=== led_dark_2 is now known as led_dark_1 | ||
doko | tsimonq2: https://bugzilla.opensuse.org/show_bug.cgi?id=1121214 seen in libqt | 09:59 |
---|---|---|
ubottu | bugzilla.opensuse.org bug 1121214 in Basesystem "GCC 9: libqt4 build fails" [Normal,New] | 09:59 |
doko | jamespage: murano-agent fails autopkg tests (with python 3.7.2) | 10:17 |
rbasak | juliank: why the dnsmasq upload? I thought Laney was sorting it in the network-manager autopkgtests? | 12:01 |
juliank | rbasak: i don't know, Laney asked me if I'd upload it | 12:02 |
Laney | The dnsmasq upload was the failure; my changes to n-m fixed some other ones, but not those which were failing on autopkgtest.u.c | 12:03 |
rbasak | I feel that this just kicks the can down the road, and I'm not keen on having to maintain this patch :-/ | 12:04 |
rbasak | (based on my current and rather incomplete knowledge of the actual problem; but I suspect that's the case with everyone else involved too) | 12:05 |
Laney | Second best approach that I could see given that we don't know how to fix it | 12:07 |
Laney | Changing the tests in network-manager to work around a bug in dnsmasq would not have been right, IMO | 12:08 |
Laney | and I *did* do some analysis and post on the upstream list, which I still hope/expect to result in a proper fix | 12:08 |
rbasak | I agree with the general principle of not having to change the tests in N-M to work around a bug in dnsmasq. | 12:10 |
rbasak | But it seemed to me that n-m tests weren't actually testing the right thing? | 12:11 |
rbasak | In that the failure couldn't identify a specific buggy use case in dnsmasq? | 12:11 |
rbasak | In that case narrowing down the n-m tests seems like a good idea. | 12:11 |
Laney | what do you mean? | 12:12 |
Laney | It was always a specific few tests that were failing in the same way | 12:12 |
rbasak | But AIUI we haven't been able to say to dnsmasq upstream "here's a bug: when we do <valid use case> it results in <incorrect behaviour> instead of <expected behaviour>" | 12:13 |
Laney | I think my thread was fine in that sense | 12:15 |
Laney | Upstream attempted to come up with a patch but it didn't work | 12:16 |
rbasak | I appreciate your work in analysis and talking to upstream btw. Hopefully that'll lead to a proper fix and we'll be able to drop the patch. I'm just concerned in case that doesn't happen, we'll end up stuck. | 12:16 |
Laney | If someone else wants to help out and create a more minimal reproducer, that would be welcome. I was suffering from a lack of knowledge. | 12:16 |
rbasak | I got the impression that upstream haven't identified a specific bug either. We only have your successful bisection, and we're all still blind to the underlying cause. | 12:16 |
rbasak | (and blind to even what the bug actually is) | 12:17 |
Laney | That's... why you post a bug report? | 12:18 |
rbasak | We know that an upstream commit caused a change in behaviour that broke us. | 12:19 |
rbasak | We don't know that there's actually a bug in upstream dnsmasq...do we? | 12:19 |
Laney | I'm confident enough. | 12:19 |
Laney | But if you're unhappy with my call, it's your team's package. | 12:19 |
rbasak | I'm not sure I agree. But anyway, I appreciate your time on this. It's probably not worth us spending more time on discussing this more unless something further happens. | 12:20 |
Laney | Put the change back in, it'll get blocked in proposed because of breaking network-manager's tests. | 12:20 |
Laney | And then, well, we can argue about who gets to debug it. | 12:21 |
juliank | That seems reasonable | 12:25 |
juliank | We really need to have a convention for changes we made that we do not want to stop syncing. So something between ubuntu and build | 12:26 |
juliank | So that I could have uploaded that hack with that version, and then a later sync would still happen and we'd see it being broken again (or it's fixed and migrates) | 12:27 |
rbasak | Something that starts with 'l' or 'm' then :-P | 12:27 |
juliank | Well, it could be any string lower than ubuntu | 12:28 |
juliank | like buntu | 12:28 |
juliank | 1.0-1buntu1 | 12:28 |
juliank | that was my silly string | 12:28 |
Laney | I should mail the person who wrote the patch in question in the first place and see what they think | 12:28 |
juliank | 1.0-1ubuntmp1 | 12:29 |
juliank | ubuntmp sorts well, but is not entirely obvious | 12:29 |
Amnesia | question, I'm trying to create a cosmic chroot, but debootstrap is stating that the InRelease file has expired: E: InRelease file http://archive.ubuntu.com/ubuntu/dists/cosmic/InRelease is expired since (Wed, 09 Jan 2019 00:00:00 +0100) | 12:38 |
Amnesia | does anyone have a clue what's goin on? | 12:38 |
juliank | So, haproxy and wget are stuck in proposed as they built against pcre2. I patched wget to use pcre3 again | 12:40 |
juliank | Not sure about haproxy | 12:40 |
juliank | Amnesia: What do you see when you open http://archive.ubuntu.com/ubuntu/dists/cosmic/InRelease (wget, curl, or browser, whatever) | 12:40 |
Amnesia | juliank: Date: Thu, 18 Oct 2018 15:17:19 UTC | 12:41 |
juliank | Amnesia: Right, but no valid-until line, right? | 12:41 |
Amnesia | correct | 12:41 |
juliank | Amnesia: so, it works for me in disco, which release are you trying that on? Also, I think you can check in the target directory you specified and look at release files there | 12:43 |
Amnesia | I'm trying it on arch.. | 12:44 |
juliank | Amnesia: ok, that's trickier | 12:44 |
juliank | Amnesia: that's debootstrap 1.0.113? | 12:44 |
Amnesia | yep | 12:44 |
Amnesia | https://git.archlinux.org/svntogit/community.git/commit/trunk?h=packages/debootstrap&id=03584f10a9edb8de54c6417f133ecea5862f4ff3 | 12:45 |
juliank | Amnesia: So, I don't see how this happens | 12:46 |
juliank | Amnesia: I think you can run debootstrap as sh -x <path to>/debootstrap | 12:46 |
juliank | that should give you a verbose log | 12:47 |
Amnesia | yep, going through it atm.. | 12:48 |
juliank | Amnesia: No, I see how this happens | 12:48 |
Amnesia | you do? | 12:48 |
juliank | Amnesia: Passing --no-check-valid-until should make it work | 12:49 |
juliank | The code assumes that Valid-Until is in the release file, which it is not for ubuntu | 12:49 |
Amnesia | hm seems to work indeed | 12:50 |
Amnesia | https://github.com/felixonmars/debootstrap/blob/4f6c6c0ccfd15145bae94b204056be3f78b08670/functions#L632 | 12:51 |
Amnesia | looks like you're right yes | 12:51 |
juliank | That's https://bugs.debian.org/918722 | 12:52 |
ubottu | Debian bug 918722 in debootstrap "debootstrap: says InRelease file expired" [Normal,Open] | 12:52 |
* Amnesia facepalms | 12:53 | |
Amnesia | I should've checked the bug tracker | 12:53 |
Amnesia | sorry for the hassle juliank | 12:53 |
=== Serge is now known as hallyn | ||
seb128 | bdmurray, hey. Could you help with https://launchpad.net/ubuntu/+source/gvfs/1.38.1-0ubuntu1.1 ? It includes 3 fixes, 2 have been verified and the 3rd one isn't working because it's relying on a samba change which we don't have in our version. The non-verified one is basically equivalent of no-change, can we validate the SRU as it? or do you want another upload with that change removed? (but then we need to revalidate the other fix/delay?) | 17:01 |
seb128 | vorlon, how did you manage to get a SRU in without bug reference? ;) (https://launchpad.net/ubuntu/+source/livecd-rootfs/2.542.1) | 17:03 |
seb128 | bdmurray, https://launchpad.net/ubuntu/+source/fontconfig/2.12.6-0ubuntu2.3 should also probably be deleted from bionic-proposed, it failed verification and it was decided that SRU was not needed after all so no fixed version has been uploaded | 17:04 |
vorlon | seb128: well, I didn't accept it, ask whoever did :) | 17:05 |
vorlon | seb128: (livecd-rootfs is a special case, but having an SRU with no linked bugs at all risks foundering on the process) | 17:05 |
seb128 | vorlon, right, unsure if the "who accepted" info is available somewhere? anyway, I mentioned it in case that was an overlook because it looks like the sort of isse that could lead the SRU to be stucked because ti's never marked as green/verified on the report page | 17:15 |
vorlon | seb128: nope, we record that information in comments on the linked bugs ;-) | 17:15 |
seb128 | k, well that usually doesn't work for me :p | 17:16 |
vorlon | well in this case there's no linked bug, hence the ';-)' | 17:16 |
seb128 | right | 17:16 |
seb128 | anyway | 17:16 |
seb128 | thx for the reply :) | 17:16 |
vorlon | bdmurray: ^^ does that happen to have been your accept? | 17:22 |
bdmurray | vorlon: Yeah, it was I who accepted the package. I imagine you or Cody will nag about getting it released. | 18:09 |
vorlon | bdmurray: ok :) | 18:09 |
bdmurray | although I guess I misremembered what you'd added to the SRU page about it. | 18:11 |
bdmurray | I blame Santa for that. | 18:12 |
bdmurray | seb128: fontconfig will require an AA which I am not | 18:15 |
bdmurray | seb128: I'm fine with letting gvfs into cosmic-updates. | 18:20 |
seb128 | bdmurray, thx, I can do fontconfig, I was just unsure if that needed some sort of SRU team ack first | 18:35 |
bdmurray | seb128: Well I'll ack it just in case | 18:37 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!