=== doko_ is now known as doko [11:20] infinity: cjwatson: for the uninstallable packages listed in the report.html in precise cdimage, reported bug #1206860 to account the smoke failure [11:20] Launchpad bug 1206860 in xf86-input-wacom-lts-raring (Ubuntu Precise) "xserver-xorg-input-all-lts-raring are unintstallable with precise alternate images" [Undecided,New] https://launchpad.net/bugs/1206860 [11:22] psivaa: Already noted at http://people.canonical.com/~ubuntu-archive/testing/precise-proposed_probs.html [11:22] Also e-mailed to us daily [11:22] Feel free to have a tracking bug if it helps you but we don't need one and it's quite plausible we'll forget to close it ... [11:22] infinity: cjwatson: ack, i am doing the both :) [11:22] I'll remember in this case, cause I'm about to fix it. [11:23] (i.e. it's actually slightly worse for us to have a tracking bug than not, for things in the report) [11:26] cjwatson: agreed, but without a tracking bug the smoke dashboard will be red with 0% pass rate with no bugs tagged [14:56] psivaa: Ideally there'd be some other way to log an explanation of what's happening [14:59] cjwatson: yes, but on the other side we have a daily image that's not installable. so a bug can only accompany that imho :) [15:00] if it's a known bug or not is a separate issue [15:26] apache2-prefork-dev 2.2.22-6ubuntu5 appears to be in saucy, built from the apache2 source. How do binary packages get removed when they're no longer relevant? [15:26] (apache2-dev 2.4.6-2ubuntu1 now replaces/provides apache2-prefork-dev and it's in saucy too) [15:27] rbasak: an Archive admin removes them by hand, following output from a report http://people.canonical.com/~ubuntu-archive/nbs.html [15:28] rbasak: if ofcourse it's time to do it / safe to do. [15:28] I see - thanks [15:29] This is bug 1206114. I'm not sure if it's causing a problem, but it certainly complicates things a bit. [15:29] Launchpad bug 1206114 in apache2 (Ubuntu) "Package apache2-prefork-dev broken in Saucy" [Undecided,New] https://launchpad.net/bugs/1206114 [15:30] AFAICT, there is a solution, the build isn't picking it. [15:30] rbasak: If all those deps reported on the nbs report are actually || deps, we can remove the package. I'll look. [15:30] rbasak: The build has no reason to pick the other option until the first is gone. [15:30] rbasak: nothing wrong with apache2 package, instead the reported should follow apache2.4 guidelines on how to port modules / extensions / webapps. [15:31] Sleep didn't work out for me anyway. :/ [15:31] :-/ [15:32] rbasak: And, the first hit there is broken. libapache2-authcookie-perl build-depends on apache2-prefork-dev, clearly still needs transitioning. [15:33] I imagine the rest are similar. [15:34] rbasak: nothing should build-dep on apache2-frefork-dev with apache2.4, as that will not work with new style packaging. [15:34] I don't understand this fully. Why is a package that build-depends on something that is now a virtual package broken? [15:34] commented on the bug report. [15:34] There's a Provides, so only ones with versioned build-deps must be transitions [15:34] ed [15:34] Thanks xnox [15:34] But when I last looked at that in NBS, there was at least one versioned build-dep still in place [15:34] rbasak: It's a real package too, until we remove it. [15:34] And Provides doesn't satisfy versioned relationships [15:34] Ah [15:35] cjwatson: oh, I didn't realise apache2-dev now provided it. [15:35] That helps a little bit. [15:35] rbasak: i was trying to say, that even if apache2-dev would be installed i daubt the package would build anyway =) [15:35] xnox: right, I got that part :) [15:35] =) ok. [15:35] I was just confused by the build-dep breaking things [15:36] So, mod-mime-xattr has a versioned build-dep. [15:36] Or more that the archive state broke the build-dep even though it looked OK to me. I didn't understand that provides couldn't provide it in this case. [15:36] The package in that bug will very likely need transitioning, certainly, so I'd be inclined to agree with that not being a valid bug in apache2 even though we should sort this out. [15:37] rbasak: Provides can't provide versioned build-deps (so mod-mime-xattr needs fixing), and while provides sure can provide apache2-prefork-dev, the real package (that's still in the archive) will be preferred. [15:37] No properly transitioned package is affected by this. [15:37] But odds are that everything on the NBS list needs transitioning to the new world order anyway. [15:37] Oh, as Colin said. [15:38] I follow now. Thanks! === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [18:06] xnox: could you respond to bug 1206114 please? I get the impression that: 1) removing the old binary won't help anyway, since he's using a versioned dependency; 2) we don't seek compatibility between release in this way, instead preferring to fix the packages; 3) if he wants that change, he should petition the Debian maintainer; but I'd prefer someone else to tell me that I'm being accurate. [18:06] Launchpad bug 1206114 in apache2 (Ubuntu) "Package apache2-prefork-dev broken in Saucy" [Undecided,New] https://launchpad.net/bugs/1206114 [18:51] rbasak: Done. I may have been a bit harsh, but his tone rubbed me the wrong way. :/ === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === james_ is now known as Guest77853 [19:08] infinity, do you take care about the other binutils-gold issues too? [19:10] doko: I haven't started looking at reverse-build-conflicts yet (other than rebootstrapping fpc, which I already did, and fixing glibc), but that's next on my list. [19:10] doko: And I need to make yade no longer ftbfs for that one to be fixed. [19:11] Hopefully, the build-conflict list isn't terribly long. [19:12] doko: I know we're also carrying some deltas in some KDE packages to remove -fuse-ld=gold because it was previously not understood. Now that we're carrying that patch to gcc, we can back those out, but I need to find them (not critical, though, they'll just keep building with bfd, which works fine). [19:13] thanks. binutils-gold wasn't very helpful with cross builds either [19:13] doko: Were you planning on backporting the -fuse-ld patch to previous GCC versions too, or have it only work in >= 4.8? [19:13] it should be there [19:13] Oh, cool. I hadn't checked. [19:14] but distro specific [19:14] Yeah, I know gcc-4.7 used to choke on the option, but RH carried the patch in their 4.7 [19:14] Which is why I asked. [19:20] (saucy-amd64)root@cthulhu:~# zgrep '^Build-Conflicts:.*binutils-gold' /var/lib/apt/lists/*Sources | wc -l [19:20] 21 [19:20] I guess that's not unmanageable. [19:26] doko: Erm, most of them are yours: http://paste.ubuntu.com/5933849/ [19:28] infinity: thanks [19:29] doko: gcc-4.8 being a false positive because it's wrong in -release but fixed in -proposed, but I suspect all the rest still need fixing. [19:31] And Laney's already fixed xpdf. [19:33] yes, the cross packages are not a priority for now [19:35] Sure, just pointing out that other than a few stragglers (and I'm chasing them all down), the rest are GCC or based on GCC packaging. [19:36] But I'll make sure all the !gcc stuff is fixed. The others will annoy you in the next rebuild test if you forget to fix them before then. ;) [20:09] Ursinha: hey, long time no talk! I have a launchpad question for you, why can't I link https://launchpad.net/phablet-image-info/i9100/+setbranch to https://code.launchpad.net/~i9100-image-dev/phablet-image-info/i9100 [20:11] sergiusens, hey, let me see [20:15] sergiusens, wild guess: it seems that project isn't set to use codehosting in launchpad (which is weird because you have branches there hehe) [20:15] sergiusens, are you admin of that project? [20:15] Ursinha: dholbach is I think... [20:16] Ursinha: oh, I am admin [20:27] sergiusens, haha [20:27] * ogra_ has a deja vu ... i had the same moment of surprise this morning [20:30] ogra_: wanted to do the same thing? [20:30] :-) [20:30] heh, no, but discovered that we're all admins