IRC log of #maemo-meeting for Monday, 2012-05-14

*** X-Fade has quit IRC02:52
*** X-Fade has joined #maemo-meeting03:09
*** Pali has joined #maemo-meeting08:28
*** javispedro has joined #maemo-meeting17:38
X-FadeHi17:59
X-FadeAnyone awake? :)18:00
javispedroI am18:01
javispedroanyone else? =)18:02
X-Fadejavispedro: hi :)18:02
X-Fadetimoph ?18:02
X-FadeOk, well ;)18:04
X-FadeI'll give a quick update then.18:04
javispedroheh18:04
javispedroI saw that you made x86 builds, something was wrong with armel ones18:04
X-FadeYeah, so there was an update to OBS with broke arm builds.18:05
X-FadeI figured out why they weren't being scheduled and got that fixed. But now also noticed the way they changed the way qemu is being used in the vm setup.18:06
X-FadeSo it is still broken and I will try to fix that this week.18:06
X-FadeI spoke with lbt about the SB2 approach just before he went to the Tizen conference.18:06
javispedroaha18:07
javispedroafter reading about it I'm afraid it's not the panacea I expected though18:07
X-FadeWe'd need to create some minimal rootstrap and use that.18:07
X-FadeIt could fix or prevent some sb-isms, but we don't seem to run into many atm.18:08
javispedrowhy minimal? why not use the sdk one?18:08
X-FadeI need to follow up on that with lbt now he is back :)18:08
javispedrooh, i was thinking maybe it got pushed around and a 1GiB transfer is too large18:09
X-FadeWe obviously need some newer tools like newer debhelper as many packages in extras depend on it now.18:09
javispedrogood point18:09
X-FadeThat was caused by introducing a newer debhelper to extras-devel for instance.18:10
javispedroso you are saying that we are now getting packages that "implicitly" depend on newer debhelper features too?18:10
X-FadeBut if you look at the build failures in i586 atm, most of them are explainable..18:10
X-Fadewell, they depend on debhelper >=718:11
javispedroah, wait, you are talking about squeeze debhelper18:11
javispedroI thought someone had already pushed something more modern =)18:12
X-FadeWell that is quite modern for fremantle already :)18:12
javispedroin which case, yes, I agree, rootstrap should be updated18:12
X-FadeSo my plan was to talk with lbt this week about how to proceed with an SB2 test.18:12
X-FadeWe might come up with a minimal rootstrap and a repo with updated 'core' tools.18:13
X-FadeBut I need to get a list of pros and cons together.18:13
X-FadeAre you interested in joining the conversation with lbt about SB2?18:15
javispedrowhen will that happen?18:15
X-FadeWell, if yes, then I'll take your availability into account too :)18:16
javispedrook, let's try, any day around this time of day is fine18:17
X-FadeOk, will contact lbt and keep you posted.18:18
javispedrotrying to think of pros are, less scratchboxism fuss; potentially faster VM startup times of we have a generous rootstrap?18:18
javispedro(and thus not have to call dpkg -i as many times during vm startup)18:19
X-FadeYeah, but also acceleration.18:19
timopho/18:19
javispedroah, yeah18:19
X-FadeWhich makes a lot of sense when you do rebuilds in a shared environment.18:20
* timoph reading backlog18:20
javispedrohi timoph18:20
X-Fadetimoph: No worries :)18:20
* merlin1991 reports in18:21
* merlin1991 reads backlog18:21
X-FadeAnyway, that is it for me :)18:21
X-FadeAnything else we should discuss?18:22
timophhmmh18:23
timophnot much to do before the obs issues are resolved18:24
X-Fadetimoph: Yep, unfortunately. But now people are back, so I expect some progress this week :)18:24
timophyeah18:25
X-FadeSo I'd sat we wrap it up for now and meet again next week?18:25
timophyep18:26
javispedroso the conclusion is that X-Fade tries to figure out the armel issue,18:26
javispedroand that we shall set some time with lba to ask him a few sb2 questions18:26
X-FadeCorrect.18:27
javispedrosome of the x86 build failures are already "interesting"18:27
javispedrohttps://build.obs.maemo.org/package/live_build_log?arch=i586&package=bootmenu&project=Extras%3ATesting13&repository=Maemo_Fremantle_1.3_standard18:27
X-Fadejavispedro: Yes, and it is those that we need to look at.18:28
javispedroyep18:28
javispedrothe build failures page is currently way too large and takes a while to load; hopefully when the armel thing is fixed it will be much more compact18:28
X-FadeBut i386 is not as well tested as armel, so I would rather do that when we have armel up and running18:28
merlin1991dang can't get the build log in question to show18:28
X-Fadejavispedro: most of them in i586 are Architecture: armel18:29
javispedro??18:29
X-FadeAnd thus the build is marked failed after it completes :)18:29
javispedrooh18:29
javispedroisn't there a way to avoid all this wasted CPU time?18:30
X-FadeInstead of : Any18:30
X-FadeMight be, but now it is not a big deal.18:30
javispedroand this reminds me that I had some gripes with arch-conditional build-depends parsing..18:31
javispedroin any case, if arch: armel stuff is being built, then it makes no sense for us to check at those now18:31
X-FadeUnfortunally I have to go now, other meeting :(18:32
javispedrook, ta X-Fade18:32
X-FadeBut I'll keep you posted on the list :)18:32
javispedro*if arch:armel only stuff is being built ;P18:32
X-FadeSame time, next week :)18:32
javispedro*on x86 target ;P18:32
timopho/18:32
javispedrocya18:32
X-FadeOk, thanks for joining. Until next week!18:33
*** javispedro has quit IRC18:38

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!