| pitti | cyphermox: hm, do you have an actual .crash file report with a backtrace? | 05:09 |
|---|---|---|
| Logan | sorry about all the last-minute uploads/syncs, trying to fix a lot of those rebuild FTBFSes :P | 05:50 |
| pitti | Laney: rejecting one of your duplicate stsci.distutils syncs, don't get confused | 06:31 |
| pitti | Laney: sorry, not you | 06:40 |
| pitti | Logan: rejecting one of your duplicate stsci.distutils syncs, don't get confused | 06:41 |
| Logan | oh yeah, my bad for requesting twice | 06:41 |
| Logan | I thought it didn't go through the first time | 06:41 |
| Logan | (Launchpad email notifications still leave something to be desired) | 06:41 |
| pitti | cyphermox, infinity: in case you pung about bug 1508075, I just followed up | 07:17 |
| ubot2 | bug 1508075 in policykit-1 "crash during network setup in ubiquity wily" [Critical,Confirmed] https://launchpad.net/bugs/1508075 | 07:17 |
| pitti | infinity: can you test core arm64 and powerpc? I tested the others, but I don't have root access to these arches | 07:57 |
| * pitti goes to test ubuntu desktop i386 | 07:58 | |
| pitti | hm, can someone confirm that ubuntu i386 desktop live session is stuck? I see the background pic and nothing else | 08:14 |
| pitti | both with default qemu vga driver and -vga vmware | 08:14 |
| pitti | everything is okay on amd64 | 08:15 |
| smb | pitti, in a VM? | 08:15 |
| pitti | yes | 08:15 |
| smb | pitti, use kvm64 cpu type | 08:15 |
| pitti | that has never been necessary -- is that some kind of mesa regression or so? | 08:15 |
| pitti | smb: "64" for 32 bit ubuntu? | 08:16 |
| smb | default is qemu64 which is an odd enough type to break llvm-pipe | 08:16 |
| pitti | ah | 08:16 |
| smb | pitti, and yes theat is "known" to people since dunno Trusty? | 08:16 |
| smb | pitti, at least when you ask them | 08:16 |
| * pitti searches for an existing bug | 08:17 | |
| smb | pitti, give me a sec and I could tell you | 08:17 |
| pitti | smb: thanks | 08:17 |
| smb | pitti, one at least is bug 1448985 | 08:18 |
| ubot2 | bug 1448985 in qemu "llvmpipe i386 crashes when running on qemu64 cpu" [Medium,Confirmed] https://launchpad.net/bugs/1448985 | 08:18 |
| pitti | smb: cheers, I'll add that to the tracker | 08:19 |
| * pitti links that in the tracker result, so that it appears in the future | 08:19 | |
| smb | pitti, np. Spent hours to understand that one and then Adam came and told me "oh yeah we know that" :-P | 08:20 |
| pitti | smb: I remember that I had to change -cpu in autopkgtest to get around that | 08:20 |
| flexiondotorg | cyphermox, We've found the the Deja Dup plugin for Caja has a namespace collision, so doesn't work anymore. | 08:21 |
| flexiondotorg | cyphermox, I've spoken with the developer (who is on the Ubuntu MATE team). | 08:22 |
| flexiondotorg | He is not available until much later today. | 08:22 |
| flexiondotorg | But is happy for me to provide and updated debdiff and tarball. | 08:22 |
| smb | pitti, ah right. Yeah either core2duo or kvm64 is ok. Only qemu64 is modeling some weird athlon with virt-extensions that never existed as real hw. And my guess is llvm-pipe assumes some things either from cpuid or id name which does not hold true when running the assembly code for it. Unfortunately the llvm-pipe error message is rather useless. At least to me | 08:22 |
| flexiondotorg | cyphermox, Does that sound like a good way to proceed? | 08:23 |
| pitti | smb: right, that rings a bell; I just used -cpu host for now | 08:23 |
| smb | pitti, yup, that should work as well. | 08:23 |
| smb | pitti, ok, fwiw I also did a quick install on an old netbook, so i386 desktop iso via life session and then install. Not sure the final reboot really got stuck with a black screen there as well or I am just getting too impatient in my old days. three finger salute did finish things... | 09:22 |
| cyphermox | flexiondotorg: sorry, what does that mean? you'd need a respin? | 09:51 |
| cyphermox | flexiondotorg: I think it probably should be a SRU at this point | 09:52 |
| cyphermox | flexiondotorg: I'll go back to my "I'm not in the release team so my opinion doesn't matter much, you could argue with them" cop out, but I think it's better to SRU the fix at this point, especially when it's "just" for dejadup. | 09:53 |
| pitti | ^ FWIW, sounds sensible | 09:58 |
| pitti | aside from that, it sounds we need a respin for Kubuntu for the kwallet-kf5 thing, right? | 09:59 |
| flexiondotorg | cyphermox, No respin required. | 10:02 |
| flexiondotorg | cyphermox, Just an upload. | 10:02 |
| flexiondotorg | cyphermox, OK, SRU it is. | 10:02 |
| pitti | flexiondotorg: oh, it's a package which isn't on teh images? which package is that? | 10:02 |
| flexiondotorg | infinity, Are the images getting spun again? | 10:02 |
| seb128 | bug #1502821 | 10:03 |
| ubot2 | bug 1502821 in usb-creator "usb-creator-* fails with message cannot install bootloader" [Undecided,New] https://launchpad.net/bugs/1502821 | 10:03 |
| pitti | flexiondotorg: likely, yes | 10:03 |
| flexiondotorg | pitti, It is seeded, in Ubuntu mATE only. | 10:03 |
| seb128 | states | 10:03 |
| seb128 | "it looks for files into directory /usr/lib/syslinux/ but (in ubuntu Wily) the correct path is | 10:03 |
| seb128 | /usr/lib/SYSLINUX/" | 10:03 |
| pitti | flexiondotorg: then it needs a respin | 10:03 |
| seb128 | ^ that might be something to fix for release? | 10:03 |
| flexiondotorg | seb128, OK, I'll prepare a debdiff. | 10:03 |
| pitti | we'll need a respin for bug 1508075 and for the "ubiquity decorations are wrong" thing that seb128 was looking into | 10:03 |
| ubot2 | bug 1508075 in udisks2 "ubiquity and others time out on polkit (killed by udisks2-inhibit)" [Critical,Incomplete] https://launchpad.net/bugs/1508075 | 10:03 |
| seb128 | flexiondotorg, for usb-creator? | 10:03 |
| pitti | smb: I get the shutdown screen with "remove install media and press enter" all the time, hmm | 10:05 |
| pitti | smb: we fought with that during the vivid release sprint | 10:05 |
| seb128 | Laney, ^ you said you would be interested to fix usb-creator issues? ;-) | 10:06 |
| * pitti claims the "first mandatory green 9/9" flag for desktop i386 | 10:07 | |
| infinity | flexiondotorg: As pitti says, we have a couple of bugs we'll probably respin for, assuming fixes happen. | 10:07 |
| infinity | The only one I consider RC is the polkit thing. | 10:07 |
| infinity | pitti: If there's any help you can lend on the polkit bug, more brains are better. Even German ones. | 10:08 |
| flexiondotorg | infinity, Thanks. I'll not burn another PowerPC DVD just yet then ;-) | 10:08 |
| pitti | infinity: was just talking to cyphermox; I asked him to test an idea which has a pretty good chance to succeed | 10:08 |
| infinity | pitti: \o/ | 10:08 |
| pitti | infinity: but I can't reproduce the hang on any of my three platforms, so I need some testing help | 10:08 |
| seb128 | anyone who knows about syslinux and want to comment on the usb-creator issue just mentioned? | 10:08 |
| seb128 | like did the path change and on purpose? | 10:09 |
| infinity | seb128: If you can hunt down mdeslaur, he touched usb-creator in funny ways this cycle. | 10:09 |
| infinity | seb128: As for paths changing, check the Debian maintainer field. :P | 10:09 |
| seb128 | infinity, he's probably still sleeping | 10:09 |
| pitti | infinity: I'd like to sync https://tracker.debian.org/news/720214 once it's imported into LP, to mop up some small fixes (most importantly, fix backportability to trusty) | 10:10 |
| infinity | pitti: It's not seeded, I'm fine with that. | 10:11 |
| pitti | infinity: and fixing the "prepare VM" scripts to give a working minimal VM again (lxc/lxd was added very late) | 10:11 |
| pitti | infinity: ok, what I thought | 10:11 |
| * pitti sighs at cloud images carrying more and more unnecessary stuff | 10:12 | |
| cyphermox | it's ever so slightly less worse, but still tends to be racy | 10:14 |
| infinity | seb128: /usr/lib/syslinux is in syslinux-common, BTW, not syslinux. | 10:14 |
| pitti | cyphermox: still getting timeouts? | 10:14 |
| cyphermox | yes | 10:16 |
| cyphermox | seb128: usb-creator was correct, and working fine this cycle unless something changed recently | 10:16 |
| cyphermox | aside from issues if there's an architecture mismatch | 10:16 |
| seb128 | cyphermox, it fails to create the bootloader for me and I found that but with the details I just copied | 10:17 |
| seb128 | testing if changing syslinux to SYSLINUX fixes it | 10:17 |
| cyphermox | well, it might if you have that path, but I think it would be good to know why it didn't find the other path | 10:17 |
| seb128 | which one? | 10:18 |
| pitti | cyphermox: ok, thanks for testing; it was worth a try, although I still don't quite understand this | 10:18 |
| flexiondotorg | infinity, I need a respin. Allegedly. | 10:18 |
| flexiondotorg | https://bugs.launchpad.net/ubuntu-mate/+bug/1505505 | 10:19 |
| ubot2 | Launchpad bug 1505505 in ubuntu-mate "ubuntu-mate-desktop does depend on vlc-plugin-pulse but that got removed" [Undecided,New] | 10:19 |
| flexiondotorg | infinity, I'll update the seeds in a sec. | 10:19 |
| cyphermox | seb128: syslinux-common | 10:19 |
| seb128 | cyphermox, infinity, Laney, in fact the path seems correct on disk, maybe that user has a local problem. My bug was because I picked an amd64 iso on a 32 bits machine I think, sorry about the noise :-/ | 10:22 |
| infinity | seb128: It's just a missing dep. We'll fix it here. | 10:25 |
| infinity | seb128: Just needs to depend on syslinux-common. | 10:25 |
| seb128 | infinity, thanks | 10:26 |
| flexiondotorg | infinity, Can you update ubuntu-mate-meta please? | 10:26 |
| flexiondotorg | Can someone please sponsor this update please? https://bugs.launchpad.net/ubuntu-mate/+bug/1508353 | 10:26 |
| ubot2 | Launchpad bug 1508353 in deja-dup-caja "deja-dup-caja plugin has a namespace collision in Ubuntu MATE 15.10 [debdiff]" [Undecided,In progress] | 10:26 |
| pitti | cyphermox, infinity: ok, give me some time for this; at this point we need an unintrusive and small fix, teaching polkitd about SIGHUP doesn't sound like that. I'll try some other things first | 10:29 |
| cyphermox | no, it's not unobtrusive enough | 10:30 |
| cyphermox | but I think your systemctl try-restart is the best we have right now | 10:31 |
| pitti | cyphermox: I thought that still leads to timeouts, and thus is useless? | 10:31 |
| cyphermox | I'm going to try it some more, see if it's good enough to get us something that looks good | 10:31 |
| cyphermox | it does sometimes lead to timeouts, but it's better than it was | 10:31 |
| cyphermox | and my testing method isn't exactly what people usually do - starting and stopping and restarting ubiquity repeatedly | 10:31 |
| pitti | unfortunately polkit doesn't support /run/polkit-1/localauthority/ | 10:32 |
| pitti | on the live system we don't actually need the bind mount though, but I would be surprised if inotify worked on aufs | 10:32 |
| pitti | anwyay, that's part of "I'll try some things" | 10:32 |
| infinity | Riddell: You around? | 10:40 |
| infinity | Riddell: If so, can you try pitti's workaround for this bug, since you reproduce it better than anyone? | 10:41 |
| pitti | not that important really, as it'll just a bit less bad, but still not "good" | 10:41 |
| pitti | I have about three ideas which I'm currently testing | 10:41 |
| infinity | pitti: If "less bad" is the best we can do, it's still better. :/ | 10:41 |
| infinity | pitti: But, yeah. Goodest would be more gooder. | 10:42 |
| infinity | pitti: The live system is overlay, not aufs, and inotify definitely doesn't work. :P | 10:43 |
| pitti | infinity: I'm currently testing that, but that's what I was afraid of | 10:43 |
| infinity | pitti: In fact, it's actively broken, which is even worse than not working (as in, the FS claims to support it, then does nothing useful when you hook into it). | 10:43 |
| pitti | I have a solution which works on my installed sytesm | 10:43 |
| pitti | hm, it actually does trigger, curious | 10:44 |
| pitti | sudo G_MESSAGES_DEBUG=all /usr/lib/policykit-1/polkitd -r | 10:45 |
| pitti | sudo touch /var/lib/polkit-1/localauthority/90-mandatory.d/foo.pkla | 10:45 |
| pitti | -> polkit reacts (that's on a live system) | 10:45 |
| pitti | not expected (but would be nice indeed) | 10:45 |
| pitti | pkcheck -a org.freedesktop.udisks2.filesystem-mount -p $$; echo $? | 10:45 |
| pitti | that should be 0 for normal mode, and 1 for inhibited mode | 10:46 |
| pitti | anyway, I still have another idea, let me test/prod more | 10:46 |
| pitti | cyphermox: can you reproduce that with ubiqutiy under the live system? i. e. where it's non-agonizing to downlaod/update a file? | 10:47 |
| pitti | Riddell: ^ | 10:47 |
| infinity | pitti: Is that path an overlay, or a tmpfs? | 10:49 |
| cyphermox | pitti: of course | 10:50 |
| pitti | infinity: it's on teh normal overlayfs | 10:50 |
| infinity | pitti: Huh. Then something's polling manually, cause inotify definitely doesn't work on overlay. :P | 10:50 |
| pitti | cyphermox: ack, I should have a new udisks2-inhibit for testing then | 10:50 |
| cyphermox | pitti: what what? | 10:50 |
| cyphermox | pkcheck I didn't try | 10:50 |
| pitti | cyphermox: I mean, in 5 mins, sorry | 10:50 |
| cyphermox | I just rebooted | 10:50 |
| cyphermox | your systemctl try-restart fix looks sufficient | 10:51 |
| cyphermox | it helps if one changes both kills, for sure. | 10:52 |
| pitti | infinity: it does use inotify,confirmed in strace | 10:52 |
| cyphermox | now, to run all this in a vm so I can actually complete the install and make sure things work correctly | 10:53 |
| pitti | cyphermox: the second kill is only when ubiquity finishes, though | 10:53 |
| cyphermox | sure | 10:53 |
| cyphermox | but that's *why* I said it wasn't quite there | 10:53 |
| pitti | cyphermox: oh, you only got hangs on the second restart? | 10:53 |
| pitti | ah | 10:53 |
| cyphermox | if it gets killed, it hangs on second starts or NM might not have a polkit ready | 10:53 |
| pitti | right, *if* the restart helps, then it needs to go to both places | 10:54 |
| cyphermox | looks good so far | 10:55 |
| infinity | pitti: Then I question that it's on overlay, unless someone fixed that in the last 6 months, but... Kay. | 10:55 |
| infinity | pitti: (I could have sworn I saw a weird polkit related mount in my last live session) | 10:56 |
| pitti | infinity: yes, you did | 10:56 |
| pitti | infinity: it *currently* doesn't rely on inotify, but it seems it actually works now | 10:56 |
| pitti | infinity: perhaps since the new version that landed in the upstream kernel (remember we had to change casper and everything for the new syntax) | 10:57 |
| infinity | pitti: Oh, actually. It might be that it works for new files, but not changed files or something. | 10:57 |
| apw | overlayfs and inotify, i'd be epically supprised if it works reliably | 10:57 |
| pitti | infinity: right, but that'd be sufficient | 10:57 |
| infinity | pitti: Yeah. | 10:57 |
| pitti | new file and rm seem to work | 10:57 |
| pitti | (repeatedly) | 10:58 |
| pitti | i. e. it seems to work for dirs, but not files | 10:58 |
| pitti | (changing files does *not* work, I tested) | 10:58 |
| infinity | New files are new inodes, changed files don't change the inode with the watch on, but instead give you a new inode and the world explodes. | 10:58 |
| apw | pitti, right because directories are a thing at the overlayfs level, files arn't, they exist in the real filesystems but your watches are not in the right world | 10:58 |
| infinity | apw: I still think the fix for that is just to copy-up when a watch is added, and attach the watch to the copy. | 10:59 |
| infinity | apw: Bit of a waste of space, but would be reliable. | 10:59 |
| pitti | new udisks2-inhibit works locally; /me tests it on live system | 11:00 |
| apw | infinity, except it is more complex in that there are are three places and you are in the thrid one, i have patches which | 11:01 |
| apw | make overlayfs a consumer of inotify on the files it uses from the lower layers, and propogate them to the real upper watches | 11:02 |
| pitti | \o/ works fine | 11:02 |
| infinity | apw: Three? Doesn't the inode on the combined filesystem either match the underlay (before a copy-up) or the overlay (after a copy-up), so there's really only two options? | 11:02 |
| cyphermox | pitti: how new is it? | 11:02 |
| pitti | cyphermox: sorry, how new is what? | 11:02 |
| infinity | apw: Oh. I guess I'm nothing thinking of multiple layer cases. | 11:02 |
| * cyphermox has been testing systemctl try-restart in an install now | 11:02 | |
| apw | but they are invasive, and they don't fix the issue that there is no way to represent copy up itself | 11:02 |
| cyphermox | how new is your udisks2-inhibit? | 11:02 |
| apw | infinity, the third place is the overlayfs inode, which doesn't exist per-see but does from a notify point of view | 11:03 |
| flexiondotorg | infinity, Can you update ubuntu-mate-meta please? | 11:03 |
| cyphermox | flexiondotorg: do you have your caja fix too? | 11:04 |
| flexiondotorg | cyphermox, https://bugs.launchpad.net/ubuntu-mate/+bug/1508353 | 11:04 |
| ubot2 | Launchpad bug 1508353 in deja-dup-caja "deja-dup-caja plugin has a namespace collision in Ubuntu MATE 15.10 [debdiff]" [Undecided,In progress] | 11:04 |
| pitti | infinity, cyphermox: attached the kill-less inotify scrip to the bug | 11:05 |
| infinity | flexiondotorg: I can do your meta update. | 11:05 |
| flexiondotorg | infinity, Thank you! | 11:06 |
| infinity | flexiondotorg: I wouldn't recommend a rebuild until we sort this ubiquity thing, since I have to rebuild everyone for that anyway. | 11:06 |
| flexiondotorg | infinity, Understood. | 11:06 |
| pitti | cyphermox, infinity, Riddell: I attached both variants to the bug now | 11:11 |
| infinity | flexiondotorg: Is there a bug number for this meta update? | 11:11 |
| infinity | flexiondotorg: Ahh, found it. | 11:12 |
| pitti | it's a "pick your poison" decision: the inotify one relies on overlayfs and might leave dangling symlinks behind if you actually use that in a real system, and the restart one might break pending calls to polkit (but not existing connections any more, which is the bug here) | 11:13 |
| pitti | I actually have a better solution in mind using temporary udev rules (which have neither problem), but that's more intrusive | 11:13 |
| pitti | if the restart one works, then my recommendation is to use that for wily as that's the least intrusive | 11:14 |
| pitti | if it doesn't work, then let's use the inotify one | 11:14 |
| infinity | Or both! | 11:17 |
| infinity | Two hammers are better than one hammer. | 11:17 |
| infinity | That's what my hammer salesman told me. | 11:17 |
| pitti | infinity: sorry, it's just one shaft with a hammer head on either end | 11:18 |
| infinity | pitti: It's PHP? | 11:18 |
| infinity | Oh, that was two claws. | 11:18 |
| infinity | https://www.flickr.com/photos/raindrift/sets/72157629492908038/ | 11:18 |
| pitti | there; I was going to work on the udev rule one and now I need to chuckle | 11:18 |
| pitti | as long as it's not like this: http://vignette4.wikia.nocookie.net/starwars/images/1/1a/Darth_Maul_lightsaber_reveal.png/revision/latest?cb=20140421143551 | 11:19 |
| pitti | infinity, cyphermox: remind me, why do we have this inhibit thing in the first place? for installations to USB drives, right? (as internal drives aren't being automounted anyway) | 11:21 |
| pitti | or, I guess it was for preventing the user from clicking on the internal drive icons and thus manually mounting them | 11:24 |
| cjwatson | automounting hasn't always been inhibited consistently on internal drives, and yes, external ones matter too | 11:25 |
| cjwatson | ubiquity needs complete control muahahahaha etc. | 11:25 |
| pitti | just tested on current live system, manual mounting is all too easy too | 11:26 |
| cjwatson | cf. http://www.chiark.greenend.org.uk/~cjwatson/blog/desktop-automount-pain.html :-) | 11:26 |
| pitti | so just setting UDISKS_AUTO=0 won't suffice | 11:26 |
| pitti | cjwatson: heh, right; but most of them won't prevent manual mounts, which AFAICS is the bigger problem? | 11:27 |
| cjwatson | the original motivation was things randomly capturing temporary mount points | 11:27 |
| pitti | that was the idea behind temporarily changing the polkit policy -- then you both disable auto- and manual mounts, and it works for all desktsop | 11:27 |
| cjwatson | but suppressing manual mounts would be good too | 11:27 |
| flexiondotorg | infinity, Thanks. | 11:37 |
| cyphermox | pitti: upload whichever one you prefer, both your fixes seem to work correctly here. | 11:37 |
| cyphermox | unless Riddell says otherwise | 11:37 |
| pitti | cyphermox: cool, thanks for testing | 11:37 |
| cyphermox | I can only test it so much, it's hard to reproduce | 11:37 |
| pitti | cyphermox: I'll flip a coin :) but indeed I'm leaning towards the restart thingy as that's the least intrusive | 11:38 |
| cyphermox | ok | 11:38 |
| cyphermox | well, the inhibit one seemed to be quite nice | 11:38 |
| cyphermox | not restarting means it's even less intrusive to other things on the session | 11:38 |
| pitti | I'm testing the temp udev rules one right now, which is kind of "best of both worlds" | 11:38 |
| pitti | but I wouldn't do that a day before release | 11:39 |
| cyphermox | right | 11:39 |
| cyphermox | but after release I'd just teach polkit to do USR1 or HUP. | 11:39 |
| cyphermox | seems like a fairly useful trick | 11:39 |
| * cyphermox takes a backup to be able to run at least one install test on real hardware | 11:40 | |
| pitti | yay, we have an X-series bug target in LP now | 11:40 |
| pitti | cyphermox: oh, don't you guys have infinity's craptop as a victim in the office? | 11:41 |
| infinity | I didn't bring my spare laptop this time, no. Didn't want the weight when I'm on the road for three weeks. | 11:41 |
| cyphermox | I considered bringing craptops, but didn't want to take more shit out at airport securities | 11:41 |
| pitti | I'd have thought in the office there sohuld be a spare or three | 11:43 |
| infinity | There might be. | 11:43 |
| pitti | ^ kthxbye lunch o'clock | 11:44 |
| cyphermox | pitti: thanks | 11:45 |
| pitti | infinity: if you review/accept it seems a a perfect lunch time for you as well, for LP/britney to do their things :) | 11:45 |
| seb128 | ^ u-s-d, that's for bug #1508327 | 11:46 |
| ubot2 | bug 1436861 in unity-settings-daemon "duplicate for #1508327 unity-settings-daemon crashed with SIGSEGV in engine_update_composite_device()" [High,In progress] https://launchpad.net/bugs/1436861 | 11:46 |
| seb128 | would be nice to get in as well | 11:46 |
| seb128 | both didrocks and I got it while testing ubiquity install-mode | 11:46 |
| pitti | https://launchpadlibrarian.net/221863227/unity-settings-daemon_15.04.1%2B15.10.20151012-0ubuntu1_15.04.1%2B15.10.20151021-0ubuntu1.diff.gz | 11:46 |
| seb128 | it leads to having wm decorations on the right and probably some other issues | 11:46 |
| pitti | ^ debdiff for that, as it's non trivial to find | 11:46 |
| flexiondotorg | This bug is confirmed not present in Ubuntu MATE but still present in Unity - https://bugs.launchpad.net/ubuntu-mate/+bug/1437764 | 11:47 |
| ubot2 | Launchpad bug 1437764 in msttcorefonts "ttf-mscorefonts-installer doesn't work from Ubuntu Software Center because of EULA, breaks APT" [Undecided,Confirmed] | 11:47 |
| flexiondotorg | I didn't do anything to fix it. | 11:47 |
| seb128 | could be a missing frontend? | 11:47 |
| pitti | seb128: is the "devices = " just for not having to write manager->priv->devices_array several times in the loop, or is there some funky refcounting thing going on? | 11:48 |
| pitti | seb128: i. e. the gist of the change is to disconnect signal handlers before freeing, right? | 11:48 |
| seb128 | pitti, it's just to improve readability | 11:48 |
| seb128 | yes | 11:49 |
| pitti | seb128: I thought one would use g_ptr_array_new_with_free_func() for such things, and disconnect handlers there | 11:49 |
| seb128 | pitti, same reasoning than https://bugzilla.gnome.org/show_bug.cgi?id=673007#c17 | 11:49 |
| pitti | but anyway, LGTM | 11:49 |
| ubot2 | Gnome bug 673007 in general "[power]: gnome-settings-daemon crashed with SIGSEGV in engine_update_composite_device()" [Normal,Resolved: fixed] | 11:49 |
| infinity | pitti: Are you reviewing u-s-d? | 11:49 |
| seb128 | pitti, feel free to comment on #ubuntu-desktop for larsu about that | 11:50 |
| seb128 | we can improve that with another commit after wily | 11:50 |
| pitti | infinity: I was reviewing, seems Laney is currently pushing the button for it | 11:50 |
| pitti | seb128: yeah, nevermind | 11:50 |
| seb128 | thanks for the review! | 11:50 |
| Laney | pitti: I didn't | 11:50 |
| seb128 | lunch now, I'm starving | 11:50 |
| seb128 | bbiab | 11:50 |
| Laney | I saw you were reviewing :) | 11:50 |
| Laney | just saying that we will take this change | 11:50 |
| pitti | ack; accepting then | 11:51 |
| pitti | "The Unapproved queue is empty." | 11:51 |
| pitti | and so is my stomach | 11:51 |
| * pitti bbl | 11:51 | |
| Laney | thanks | 11:51 |
| pitti | ^ oooh! | 12:03 |
| pitti | we have an adjective? | 12:03 |
| cjwatson | shh | 12:03 |
| Laney | sssshhhhhh | 12:03 |
| cjwatson | (also, if we're fired, I blame infinity) | 12:05 |
| doko | promoted the slimit binary, only the source was promoted | 12:15 |
| infinity | Huh, I thought I got that in the queue yesterday. Thanks. | 12:15 |
| smb | pitti, really it might be just me being too quick... I am not used to a puny netbook with an old Atom CPU. | 12:18 |
| pitti | Laney: vim/amd64 WTH? | 12:55 |
| Laney | pitti: ? | 13:00 |
| Laney | did it fail? | 13:00 |
| pitti | Laney: yes, in a really weird way; I retried it | 13:00 |
| pitti | Laney: https://launchpadlibrarian.net/221870097/buildlog_ubuntu-wily-amd64.vim_2%3A7.4.712-2ubuntu4_BUILDING.txt.gz | 13:00 |
| * Laney blinks | 13:00 | |
| pitti | Laney: seems the new build is past that point already | 13:01 |
| Laney | worrying | 13:02 |
| Laney | but thanks for retrying :) | 13:02 |
| didrocks | the new build was just a better build, that's it :p (hem… ;)) | 13:03 |
| infinity | pitti: Can you hunt down why the kio/i386 autopkgtest isn't autopkgtesting? | 13:13 |
| pitti | https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-wily/wily/i386/k/kio/20151021_095959@/log.gz | 13:14 |
| pitti | argh that again | 13:14 |
| * pitti radiates hate towards lcy01 | 13:14 | |
| pitti | infinity: yep, on it | 13:14 |
| pitti | http://autopkgtest.ubuntu.com/status/alerts/ - eww | 13:14 |
| infinity | pitti: I retried it. But I'm not convinced it's retrying. | 13:14 |
| pitti | infinity: next britney run will re-trigger them | 13:17 |
| * pitti disabled lcy01 again, next runs will use lgw01 and clean up the mess | 13:18 | |
| pitti | infinity: http://autopkgtest.ubuntu.com/running.html -> kio/i386 is actually running | 13:18 |
| infinity | pitti: I always forget the URL to that page. | 13:21 |
| pitti | infinity: "Running" in the menu on a.u.c. | 13:21 |
| infinity | ... which is linked from the top level. | 13:21 |
| pitti | infinity: http://autopkgtest.ubuntu.com/status/alerts/ shows all the tmpfails, FYI; i. e. problems with the testbeds | 13:21 |
| doko | are uploads for wily-proposed already accepted? | 13:26 |
| flexiondotorg | cyphermox, Will you have time for the Deja Dup extension fix for Caja? | 14:17 |
| cyphermox | can you point me to the bug again? I was on something else | 14:18 |
| flexiondotorg | cyphermox, Sure. | 14:18 |
| flexiondotorg | cyphermox, https://bugs.launchpad.net/ubuntu-mate/+bug/1508353 | 14:18 |
| ubot2 | Launchpad bug 1508353 in deja-dup-caja "deja-dup-caja plugin has a namespace collision in Ubuntu MATE 15.10 [debdiff]" [Undecided,In progress] | 14:18 |
| cyphermox | flexiondotorg: are you sure this is good? the commented out part in setup.py? | 14:25 |
| Riddell | pitti: infinity: still need testng? | 14:40 |
| flexiondotorg | flexiondotorg, Yep. | 14:41 |
| flexiondotorg | cyphermox, Thanks. | 14:41 |
| flexiondotorg | Did you change debian/install to executable? | 14:42 |
| infinity | Riddell: I'm going to respin you shortly, you can just test the ISO instead. :) | 14:43 |
| Riddell | lovely | 14:43 |
| pitti | Riddell: if you want, you can replace udisks2-inhibit with what's attached to the bug, to get it early | 14:50 |
| pitti | also, the fixed udisks2 is just being published | 14:50 |
| cyphermox | ^ lolz | 14:52 |
| cyphermox | flexiondotorg: deja-dup-caja is atrocious | 14:53 |
| flexiondotorg | cyphermox, In what sense? | 14:57 |
| cyphermox | in the sense that the packaging was broken, and that your patch means the upstream tarball should get fixed | 14:58 |
| cyphermox | (and we rejected twice before getting it right) | 14:58 |
| flexiondotorg | cyphermox, I'm working with the upstream author to fix the tarball either tonight or tomorrow. | 14:58 |
| cyphermox | ok | 14:58 |
| flexiondotorg | Is there anything else you want me to improve. | 14:59 |
| flexiondotorg | Because I'v not ben involved with the development or packaging until tody. | 14:59 |
| flexiondotorg | cyphermox, ^^^^^^^^^^^^^^^^^^^^^^ | 15:03 |
| === fginther` is now known as fginther | ||
| cyphermox | flexiondotorg: nothing critical now; if you can just make sure that the upstream developer fixes the filename in their tarball, then we'll be pretty good | 15:14 |
| cyphermox | oh, and the README too; there's a small diff | 15:14 |
| cyphermox | basically, what's in deja-dup-caja.patch | 15:14 |
| cyphermox | once that is done, there's nothing special to do but to drop the patch with the upload, the rest is not an issue | 15:15 |
| apw | infinity, i've done a quick spin through the release notes, dropped the bugs which appear fixed etc. we are missing updated release notes for most of the flavours at the moment (many are /beta2 still) | 15:21 |
| apw | infinity, we are also missing server | 15:22 |
| flocculant | apw - ours are done - just not changed that page yet | 15:24 |
| apw | flocculant, cool thanks | 15:27 |
| flocculant | apw - though if I'd have waited I could have said I'd done the main page ... | 15:28 |
| flocculant | must be install testing fever | 15:29 |
| infinity | flexiondotorg: Will rebuild MATE once that deja-dup thing publishes in the next 30m. | 15:41 |
| infinity | **** WARNING: REBUILDING THE WORLD **** | 15:41 |
| * ogra_ searches for cover | 15:43 | |
| davmor2 | infinity: \o/ imagesssss the precccciiiouussssss | 15:43 |
| flexiondotorg | infinity, Thanks. | 15:45 |
| flexiondotorg | I'll get my DVD burner at the ready for PowerPC ;-) | 15:45 |
| flexiondotorg | cyphermox, Noted regarding the tarball. | 15:46 |
| cyphermox | flexiondotorg: oh, that's right. getting a ppc machine next week means I'll really need to do my netboot setup. | 15:47 |
| flexiondotorg | cyphermox, What PPC machine you getting? | 15:53 |
| cyphermox | flexiondotorg: some random G5 I might be able to aquire. | 15:54 |
| cyphermox | flexiondotorg: I've been meaning to setup a proper netboot environment regardless. | 15:54 |
| infinity | pitti: How long after something disappears from running.html until is shows up in the test results? | 15:57 |
| infinity | s/is/it/ | 15:57 |
| infinity | pitti: The juju-core tests went away from running.html minutes ago, but I don't see results. Weird. | 16:00 |
| infinity | pitti: Bah. And now they both show up. Finally. As failures. | 16:00 |
| flexiondotorg | cyphermox, Look on ebay, they G5 powermacs are really cheap now. | 16:06 |
| cyphermox | yeah | 16:06 |
| cyphermox | it shouldn't be too hard to find, I just need to email someone actually | 16:07 |
| cyphermox | I just postponed until next week since I'm not in Canada | 16:07 |
| flexiondotorg | cyphermox, There is also the Amiga ONE X500 and X1000 too. | 16:08 |
| flexiondotorg | Which are new. | 16:08 |
| cyphermox | flexiondotorg: which are probably quite expensive then | 16:18 |
| pitti | infinity: up to 5 minutes | 17:06 |
| pitti | infinity: I retried them, in case it helps; but if it's urgent, feel free to hint away | 17:07 |
| infinity | pitti: You're way behind. | 17:07 |
| infinity | pitti: I already hinted them, and the bug has been diagnosed and is being fixed. | 17:07 |
| pitti | panic: osVersion reported an error: Could not determine series ? | 17:07 |
| pitti | behind> heh, indeed I am, just returned quickly to laptop before basketball | 17:08 |
| pitti | infinity: is that the old "no current development series in Launchpad" thing agaain? | 17:08 |
| pitti | ah, autopkgtest finally imported; /me pushes sync button | 17:09 |
| === robru is now known as robru|sick | ||
| infinity | pitti: No, it's a bug in the test. | 17:10 |
| * pitti will be back in some 3 hours after basketball; anything I should be looking out for then? | 17:11 | |
| pitti | (I assume/hope you'll be out by then) | 17:11 |
| pitti | 19:11:25 rbasak | Juju fix uploaded. | 17:11 |
| pitti | \o/ | 17:11 |
| infinity | pitti: Just ISO testing. This spin should be "final" unless something horrible happens. | 17:12 |
| pitti | infinity: ack; I guess I'll give them an initial spin tonight then, and do the finer-grained install cases tomorrow | 17:12 |
| pitti | already went through all of them today after all, and we didn't change that much | 17:12 |
| pitti | (famous last words, I know) | 17:12 |
| flocculant | infinity: assuming that nothing particularly horrid has been seen - these new images just need a quick check? | 17:13 |
| pitti | infinity: this spin> the ones that just landed 45 mins ago, or is another one going on? | 17:14 |
| pitti | i. e. 20151021 or an upcoming 20151021.1? | 17:14 |
| jibel | pitti, final build should be 20151021 | 17:16 |
| infinity | flocculant: Pretty much, yeah. | 17:16 |
| infinity | pitti: What jibel said. | 17:17 |
| pitti | hmm, and still bug 1471476 and bug 1508000 | 17:17 |
| ubot2 | bug 1471476 in libimage-exiftool-perl "[MIR] libimage-exiftool-perl" [High,New] https://launchpad.net/bugs/1471476 | 17:17 |
| ubot2 | bug 1508000 in cmake-extras "[MIR] cmake-extras" [Undecided,New] https://launchpad.net/bugs/1508000 | 17:17 |
| infinity | pitti: Yeah, neither of those affect images, so we just need them sorted by tomorrow afternoon. | 17:17 |
| pitti | the latter could be promoted now, after doko's review and latest comments | 17:17 |
| flocculant | infinity: thanks - that was exactly what I needed to read :p | 17:18 |
| infinity | pitti: Would perhaps be nice if those mythical bugfixes were actually in wily instead of a PPA, but meh. | 17:18 |
| jderose | infinity: will there be 20151021 server ISOs, or did only the desktop need a respin? | 17:21 |
| infinity | jderose: Server's building right now. | 17:21 |
| jderose | infinity: gotcha, thanks! | 17:21 |
| wxl | infinity: so what are we trying to fix with today's global respin, sigh? | 17:37 |
| wxl | …bug numbers might be helpful | 17:38 |
| jibel | wxl, bug 1436861 and bug 1508075 | 17:38 |
| ubot2 | bug 1436861 in unity-settings-daemon "unity-settings-daemon crashed with SIGSEGV in engine_update_composite_device()" [High,Fix released] https://launchpad.net/bugs/1436861 | 17:38 |
| ubot2 | bug 1508075 in udisks2 "ubiquity and others time out on polkit (killed by udisks2-inhibit)" [Medium,Fix committed] https://launchpad.net/bugs/1508075 | 17:38 |
| wxl | jibel: 1436861 required a global respin? | 17:38 |
| infinity | wxl: No, the other bug did. | 17:39 |
| wxl | okie dokie | 17:39 |
| jibel | wxl, 1508075 did | 17:39 |
| jibel | what infinity said | 17:40 |
| infinity | wxl: And no sighing, we almost always have a respin the day before release. :P | 17:40 |
| infinity | I can't remember the last time we just sat around twiddling our thumbs on a Wednesday. | 17:41 |
| jibel | maybe because there was no last time? | 17:41 |
| infinity | OR MAYBE BECAUSE WE HAVE NO THUMBS. | 17:42 |
| jderose | pitti: small inconsistency i noticed: when installing from the desktop ISO, /var/lib/dbus/machine-id is a symlink to /etc/machine-id, but from the server ISO /var/lib/dbus/machine-id is a file (with the same UUID value copied from /etc/machine-id) | 17:46 |
| utlemming | infinity: can you add wily final to cloud.qa.ubuntu.com? | 17:49 |
| wxl | infinity: upgrades haven't been updated. don't the unity ones need to be? | 18:20 |
| wxl | s/updated/respun/ | 18:21 |
| utlemming | slangasek, infinity, stgraber: cloud image test results can't be populated because the product is not on cloud.qa.ubuntu.com yet, and switching to iso.qa.ubuntu.com is erroring out. | 18:24 |
| slangasek | utlemming: "erroring out" how? | 18:28 |
| slangasek | the webpage loads for me | 18:28 |
| utlemming | slangasek: when I try to populate the products it returns a failure | 18:29 |
| utlemming | slangasek: http://paste.ubuntu.com/12887531/ | 18:30 |
| utlemming | slangasek: we moved over to cloud.qa.u.c, for the last release so that the cloud images didn't take over the dashboard. So the real fix is for someone to create the milestone of "Wily Final" there | 18:31 |
| slangasek | ok; I don't know anything about cloud.qa.u.c, would I have admin access on it? | 18:31 |
| slangasek | doesn't look like it | 18:32 |
| utlemming | slangasek: actually, I think I can do it...I was looking at the wrong place...standby | 18:32 |
| utlemming | slangasek: okay, I found that I can do it | 18:33 |
| slangasek | \o/ | 18:33 |
| utlemming | General FYI -- cloud image test results are going to http://cloud.qa.ubuntu.com | 18:34 |
| davmor2 | infinity, jibel: boot from first drive is failing | 18:43 |
| davmor2 | not sure how critical it is | 18:43 |
| davmor2 | it does cause the system to reboot so I guess that will get you to the first hard-drive ish | 18:45 |
| infinity | davmor2: Longstanding issue with some BIOSes. It's hard to fix without breaking other people instead. | 18:49 |
| infinity | davmor2: Unless your contention is that this works on the same machine with an older image. | 18:50 |
| davmor2 | infinity: I'll grab an older image but I think it used to, but this bios has been updated so now might not, it might also be dvd vs pendrive so I'll try that too | 18:51 |
| infinity | davmor2: It's less likely to work with USB drives than with optical drives. | 18:52 |
| infinity | davmor2: As some BIOSes boot USB sticks by remapping them to the logical location of the first hard drive. | 18:52 |
| davmor2 | infinity: indeed that's why I say I'll double check that too once this install finishes | 18:53 |
| cyphermox | go davmor2, break stuff! | 19:11 |
| davmor2 | cyphermox: shhh that's what I'm trying not to do | 19:12 |
| davmor2 | infinity: same issue on dvd let me get a 14.04 image burned | 19:37 |
| davmor2 | infinity: 14.04 daily image dvd boots from first hard drive | 19:47 |
| davmor2 | jibel: ^ | 19:48 |
| davmor2 | I'll write up a bug and blame cyphermox | 19:48 |
| davmor2 | now where does that menu live I guess casper maybe? cyphermox | 19:49 |
| davmor2 | infinity, jibel: https://bugs.launchpad.net/ubuntu/+source/casper/+bug/1508589 | 20:05 |
| ubot2 | Launchpad bug 1508589 in casper "15.10: On the Ubuntu menu on a non uefi install Boot from first disk fails" [Undecided,New] | 20:05 |
| davmor2 | jibel: there is no test for that option on the tracker so shall I attach it anywhere or do you just want to tag it for info | 20:12 |
| flocculant | davmor2: I noticed that a while back - then promptly forgot about it | 20:17 |
| davmor2 | flocculant: feel free to confirm the bug then :) | 20:22 |
| flocculant | davmor2: do you think I wouldn't have ;) | 20:23 |
| flocculant | davmor2: or do you mean actually write something instead of just me tooing? | 20:24 |
| davmor2 | flocculant: no that's fine I didn't look at the bug too busy looking at the other machine for the next install | 20:24 |
| flocculant | ok - then all done :) | 20:25 |
| pitti | jderose: interesting, can you please file a bug about it? a symlink sounds better | 20:32 |
| cyphermox | davmor2: it's in debian-cd actually, I'll reassign | 20:32 |
| pitti | jderose: might be that this is because server doesn't install dbus by default? | 20:32 |
| davmor2 | cyphermox: thanks dude | 20:32 |
| jderose | pitti: yeah, i'll file a bug about it later today or tomorrow. er, i thought dbus (system-wide anyway) was installed by default on the server? | 20:34 |
| pitti | jderose: ah, maybe it is now | 20:35 |
| jderose | pitti: just double checked... yup, installed by default on the server | 20:38 |
| davmor2 | okay that is odd, screen reader not working on i386 dvd but is from i386 thumbdrive meh | 21:10 |
| * Laney peeks in | 21:21 | |
| davmor2 | Laney: mad fool it's all your fault now | 21:22 |
| Laney | /quit | 21:22 |
| flexiondotorg | cyphermox, Should there be a network applet in oem-config? | 21:44 |
| jderose | pitti: filed a bug - https://bugs.launchpad.net/ubuntu/+source/dbus/+bug/1508697 | 21:53 |
| ubot2 | Launchpad bug 1508697 in dbus "Wily: /var/lib/dbus/machine-id is symlink on desktop, file on server" [Undecided,New] | 21:53 |
| === maclin1 is now known as maclin | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
