=== maclin1 is now known as maclin [10:12] How do I go about having an obsolete package removed from the Yakkety archive? [10:13] file a bug and subscribe ~ubuntu-archive [10:13] Thanks. [14:13] bdmurray: hi. looking at the gnome-software error logs, i'm not sure that those errors are regressions. they just seem to be binned separately from https://errors.ubuntu.com/problem/77807d3502a3e74469d5b5dca0a34e9197c647d6 [14:25] attente: I'm sorry which crash is binned separately from 77807d35... [14:26] bdmurray: this one https://errors.ubuntu.com/problem/dc83afe90b31d0984940db9de2ff06b2c6ee8b78 [14:26] which seems to be the reason why it looks like a new regression with increased rate of occurrence [14:29] attente: okay, thanks for digging. I'll take care of it now. [14:30] bdmurray: thanks again [14:31] attente: what about https://errors.ubuntu.com/problem/111ba23deb7463e9f1203f30099e0c2ca6c2af80? [14:33] bdmurray: hadn't seen that one. does look like a regression though [14:37] attente, do you say that from code changes making it likely? or just because it was not reported before? [14:39] just because it wasn't reported before [14:44] attente, there is a similar 3.18 report on the redhat bugzilla, I think it's not a regression, just not a common issue [14:44] ok, i guess you are right. that upstream bug report doesn't seem to have any resolution === dgadomski_ is now known as dgadomski [14:51] no, I tried to play with nm status changes during install and co, but can't trigger it [14:51] anyway I don't see any sign of an obvious regressions in those reports [14:53] but e.u.c retracings are unreliables atm so difficult to say, if 2/3 of the weekly view was not missing symbols we would have more datas on what has already been reported [14:53] bdmurray, can you unblock the SRU? we believe there are no reason to block it due to e.u.c reports [15:01] seb128: will do [15:01] bdmurray, thanks [15:01] attente, ^ [15:01] attente, thanks for following up on the issues btw! [15:03] seb128: i was just the messenger here, thanks for finding that upstream report :) [15:03] attente, yw, good team work! [16:50] kirkland: Do you care that your changelog format is ugly (extra newline)? [16:50] kirkland: http://launchpadlibrarian.net/258203000/pollinate_4.15-0ubuntu1_4.15-0ubuntu1.2.diff.gz [16:50] kirkland: Also, never upload to series-updates. :P [16:51] uploads to xenial -> are redirected by launchpad into the right place already. [16:51] I'm well aware. [16:51] uploads to updates are not, however, rewritten. [16:52] infinity, that's not for you =) but kirkland. [16:52] kirkland: But yes, you're living 5 years in the past. If you always just upload to "$series", LP does the right thing. [16:58] sgclark: Talk to me about kde-l10n-* [16:58] sgclark: In the past, Scott/Riddell did the SRU admin side (I'm happy to take that over), but are you still following whatever previous practise was for building and testing them? [16:59] infinity: hi, when they were sponsered for me they missed a chunk, which randomly break installs across the world, no pun intended [16:59] sgclark: Sure, I'm not questioning the necessity of the uploads. We do them for point releases anyway, historically. [16:59] they were heavily tested, it was the upload to archive at the end that just missed a chunk [17:00] sgclark: More questioning how people plan to test and promote, etc. [17:01] We have several bugs with willing testers? Sorry I am not sure what the question is [17:02] sgclark: Well, updating *all* the lang stuff tends to imply that either you're only spot-checking one or two, or you have an automated test plan to make sure they don't suck. [17:02] sgclark: I don't know how that was handled before, but scott/riddell did it without me intervening, so I didn't need to know. [17:03] Yeah and Jonathan took all his knowledge with him. Scott still helps me out when he can (very busy) and yofel has much of the knowledge. [17:03] yofel uploaded the last languages correctly afaik, he is also very busy [17:04] Indeed, I'm not questioning the uploads, I'm questioning how we get them all tested so they can move from proposed to updates. [17:05] right. so if I conjure up testers, what do I need to have them do? I do not use language pack to be perfectly honest [17:07] sgclark: I'm not sure manual testing is the right answer, unless you can find 60+ testers who all speak the right languages. [17:07] I see yoour point [17:07] sgclark: And maybe the answer is "in the past, we never actually tested, but just spot-checked a few languages". [17:07] sgclark: But I'd like to know what Scott was doing in the past, so we can either emulate or improve, but not regress. :P [17:08] infinity: ok I will ping him and ask. [17:08] Ta. [19:27] infinity: xnox: :-) thanks dudes. [19:28] * xnox did nothing =) [19:36] Who accepted what-utils? [19:37] kirkland: You didn't address my reject message on that upload (s,ports.ubuntu.com,ports.ubuntu.com/ubuntu-ports,) [19:38] infinity: hmm [19:38] infinity: it seems to work okay, without [19:38] kirkland: Both work, technically, but the symlink from ubuntu-ports -> isn't necessary for mirrors, so it has the potential to be wrong. [19:38] ⟫ how-many-binary -a s390x [19:38] 50855 [19:38] interesting [19:38] "ubuntu-ports -> ." that is. [19:38] fixing now. [19:39] (IMO, it should just be /ubuntu/ to be consistent, but yay history) [19:39] Maybe we should fix that some day. [19:40] kirkland: That said, why does it use the archive at all? That seems amazingly slow. [19:41] infinity: yes, requires network access [19:41] https://launchpad.net/ubuntu/xenial/s390x/ -> 67302 binary packages [19:41] kirkland: No, I mean, using the archive implies downloading Packages and parsing.. [19:42] Not that I've read the source. [19:42] infinity: download package list, and count [19:42] Yeah, could be a single API call. [19:43] infinity: hrmm, well... [19:43] infinity: i'm curious about the discrepancy [19:44] https://launchpad.net/ubuntu/xenial/s390x/ says 67302 [19:44] kirkland: Also, to avoid lpapi clients: wget -O- https://api.launchpad.net/devel/ubuntu/xenial/s390x/package_count 2>/dev/null [19:44] kirkland: The count would be the difference between all binaries in all pockets and whatever you're parsing in your script, I imagine. [19:44] This page contains the following errors: [19:44] error on line 1 at column 1: Document is empty [19:44] Below is a rendering of the page up to the first error. [19:45] kirkland: Yeah, hence wget. ;) [19:45] wget -O- https://api.launchpad.net/devel/ubuntu/xenial/s390x/package_count 2>/dev/null && echo -en "\n" [19:45] 67302 [19:45] (base)adconrad@nosferatu:~$ [19:45] yeah [19:45] so, now, I'd like to see that list of 67K packages [19:46] ie, how-many-binary --verbose [19:48] for i in main universe restricted multiverse; do [19:50] You're at least missing a "for release in $release{,-updates,-proposed,-backports}; do" after that. [19:50] Which will get your list higher.