdog: https://wiki.ubuntu.com/DeveloperGuide/Sponsorship10:54
jazzdogKeybuk: thanks. could you please explain the 1st and 4th point under "attach your work" and the difference between them? sorry I'm new to ubuntu, coming from slackware11:03
Keybukjazzdog: if you're looking to get involved, try #ubuntu-motu11:05
lifelessTheMuso: what are the parameters to run dmraid-activate with ? I happen to have needed to reboot...11:09
TheMusolifeless: Probably the best way to get a log is to modify the file, adding set -x and dumping to a log file in /dev/.initramfs, regenerating the initramfs and rebootingv.11:12
lifelessTheMuso: yeah, I will11:12
lifelessTheMuso: the reboot was hardware :P11:12
lifelessTheMuso: but as I had it at the prompt I thought I'd ask11:13
TheMusolifeless: ah ok.11:13
TheMusobest bet to see whats going on at a glance is to run sh -x /sbin/dmraid-activate sda or dmraid-activate sdb11:13
TheMusosorry dmraid-activate /dev/sda/b11:14
TheMusoso either drive11:14
lifelessyah11:14
lifelessI tried that11:14
lifelessdmraid-activate /dev/sda11:14
lifelessdmraid-activate /dev/sdb11:14
lifelessno output, no activation11:14
TheMusoand nothing?11:14
TheMusocdid you run with sh -x?11:15
TheMusos/cdid/did/11:15
lifelessno didn't think to run it that way11:15
TheMusoright11:15
Keybukcjwatson: have you ever seen make utterly fail to traverse deps?11:24
cjwatsonKeybuk: not in a way I ever proved to be make's fault :-)11:26
cjwatsonusually turned out to be me misunderstanding some abstruse feature of GNU make11:27
KeybukI think it's bzr's fault11:27
Keybukshelve has done weird things to my timestamps11:27
loolOk so my strace/valgrind issues were: setuid/setgid executable can't be run by valgrind; strace was crashing in jaunty, I filed bug #376858, but the karmic version fixes this11:39
ubottuLaunchpad bug 376858 in strace "strace: malloc(): memory corruption (fast): 0x00000000026452c0" [Undecided,Fix released] https://launchpad.net/bugs/37685811:39
loolOtherwise it's perfectly fine to run python or strace under valgrind11:41
cjwatsonlool: I was going to suggest set-id problems earlier, but you gave ls as your example so I decided not to bother ...11:44
cjwatsonI don't suppose that a similar set-id-handling trick to the one mentioned in strace(1) works with valgrind?11:45
loolcjwatson: having an alternate suid valgrind?  (there's also the -u flag, but didn't see the corresponding valgrind flag)    I guess I can try out11:59
cjwatsonthat's what I meant, yes11:59
loolcjwatson: It didn't work to run valgrind suid root12:13
loolAlbeit interestingly I didn't get the warning12:14
loolthat said, I'm not running the suid root program directly, but via pyhton12:16
Riddellpitti: we have a couple of non-trivial SRU requests12:33
Riddellpitti: firstly plasma-widget-network-manager is broken in many ways in jaunty which is quite a big problem for people12:34
Riddellwe have a new version which fixes a lot of scenarios but it's a new snapshot and not a simple patch12:34
directhexwho should i poke if a link on summit.ubuntu.com is 404ing?12:52
Hobbseedirecthex: Keybuk, usually12:55
pittiKeybuk: ah, thanks for the magic :)12:55
directhexKeybuk, the link for "Xorg: Driver Selection for nvidia hardware" points to +spec/driver-selection-for-nvidia-hardware instead of the correct +spec/desktop-karmic-xorg-driver-selection-for-nvidia-hardware/12:56
pittiRiddell: my primary requirement is "does it cause regressions"12:57
pittiRiddell: so if not a lot is working with the current version, and it's a major regression wrt. intrepid, it sounds feasible12:57
pittihow intrusive are the chnages?12:57
pittijust 100 individual bug fixes or a complete rewrite?12:58
Riddellpitti: 100 bug fixes12:58
Riddellbeta quality vs alpha12:58
=== azeem_ is now known as azeem
Riddellwe've been testing it quite a bit and havn't heard of any regressions12:59
pittiRiddell: sounds like we should let it mature in -proposed for a while then12:59
Riddellyes12:59
Riddellpitti: the other SRU is KDE 4.2.313:00
Riddellwhich has also been tested in PPAs and seems  to have no problems13:00
directhexthe whole thing?13:02
pittiRiddell: that sounds quite a bit more ambitious, and much harder to test for regressions..13:02
Riddellnothing we have done before for intrepid13:02
Riddelldirecthex: sure why not, if  upstream make bugfix releases it seems impolite not to use them13:03
pittiyeah, unfortunately we were pretty lax with intrepid SRUs because that was so broken in the first place13:03
pittibut jaunty itself should be much better?13:03
directhexRiddell, it's never been 100% clear to me the attitude towards bugfix releases versus bug fix patches13:04
pittithey just bind a lot of work and cause instability, so we shouldn't do them too liberally13:04
pittican we at least limit it to individual components were people filed actual bugs against?13:06
Riddellok, maybe ScottK will want to argue the case more or knows of paticular packages which have important fixes13:06
* ScottK looks up.13:06
ScottKpitti: I would argue that our experimentation in Intrepid showed that we we can reasonably reliably deliver KDE point releases without regression.  Equally it shows that users benifit from the update (I certainly did).  I think that if we can repeat the process for Jaunty it would be of benifit to the users.13:10
ScottKpitti: I also think that we should minimize proliferation of of unofficial repositories and deliver this via the Ubuntu archive.13:10
pittiScottK: that's a knockout argument13:10
pittibecause it applies to all packages13:11
pittiand kind of defeats the purpose of a stable release13:11
pittiwe had that major exception in intrepid because it was particularly bad when released, but jaunty wasn't13:11
ScottKI can also see this being a case where you might want the tech board to decide.13:11
pittiand if we keep concentrating on fixing stables instead of fixing development releases, we invest in the wrong direction13:12
ScottKI do think the benifit from from 4.1.4 was less than 4.1.3.13:12
pittiwe stretched the SRU policy a lot in intrepid, but we can't maintain this forever13:12
ScottKI think it might be reasonable just to deliver the first update and not the 2nd.13:12
pittiScottK: of course you are welcome to challenge this on the TB, of course I'll comply to a TB decisisuionm13:13
ScottKpitti: I'd rather not have it be a challenge against what you say, I was hoping more we could come up with something you could support.13:13
ScottKTB specifically is the authority to grant microversion update exceptions.13:14
directhexcompromise: ubuntu-backports ?13:14
loolIs there no special provision for point updates to KDE sources like there is for GNOME packages?13:14
Riddellwe want backports for new KDE versions13:14
ScottKSo if Intrepid was an experiment, then we get official approval before doing it in Jaunty (or not).13:14
ScottKdirecthex: That would be my plan B, but as Riddell says has other complications.13:15
Riddelllool: what special privision does gnome have?13:15
ScottKlool: There is not (I think we should get one).13:15
pittiScottK: it's not "against me", it's "against SRU policy"13:15
ScottKpitti: OK.13:15
pittiit's not like I'd arbitrarily decide that :)13:15
pittilool: we don't update GNOME packages either13:16
loolI thought GNOME had special exception to not have a too heavy SRU process just after release13:16
pittilool: that only applied to hardy13:16
pittia decision by TB13:16
loolHmm ok, for some reason I thought it was permanent, nevermind then13:16
pittiand we really, really, REALLY need to stop throwing so much time at stables13:16
ScottKRiddell: How about jaunty-backports for now and then ask TB to add KDE to https://wiki.ubuntu.com/StableReleaseUpdates/MicroReleaseExceptions13:16
pittiwe do not have enough development power for that, and users don't appreciate installing 100 MB of updates every day either13:17
loolThe thing is that we don't get enough QA and we push a lot of changes until the last minute, I'm in a bad position to criticize this as I had to do so myself to reach some goals in previous releases13:17
ScottKpitti: For the KDE updates it isn't a manpower issue.  We will package them.  The only question is where do we deliver them.13:17
RiddellScottK: can do although if they don't approve Gnome I doubt they'll approve KDE13:18
ScottKDid Gnome ever ask?13:18
loolWhen I say QA above, I mean community testing; I think QA works well on some key stacks, but not on the less used packages  :-/13:18
ScottKRiddell: I wouldn't have dared ask for KDE in the KDE3 timeframe as they did not do bugfix only releases.  I think in 4.1/4.2 they have shown discipline about it.13:19
pittiScottK: you also need to test them heavily, and the bigger the updates get, the harder it is to ensure that it doesn't have major regressions13:19
pittiand there's also the user inconvenience of downloading tons of updates13:19
ScottKpitti: This is true, but I think we've shown we can do the testing.13:19
ScottKWe also have a lot of user demand for the updates.13:19
seb128lool: the issue is that stable is what most users are using13:19
loolseb128: Yeah13:20
pittiScottK: right, but it still binds manpower13:20
ScottK4.2 is a lot better than 4.1, but it's not nearly as stable as KDE3 yet.13:20
seb128I had the discussion this week with pitti about GNOME stable updates13:20
pittisome people demand updates all the time, but *shrug813:20
loolseb128: But we have a really large window where people are encouraged to upgrade to shake out bugs13:20
ScottKpitti: Yes, but we are going to do the work anyway.13:20
pittiScottK: see, and that's what I'd like to see stopped :)13:20
ScottKIn this case we've already done it, so it's a sunk cost in any case.13:21
loolseb128: Usually, from beta to final I am always too busy to take new stuff on my plate, such as new bugs discovered by people upgrading13:21
seb128lool: I consider jaunty as a good version and worth stabilizing it a bit, especially where karmic shape to be a bumpy one13:21
seb128lool: I expect lot of people will want to stay on jaunty over karmic13:21
ScottKOh dear.13:21
loolseb128: Makes sense13:21
ScottKseb128: Hopefully not Intel video users.13:21
ScottKI hope that gets better.13:21
seb128I'm not interested in trolls13:22
ScottKseb128: I'm not trollilng.  The Intel video experience on Jaunty is sub par and saying so is just stating facts.13:22
Hobbseeis karmic any better yet, though?13:22
pittiScottK: agreed13:22
pittiHobbsee: it will be differently broken13:23
seb128it's not an issue problem and not something ubuntu has total control on13:23
Hobbseepitti: well, yeah, I guess that's normal :)13:23
seb128issue -> easy13:23
pittiHobbsee: intel, I mean, with all the new UXA/KMS stuff13:23
ScottKseb128: I wasn't laying blame.  I was hoping next time it would be better.13:23
seb128hey Hobbsee, long time that I didn't talk to you, how are you?13:23
Hobbseepitti: that's true.13:23
pittiyeah, intel wise jaunty was quite a dent :(13:23
* directhex thinks calm & beer should be liberally issued13:24
seb128I expect that karmic will not be a good luser version13:24
tjaaltonScottK: it can't get any worse ;)13:24
Hobbseeseb128: heya!  I'm doing OK.  Not really active, focussing mainly on uni work, to get it finished.  Done at the end of the year, into the big wide world - woot!13:24
Hobbseetjaalton: sure it can.  Hardy I could'nt play video most of the time ;)13:24
Hobbseetjaalton: intrepid plays video.13:24
Hobbseeer, wait.  s/hardy/intrepid; s/intrepid/jaunty/13:25
tjaaltonHobbsee: lucky you :)13:25
Hobbseetjaalton: oh, indeed.  I just didn't watch any movies on this laptop for 6 months or so ;)13:25
jazzdogcomeon, everything plays video if one builds mplayer himself :)13:26
Hobbseejazzdog: actually, the only thing that would work for me was xine-ui ;)13:27
Hobbseewhich was...interesting13:27
ScottKpitti and Riddell: Did we reach some resolution on a path forward?13:27
pittiScottK: I'd really like you to reconsider and focus efforts on karmic, but as I said, if you think you want an exception, you should feel free to ask the TB13:28
ScottKpitti: OK.  Thanks.13:29
RiddellI'm pretty sure the kubuntu team will always want to make backports, they're very popular (at least we get a lot of complains from being late with 4.3 beta)13:29
ScottKRiddell: ?  Backports and then think about it (backports won't conflict until 4.3.0 so we have some time)13:29
RiddellScottK: yeah go for  it13:29
* Riddell out for a bit13:30
HobbseeRiddell: not to mention they've routinely made backports on kubuntu.org, too13:30
Hobbseeor at least, back when I ran it13:30
directhexHobbsee, exciting career plans yet?13:32
Keybukdirecthex: poke rickspencer or pitti ;)13:32
Hobbseedirecthex: none yet.  Hoping to get bites when I've finished the degree, or close to it.13:32
* directhex redistributes buck13:32
ScottKHobbsee: That's a good point (re kubuntu.org).13:33
ogracjwatson, poke ... http://ports.ubuntu.com/ubuntu-ports/pool/universe/l/linux/ seems all linux packages for armel landed in universe again14:06
cjwatsonogra: OK, I'll have a look, thanks14:11
ograthanks for looking14:11
cjwatsonogra: should be fixed for the next relevant publisher run (visible a bit over 1.5 hours from now)14:17
ogracjwatson, thanks a lot14:17
sorenI see this is accepted for UDS, but I can't see it on the schedule anywhere? https://blueprints.edge.launchpad.net/ubuntu/+spec/foundations-karmic-swapfile14:29
cjwatsonsoren: we took it off as it's already been discussed twice and just needs time to implement14:33
sorencjwatson: Alright. Thanks.14:34
=== asac_ is now known as asac
pittiwhat the heck went wrong with the PPAs?15:11
pittiamd64      6010 builds waiting in queue15:12
pittii386      11538 builds waiting in queue15:12
pittiif that isn't a DoS, what is?15:12
directhex11539 builds is.15:12
directhex11538 is merely popularity15:12
directhexcjwatson/test-rebuild-2009051315:13
directhexthat answer your question?15:13
pittiah, I see15:13
pitti*phew*15:13
pittiI thought something was going crazy and reuploaded infinitely or so15:14
directhexi can do that if you want15:14
sistpoty|workoh, pitti: did your copying from jaunty-updates forward to karmic somehow confuse MoM? faumachine is now listed in universe-manual despite an identical tarball?15:16
sistpoty|work(not that it matters in regards to faumachine, but maybe other packages are affected as well)15:17
pittisistpoty|work: hm, entirely possible, I don't know15:28
pittibut we have done that for a long time already15:28
* apw looks for a main-sponsor to upload a grub change to karmic/jaunty, on bug #37687915:40
ubottuLaunchpad bug 376879 in grub2 "grub2 installer modifies grub 0.97 menu.lst incorrectly and fails to chainload grub2" [Medium,In progress] https://launchpad.net/bugs/37687915:40
* seb128 points apw to ubuntu-main-sponsors on launchpad to subscribe to the bug15:41
apwseb128, t'is done15:43
seb128apw: ok something will look at it while doing sponsoring I guess15:45
pittioh, argh15:45
pittilool: you uploaded sqlite-3 to -updates instead of -proposed, and I failed to spot that15:45
loolpitti: Crap; sorry15:46
loolI really am not in a good shape today15:46
pittiI removed the package again15:46
loolpitti: Thanks15:46
loolpitti: Do you have the .Dsc?  I had removed it since it was accepted15:47
loolotherwise I'll rebuild it15:47
pittilool: please reupload15:47
pittilool: yes, you can grab it from launchpad15:47
loolIn "Done"?15:47
pittihttps://edge.launchpad.net/ubuntu/+source/sqlite3/3.6.10-1ubuntu0.115:48
loolThanks15:48
pittiI set the build score to -1 and will reject any binaries that still make it15:48
loolpitti: I'm upload a .2 built with -v to include both versions15:49
pittilool: just rename it .215:50
pittilool: but your's is fine as well15:50
pittias you like15:50
loolpitti: uploaded and15:50
loolsorry not accepted yet15:51
loolI need a break, bbl15:51
pittiI guess they'll just fail to upload, but I'll watch it anyway15:52
loolpitti:  subject: [ubuntu/jaunty-proposed] sqlite3 3.6.10-1ubuntu0.2 (Waiting for16:01
pittilool: accepted16:02
cjwatsonapw: your kexec-tools patches all look fine, thanks (and sorry for the delay); the only change I'm making is that the upload target in debian/changelog for jaunty needs to be "jaunty-proposed"16:05
apwcjwatson, doh thanks16:05
cjwatsontrivial to fix at sponsorship time though :)16:06
apwi get caught by that automatic installation of the current distro a lot16:06
apwi need to change ti to YOUARESTUPID or something by default16:06
apw.. it is all a learning expierence.  one day i'll get one right first time16:07
cjwatsonapw: by dch -i?16:07
cjwatsonor dch -r?16:07
apwyeah it puts in your local distro release by default (dch -i)16:07
cjwatsonif you put DEBCHANGE_RELEASE_HEURISTIC=changelog in ~/.devscripts then dch -i will put UNRELEASED in there instead; that said, dch -r will still use the current release (but you could set it by hand as an alternative)16:08
apwyeah that would be fine, as it goes nice and RED in there to tell me i've not set it if its UNRELEASED16:08
apwcjwatson, perfect thanks16:09
loolDo people like the dch -r defaulting to Ubuntu dev under Ubuntu?16:10
loolI keep mixing the Debian and Ubuntu behaviors because I work on Ubuntu packages from Debian and vice-versa16:11
loolUsually Debian's dch -r does what I want; Ubuntu's is wrong half of the time16:11
seb128I almost never use dch -r, I usually use -i or -v new_version16:12
* ScottK would like a dch -rd or something like that to get unstable16:12
cjwatsonI use dch -r all the time and like its behaviour16:12
cjwatsonwhen I'm working on Debian packages I do so in a Debian chroot, since I need to do binary uploads anyway16:13
ScottKI find dch -r works well for me for Ubuntu16:13
cjwatsonStevenK: can you or one of your team merge libhildon? I'd have no idea where to start16:15
pittilool, ScottK: -d exists, it's called -U16:35
Keybukhmm16:36
ScottKpitti: Thanks.16:36
Keybukkernel panic on boot, at exactly the same time, thunder rolls outside16:36
geserwhere is your Igor?16:36
dokoplease could somebody process binutils in NEW (binutils-gold -> universe)16:38
AwsoonnThank you seb12816:42
loolpitti: That's just for the version number, right?  jaunty is set unconditionally16:44
lool(well unless -Dfoo is used)16:44
pittiah, right16:45
pitti-Dunstable would be the target16:46
seb128Awsoonn: you're welcome, thanks for the work on the bug16:46
=== dpm is now known as dpm-afk
=== ArneGoet1e is now known as ArneGoetje
cjwatsonasac: I've synced wpasupplicant into karmic, since Kel incorporated my patches; you might care18:29
* asac *nods*18:55
=== rickspencer3 is now known as rickspencer3-afk
ebroderAnyone from motu-sru who could provide an opinion on bug #371581?20:37
ubottuLaunchpad bug 371581 in erlang "erlang-base conflicts with old erlang-doc-html" [Undecided,New] https://launchpad.net/bugs/37158120:37
ion_Anyone feel like uploading command-not-found with the patch from LP #377065, or should i just wait for mvo? :-)21:56
ubottuLaunchpad bug 377065 in command-not-found "zsh_command_not_found should use the pre*_functions arrays instead of overwriting the singular preexec/precmd functions" [Undecided,New] https://launchpad.net/bugs/37706521:56
directhexhow would i define which files should be uploaded by apport in the event of an app dying?22:07
ion_directhex: /usr/share/apport/package-hooks22:09
directhexi should look into what would be needed for apport to get triggered by Managed exceptions in mono apps22:12
ajmitchdirecthex: I remember looking at that awhile ago, back then we discussed havign a small patch to the mono runtime to dump out the right info22:18
directhexajmitch, i'll ask miguel what he thinks22:21
=== jldugger is now known as pwnguin

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