[07:44] Could one of the ubuntuwire admins please clean up http://qa.ubuntuwire.com/bugs/rcbugs/ === cpaelzer is now known as cpaelzer_afk === cpaelzer_afk is now known as cpaelzer === cpaelzer is now known as cpaelzer_afk [16:00] rbasak, hi you there? I finally made llvm-3.7 build on i386 and updated the bug report. sorry for the long wait, but I had to upload 3 different versions to get one working :) [16:59] LocutusOfBorg1: hi [17:09] LocutusOfBorg1: looking [17:09] hi you both :) [17:15] LocutusOfBorg1: in your PPA test you disabled both LLDB build and test, but your submitted debdiff only disables test. [17:15] I'm not sure how I feel about this. [17:15] If it works then it's better than a complete FTBFS I suppose. [17:16] But are we shoving something under the carpet that will be a maintenance burden for us later? [17:16] LocutusOfBorg1: let me comment in the bug. If doko doesn't object then I'll sponsor. [17:16] LocutusOfBorg1: in the meantime do you mind confirming in your PPA that disabling just the LLDB test works please? [17:17] Also, is it suitable to build it if it is not being tested? Or if the test is a problem perhaps we _should_ also disable the build? [17:17] I really appreciate your help here. My problem is that I don't know enough about LLVM to be completely confident sponsoring. So let's see what doko thinks. [17:25] rbasak, you are probably right, I uploaded on my ppa a -5 revision with just the debdiff applied [17:25] I don't remember if I rolled back between my previous attempt [17:25] (not enough space on my pc to do the builds) [17:26] LocutusOfBorg1: if you think it's a buildd issue you probably should use a PPA to eliminate differences. [17:26] Although AIUI a PPA is not quite identical to a buildd environment (yet?), at least it's close. [17:27] true story [17:28] about testing this is a problem common on many architectures [17:28] LLDB_DISABLE_ARCHS := arm64 hurd-i386 ia64 ppc64el s390x powerpc alpha [17:28] and later armhf and armel [17:29] well armhf and armel have lldb enabled but testsuite disabled [17:29] and I did the same for i386 [17:30] I don't have an opinion about my build fix, probably some more deep issue is just hidden, I hope doko will have the final and correct decision on the matter === cpaelzer_afk is now known as cpaelzer [17:37] LocutusOfBorg1: yeah copying armhf and armel was reasonable. I just don't like to upload things that I don't understand, and I'm not sure why that's the case already. [17:38] If doko doesn't have an opinion then I'm happy to upload based on the fact that you're looking after it and it'll rot otherwise, so it's better than doing nothing. [17:41] I did ~10 uploads of llvm stuff for Debian this year, fixed various RC and other bugs, but I still don't understand it :) [17:41] I guess instead of just disabling it, it might be better to disable the testsuite. [17:42] anyway, not many people should use LLDB features, and even less on i386 (because it is disabled mostly everywhere this LLDB), so at the end we can say, who cares about something that not many people use, on a particular version of llvm stack, on a particular architecture, on a particular ubuntu version? [17:43] LocutusOfBorg1: if there is consensus on that (ie. nobody objects) then I'd be fine with disabling both LLDB and the testsuite. [17:43] (the testsuite for LLDB that is) [17:43] Building something but disabling that thing's tests I'm less happy with. [17:44] (I should even check if I did disable them both or just the testsuite in my ppa lol) [17:44] this seems to be a fact for armel and armhf [17:45] That seems to have come from Debian though and I have no say in it. [17:47] (it comes from the Debian maintainer, and he is also upstream) [17:47] upstream did a lot of work on arm* recently, but some tests are still hanging/failing [17:47] See, you know far more about this than I do. I'm unqualified to make a decision. :-) [17:50] I think the only one that can make a decision is doko :) === cpaelzer is now known as cpaelzer_afk === cpaelzer_afk is now known as cpaelzer === plars is now known as plars-off [18:52] https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa/+build/8743116 [18:52] it was good the debdiff === cpaelzer is now known as cpaelzer_afk [20:15] rbasak: For architectures that have completely moved to scalingstack (amd64, i386, ppc64el), they're identical except for a few configuration details affecting pkgbinarymangler / pkg-create-dbgsym. [20:41] cjwatson: useful to know, thanks. I think I do remember you saying so in a recent update but I'm not very good at distinguishing roadmap from delivery in my casual memory as things keep moving from one to another :)