/srv/irclogs.ubuntu.com/2015/08/21/#ubuntu-devel.txt

=== salem_ is now known as _salem
=== _salem is now known as salem_
=== salem_ is now known as _salem
tjaaltonjderose: yo, still there?03:05
pittiflexiondotorg: it's okay, NEW delays aren't taken into account for FF07:37
Riddellcjwatson: you say I can update packagekit up to some version?09:44
cjwatsonRiddell: 1.0 will break click until mvo gets round to landing that branch09:55
cjwatsonI know not where that stands09:55
Riddellmvo: hola, where do you stand? I can update to 0.9 I think should be enough for what I need (my gsoc project to install bits) but obviously latest would be nicer09:56
mvoRiddell: at debconf right now, I can try to get to it next week but time is very limited currently due to other duties, the branch should basicly be ready, its missing tests and some more manual testing09:57
Riddellmvo: groovy, I'll see if I can update to 0.9 to make feature freeze then we can test your stuff next week09:57
LaneyRiddell: to make feature freeze?09:58
Laneyit was yesterday09:58
Riddellah, hmm09:59
Riddelloops :)09:59
LaneyHappy to see you guys appstreaming though as I'm kind of interested in that stuff for us too09:59
Laneyximion showed me appstream.kubuntu.co.uk09:59
RiddellLaney: yeah I need to chat to him about what the next steps are for that10:00
LaneyPPAs seem a bit hard10:00
LaneyI guess you care about that too10:00
Riddellmm, I've not through about that10:01
Laneythink it will need some kind of LP integration10:02
Laneyeven if that's a dispatch/collect job or something10:02
* Laney doesn't really know10:02
rbasaksil2100: I'll sponsor your patch in bug 1471903 if there isn't any further objection. ogra_, do you have any other comment in response to bdmurray's response?10:08
ubottubug 1471903 in live-build (Ubuntu) "-updates, -security missing from apt lists" [High,In progress] https://launchpad.net/bugs/147190310:08
ogra_rbasak, you mean beyond asking to wait for signoff from the product team (pat mcgowan is back on monday) ?10:08
sil2100rbasak: hey! Thanks, no, I didn't hear anything new... also, infinity seemed to be in favour of the same change as well10:09
rbasakogra_: signoff for what reason?10:09
ogra_rbasak, note that mako already has probs installing from scratch today due to the smaller cache size10:09
ogra_rbasak, the product team usually makes all size related decisions for these images10:10
sil2100ogra_: this can basically land to wily now anyway, the PT will decide if we want the change in the stable images or not10:10
sil2100Since this is not only about the phone images but all images everywhere10:11
ogra_sil2100, well, you deal with the support in #ubuntu-touch for mako users then :P10:11
sil2100uh oh, we recommend using rc images for mako as well!10:11
rbasakogra_: so you're saying that I can't upload any bugfix that will increase size without product team approval? I realise this is a bit special, but that doesn't sound reasonable in the general case for me.10:11
ogra_i'm not really fond of bypassing the product team with that one, but up to you10:12
ogra_rbasak, for the phone images they are our consumer and need to make sure what we pdoruce fits on the vendors devices10:12
ogra_*produce10:12
ogra_if we add bits to the seed of the phones we usually coordinate with them10:13
rbasakScottK: around? I'm reviewing kickinz1's merge of amavisd-new and am a bit confused about debian/22-amavisd-new-postfix. It doesn't seem to be used or end up in a binary package in the current version in Wily. kickinz1 has moved it to debian/etc/conf.d/ where I think it will get used. Does this sound right to you?11:07
rbasakI wonder if Debian restructured debian/ and this part of the delta got left behind or something. But clearly if we do this the binary package delta will change from the user's perspective. Maybe this is an accidentally disappeared confflie from the past?11:08
rbasakOr should it ship with amavisd-new-postfix only?11:08
rbasakivoks: ^^ looks like you touched this back in the day. Any ideas please?11:09
ivokschecking11:09
ivokswithout looking at the source11:11
ivoks22-amavisd-new-postfix should be amavisd-new-postfix's amavis configuration11:11
ivoksit should ship only with amavisd-new-postfix package11:11
rbasakOK, thanks11:12
ivoksbut, i haven't touched this in 5 years :)11:12
rbasakWhere should amavisd-new-postfix drop that file when installed?11:12
ivokswhere other amavis' configuration is. iirc that's /etc/amavis/conf.d11:12
rbasakOK that makes sense.11:13
ivokslooking at source11:13
ivokshm11:14
ivoksthat file was never utilized11:14
rbasakI wonder how that got lost though. Seems unlikely that amavisd-new-postfix.install would get lost in a merge.11:14
ivoksat least not in 14.1011:14
ivokser 15.0411:14
rbasakYeah - hence my confusion11:14
ivoksthat file enables antispam and antivirus functionality :D11:14
ivokswithout it amavisd-new-postfix is not very useful11:15
ivoksbut11:15
rbasakNor is it in Precise.11:15
ivokshm11:15
ivoksmaybe default changed11:15
ivoksgive me a second11:15
rbasakThanks11:15
ivoksit didn't11:16
ivoksyeah, i built that in 201011:17
ivoksbut has been merged few times since then11:17
ivoksrbasak: bottom line, it needs to be shipped with amavisd-new-postfix, in /etc/amavisd/conf.d/11:20
ivoksrbasak: but you might want to double check with ScottK11:20
rbasakOK, thanks.11:21
ivoksdidn't help you there :)11:21
rbasakkickinz1: ^^11:21
=== marcusto_ is now known as marcustomlinson
rbasakWell, it's nice to know that someone knows what's going on here.11:21
rbasakIMHO, we should retire the deltas relating to mail-delivery-stack etc.11:21
rbasakBut that's a discussion for another day.11:21
ivoksidea behind m-d-s was to have single packet that will set 'best practices' for mail server11:22
ivoksat first, amavisd-new-postfix was not included in that because antispam and antivirus functions impact mail delivery11:22
rbasakYeah. Trouble is that it tries to shoehorn both service orchestration and configuration management into packaging, which doesn't really work well IMHO. Made sense at the time, but we have better tools nowadays.11:23
ivoksand when you go into realm of not delivering an email for some reason, then you need to be sure you have no false positives (which is impossible)11:23
ivoksyeah, i get that11:23
rbasakI'd like to see all this stuff implemented into charms and then pulled out of packaging.11:23
rbasak(in all that free time I have)11:24
ivoksi did start that effort long time ago11:24
ivoksnever finished it :/11:24
ScottKrbasak: it's been even longer for me the ivoks since I touched it.  The mail stack stuff is unmaintained for a very long time. I agree about dropping it.11:35
ivoksthere you go11:38
rbasakScottK, ivoks: OK, thanks. For this cycle, I think maybe we can merge without changing anything from what is there already - so let's not ship 22-amavisd-new-postfix at all just like we weren't doing before since that won't regress anyone.11:41
rbasakWe can talk about dropping this stuff next cycle.11:41
ivoksyeah, it should be announced11:41
ivoksand before we drop it, at least have a replacement strategy11:42
rbasakkickinz1: ^^11:42
ivoksjuju this and that to get it like you had11:42
ivoksfor that we need charms11:42
=== _salem is now known as salem_
=== Tm_Tr is now known as Guest13976
dholbachtkamppeter, do you know who can take a look at https://bugs.launchpad.net/hplip/+bug/1486732?12:51
ubottuLaunchpad bug 1486731 in HPLIP "duplicate for #1486732 hplip-3.15.7-plugin.run doens't return" [Undecided,Confirmed]12:51
dholbachI'm seeing the same thing12:51
jderosetjaalton: was off to bed, but here now :)13:17
jcastroivoks: wrt. charming the mail stack, I can put it on our team's wishlist, is your code pushed to the store?13:35
ivoksjcastro: let me take a look13:46
ivoksjcastro: ah, now i remember13:46
ivoksat that time charms were unable to fetch information that was required to do this13:46
ivokshttps://code.launchpad.net/~ivoks/charms/precise/mail-stack-delivery/trunk13:47
ivoksi think the problem was the fact that juju was unable to tell me which node departed13:47
ivoksso, i was suposed to come up with a workaround for which i had no time then13:47
jcastroivoks: I'll toss it on our list, can't promise everything, but it'd be nice to provide that for people13:48
ivoksjcastro: ok13:51
tkamppeterdholbach, the two bug reports are for Arch Linux, but I contacted the HP guys directly anyway.14:04
dholbachtkamppeter, thanks14:05
tjaaltonjderose: ah, nothing special, just that skl on wily needs work on the kernel still15:01
jderosetjaalton: ah, right.  so was i correct that this patch set you proposed is why skylake is shiny on Vivid, Trust+Vivid HWE and badly broken on Wily?15:03
tjaaltonit only fixes a known gpu hang and probably prevents some others15:04
tjaaltonyou need to give i915.preliminary_hw_support=1 to actually use the kernel module15:04
jderosetjaalton: ah, thanks, great tip! i did notice that kms doesn't seem to be used on Wily... so the horrible GPU performance is actually because the GPU isn't used at all?15:07
tjaaltonright15:07
jderoseokay, that's very helpful. means i can do reasonable testing in the mean time :D15:07
tjaaltonyep15:08
jderosetjaalton: if you don't mind my asking, what's your assessment of Ubuntu on skylake hardware overall? any show stopping issues?15:21
ejatanyone can explained why there is 153 packages not upgrade ?17:20
ejathttp://paste.ubuntu.com/12142510/17:20
cjwatsondist-upgrade not upgrade17:21
cjwatsonupgrade on its own is only very slightly useful, especially when we're in the middle of a C++ transition requiring a bunch of library packages to be renamed17:22
ejatcjwatson: if i do dist-upgrade .. there is 97 to remove and 5 not upgraded.17:22
cjwatsonyes, you have to inspect the list to be removed17:22
cjwatsonfeel free to pastebin that17:22
cjwatson(well, the full output)17:23
ejathttp://paste.ubuntu.com/12142547/17:23
ejatscary after doing the inspection17:23
cjwatsonso a lot of those are correct but there's evidently something up with libreoffice17:24
cjwatsonI mean, the libraries, those are basically just swapping existing ones out for v5 variants17:24
ejatowh ..17:25
ejatso ok for me to proceed removing all that 97 ?17:25
cjwatsonno17:25
cjwatson"there's evidently something up with libreoffice"17:25
ejatowh .. okie .. sorry bout it ..17:25
ejatwhat should i do know ?17:26
ejatfile a bugs ?17:26
cjwatsonpastebin 'apt-get -oDebug::pkgProblemResolver=true dist-upgrade'17:26
cjwatsonalso make sure that you only have wily in sources.list, not wily-proposed17:27
cjwatsonpitti: language-pack-kde-{ia,wa} are uninstallable in wily - missing -base and non-kde packages by the looks of it?17:28
ejatcjwatson: http://paste.ubuntu.com/12142592/17:30
ejatno proposed17:31
cjwatsonejat: I suspect you have some old package installed that it doesn't want to remove for some reason, since our analysis tools say libreoffice is installable in wily.  What does 'apt-get -oDebug::pkgProblemResolver=true install libclucene-core1v5' say?17:32
willcookeg'night all17:33
cjwatsonInvestigating (0) libstreamanalyzer0v5 [ i386 ] < none -> 0.7.8-1ubuntu5 > ( libs )17:34
cjwatsonBroken libstreamanalyzer0v5:i386 Conflicts on libstreamanalyzer0 [ i386 ] < 0.7.8-1ubuntu2build1 > ( libs )17:34
cjwatson  Considering libstreamanalyzer0:i386 103 as a solution to libstreamanalyzer0v5:i386 017:34
cjwatson  Holding Back libstreamanalyzer0v5:i386 rather than change libstreamanalyzer0:i38617:34
cjwatsonI think that's where it starts going wrong17:34
ejatcjwatson: http://paste.ubuntu.com/12142634/17:35
cjwatsonI wonder why libstreamanalyzer0 has such a high score17:36
ejatis it only me who face it ? or got some others reporting it too ?17:37
zulmterry: ping17:38
mterryzul, hello17:40
zulmterry: im just looking at the python-pymysql MIR, the server team is subscribed to the bug report, however the debian bug report, the reason why the test is failing in the debian bug is because mysql not installed17:41
cjwatsonejat: please could you create a tarball of /var/lib/apt/, /var/cache/apt/, and /var/lib/dpkg/, and attach it to a new bug against Ubuntu apt along with the debug output from dist-upgrade above?  I've spent as much time on this as I can for a Friday evening, but it smells like an apt bug to me, it's odd that it doesn't seem to be preferring libstreamanalyzer0v5 over the non-v5 version17:41
cjwatsonwith such a tarball it should be possible to recreate the bug elsewhere17:42
mterryzul, I was also seeing the failure locally.  Is there a missing build-dep then?17:42
zulmterry: other than mysql not being installed17:43
zulmterry: there isnt afaik17:43
mterryzul, maybe I'm missing something, but "mysql not being installed" sounds like a missing Build-Depend?17:44
zulmterry: yeah mysql17:44
zulmterry: ill see what i can do17:45
ejatcan upload such big file /var/cache/apt/ to launchpad?17:48
ejatcjwatson: ?17:49
anna__I get this: Skipping dist/deb_dist/nuitka_0.5.14~pre9+ds-1_amd64.changes: Can't locate Date/Parse.pm in @INC (you may need to install the Date::Parse module)17:59
anna__Any idea, why all my pbuilders starting doing that recently? The package libtimedate-perl appears to be installed... so that's strange18:00
ejatcjwatson: bug #148756918:11
ubottubug 1487569 in apt (Ubuntu) " a lot of files to be removed when execute apt-get dist-upgrade" [Undecided,New] https://launchpad.net/bugs/148756918:11
tjaaltonjderose: not at all.. it should be pretty stable, but the display side is lacking. new bios code can break current modesetting etc..18:46
jderosetjaalton: gotcha, thanks. and thanks for the FYI about bios being able to break kms, very good to know.18:57
jderosetjaalton: when you say new bios can break KMS... do you mean newer Intel video bios?19:30
mterrybdmurray, FYI, ~ubuntu-server is looking after bandit package in main now19:57
tjaaltonjderose: right21:00
jderosetjaalton: gotcha, thanks!21:01
=== salem_ is now known as _salem
=== _salem is now known as salem_
=== salem_ is now known as _salem
slangasekmdeslaur: I'm doing some cleanup of the edk2 package, including providing OVMF_{CODE,VARS}.fd as split files which is apparently what's needed for proper per-VM EFI variable support.  However the virt-manager we have doesn't appear to have proper integration with this (though our libvirt does).  You don't happen to have any prerelease packages based on the 1.2.0 virt-manager in Debian experimental,23:19
slangasek do you?23:19
mdeslaurslangasek: I don't, no23:20
mdeslaurslangasek: I can look at this next week if you'd like23:20
slangasekmdeslaur: ok.  I may just upload this untested then, since it's backwards-compatible with what we're currently doing and somebody else can figure out the new stuff23:20
slangasekmdeslaur: that would be keen but it's by no means high priority23:20

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