=== nickv1985_ is now known as nickv1985 [07:24] vorlon, retrying n-m autopkgtests isn't going to work, the log is having those kernels errors that seem the issue/a fallout from the new gcc [07:24] error: ‘-mindirect-branch’ and ‘-fcf-protection’ are not compatible [07:24] (e.g https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-eoan/eoan/i386/n/network-manager/20190708_231002_6fa4a@/log.gz) [07:26] also dkms/nvidia drivers seem broken now, would be good to have some autopkgtest to prevent gcc migrating if it breaks drivers next time maybe? [07:26] doko, ^ [07:29] vorlon, the n-m warnings look like bug #1835764 [07:29] bug 1835764 in virtualbox (Ubuntu) "Multiple DKMS kernel drivers fail to build with gcc-9 [error: ‘-mindirect-branch’ and ‘-fcf-protection’ are not compatible]" [Critical,Confirmed] https://launchpad.net/bugs/1835764 [08:03] seb128: please see the announcement from amurray and contact the security team for the hardening changes [08:03] doko, what announcement de amurray? [08:03] from* [08:04] seb128: https://lists.ubuntu.com/archives/ubuntu-devel/2019-June/040741.html [08:05] amurray, ah, thanks, that's some week old and scrolled out from my first page [08:05] doko, would have been nice to include a reminder to that one in your gcc transition email, maybe a tip for next time :) [08:06] since that only got active with the gcc change [08:06] the best fix for this would be to update the eoan kernel to add -fcf-protection=none I think [08:06] seb128: it *should* have been in gcc-8 already [08:06] amurray, doko, it seems a bit backward to have landed the changes before fixing the kernel [08:06] now we are in a situation were drivers are broken on eoan archive [08:07] nothing even blocked that breakage to limit it to proposed :-/ [08:07] dude, the kernel team is just binary copying kernels to eoan, and yes, they know about that [08:07] seb128: apologies, it slipped off my radar with some embargoed security stuff which just went out the door [08:07] what has how they upload to do with that? [08:08] amurray, no worry, I just wodner if we need some sort of autopkgtest to avoid having $things migrated when dkms gets screwed next time [08:08] doko: oh well that fixes my mental model then - I was wondering why the eoan kernel builds weren't failing as well but if they are being done on disco gcc that would make sense [08:09] * amurray will look at cooking up something for the eoan kernel and will discuss with kernel-team [08:09] ta [08:09] amurray, thx [08:10] * seb128 brb, need to change location for a meeting [08:12] * amurray dinner etc - back later [08:53] cyphermox, hey, what's the status of the grub2 buildfix for eoan? any chance that lands this week? [08:58] bout time [09:17] Hrm, it seems to me when you select a compressor in /etc/initramfs-tools/initramfs.conf, mkinitramfs shouldn't silently bail if the compressor is missing. [09:19] Something like http://paste.openstack.org/show/M6SOlCnADce7eUXS1W4c [09:22] powersj, hey, unsure if you are the right person to ping about that but do you know if daily ISO utah test env is having issue? seems like bionic/eoan test are bailing out on an error 7 [10:24] Could someone review https://code.launchpad.net/~racb/ubuntu-transition-tracker/mysql-8/+merge/369878 please? I've not done this before, and I can't figure out how to test locally (seems that ben can't download from Ubuntu and the go script is a workaround but not suitable for use outside the machine it runs on) [10:34] paride: see seb128's question from about an hour ago please - can you help? ^ [10:36] seb128, oh yes jibel and me were looking at the issue [10:36] seb128, error 7 is defined as a command line parsing error, but it seems that what's actually happen is that a URLChecker method is failing, probably because of networking issues [10:36] and the error is deceptive [10:37] (thanks rbasak for the ping) [11:23] rbasak, paride, thx! do you know what needs to be done? [11:38] seb128, looks like it's back to normal, but let me re-launch another couple of jobs manually [11:39] paride, thx [11:40] paride, I'm curious but what part of the log hinted out that it was an infra issue? just the knowledge of what is behind error7 ? (can you point me to where those are defined?) [11:46] seb128, the log wasn't really useful. We observed a couple of things: (1) nothing was changed on venonat, utah, or the job definitions (2) the build didn't always fail, but they were very flaky (3) utah took a long time to parse its arguments, hinting it was trying to do something over the network (4) there is this URLChecker class which the parser [11:46] uses to check if the URL which are passed to UTAH are valid, but which actually makes things more difficult to debug [11:46] and so that's what we're blaming and why [11:46] paride, k, thx for the details, let's see how the retries goes then! [11:47] seb128, if you want to have a look, `git clone lp:utah`, grep URLChecker and follow it back [11:47] k [11:47] thx! [11:47] yw! [11:50] paride, seems it worked, I got 'back to normal' emails for bionic and eoan now ;-) [11:51] and the pending ISO migrated to be current [11:51] excellent :) === balkamos_ is now known as balkamos [12:52] Could someone review https://code.launchpad.net/~racb/ubuntu-transition-tracker/mysql-8/+merge/369878 please? I've not done this before, and I can't figure out how to test locally (seems that ben can't download from Ubuntu and the go script is a workaround but not suitable for use outside the machine it runs on) === ricab is now known as ricab|lunch [13:37] rbasak: it looks reasonable, I suggest just committing it and seeing what happens [14:03] Laney: thanks! [14:06] seb128: I have the fix staged, I'm finishing up on testing the new grub post-merge [14:07] cyphermox, ok, so likely this week? [14:08] yes, likely [14:08] ups [14:08] cyphermox, great :) [14:09] cyphermox, oh, also what's the status for the plymouth patches upstreaming/new version update from your side? I had it on my june backlog but failed to got to it, I plan to try again this month! [14:13] vorlon: can you please promote xdg-desktop-portals and xdg-desktop-portals-gtk in bionic per bug 1750069 and bug 1749672 [14:13] bug 1750069 in xdg-desktop-portal-gtk (Ubuntu Bionic) "[MIR] xdg-desktop-portal-gtk" [Undecided,Triaged] https://launchpad.net/bugs/1750069 [14:13] bug 1749672 in xdg-desktop-portal (Ubuntu Bionic) "[MIR] xdg-desktop-portal" [High,Triaged] https://launchpad.net/bugs/1749672 [14:14] vorlon: i'm trying to update the seed now, but Laney said i need them promoted first === ricab|lunch is now known as ricab [14:24] seb128: I have not yet had time to look at plymouth this cycle [14:26] cyphermox, do you think you are going to be able to squeeze some time for that in the next weeks? [14:34] cyphermox, I will try to have a go at that update since we would like to see the new version in but let me know if you start poking as well so we don't dup work [14:41] xnox, hello, your dh-cargo merge il blocking a bunch of rust packages... === Foxtrot is now known as foxtrot [14:55] doko: dealing with some gcc9 fallout, [14:55] doko: stuff like this: https://pastebin.ubuntu.com/p/nSBdGN3xy4/ and https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88780 [14:55] gcc.gnu.org bug 88780 in middle-end "[8/9/10 Regression] bogus -Wstringop-truncation for copying as many bytes from a string as its length" [Normal,Assigned] [14:55] doko: googling brought up https://bugs.squid-cache.org/show_bug.cgi?id=4969, is that a thing for gcc9? [14:55] er, switch the bugs, "is that a thing" is the gcc bug [14:56] let me refrase [14:56] doko: stuff like this: https://pastebin.ubuntu.com/p/nSBdGN3xy4/ and https://bugs.squid-cache.org/show_bug.cgi?id=4969 [14:56] doko: googling brought up https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88780, is that a thing for gcc9? [14:56] gcc.gnu.org bug 88780 in middle-end "[8/9/10 Regression] bogus -Wstringop-truncation for copying as many bytes from a string as its length" [Normal,Assigned] [15:00] ahasenack: are you seeing this with GCC 9 only? according to the report it should be in 8 as well [15:01] yeah, the squid build works with gcc8 [15:01] the gcc 9 release notes say that this stringop-truncation check was expanded a bit [15:01] I'm still digging, for example, it only happened with -O2 or higher [15:02] well, -Wno-error=stringop-truncation is your friend ... [15:02] with just -Wall -Werror -Wstringop-truncation it works, but if I add -O2 then it fails [16:05] seb128: hopefully the new kernel in eoan-proposed has fixed this - I'm retrying those nm tests with --all-proposed to see [16:07] kenvandine: bionic only, or both bionic and xenial? [16:08] please promote both [16:08] vorlon: i haven't prepared the seed for xenial yet, but i will be this week [16:09] kenvandine: also, it is normally the case that we add things to seeds /first/ and then change the archive, so that we can use component-mismatches as a guide for what needs promoting; we don't generate c-m reports for !devel, but I'm not sure why it should be promoted before the seed change anyway [16:09] the update script won't find it [16:09] update script? [16:09] Laney said we need to promote it first [16:09] in ubuntu-meta [16:09] oh, you do need to promote it before reuploading ubuntu-meta, yes [16:09] yes [16:10] but the ordering is (normally) seed change -> archive promotion -> ubuntu-meta upload [16:10] right, for devel [16:19] vorlon, thx [16:24] kenvandine: the xdg-desktop-portal dependencies look a bit strange on xenial; Depends: default-dbus-session-bus | dbus-session-bus | dbus-x11 but only the third of these exists in the release [16:24] yeah [16:25] there are other packages in xenial that followed that pattern as well [16:25] so the packaging matched debian, for example [16:25] mm [16:26] i can change it [16:26] anyway, promotions done on both series [16:26] thanks [18:08] bdmurray: Just replied to your question at bug #1834406. Any other idea how to reach out to CJK users willing to help out? [18:08] bug 1834406 in fonts-noto-cjk (Debian) "Upgrade Noto Sans CJK fonts to version 2.001" [Unknown,New] https://launchpad.net/bugs/1834406 [18:12] GunnarHj: I think the ubuntu-quality list used to be used for calls for testing but I'm not sure how active that list is now [18:17] bdmurray: I don't know either, but that's an idea. Do you think the statements so far are insufficient, so I should try that way? (Personally, and unlike my last dubious SRU proposal, I feel this is a safe one.) [18:18] GunnarHj: I was following along with sil2100's comment. [18:18] https://bugs.launchpad.net/ubuntu/+source/fonts-noto-cjk/+bug/1834406/comments/4 [18:18] Launchpad bug 1834406 in fonts-noto-cjk (Debian) "Upgrade Noto Sans CJK fonts to version 2.001" [Unknown,New] [18:19] bdmurray: Right, and that's the reason for the efforts so far. [18:22] GunnarHj: Okay, I'd asked about the call for testing because I wasn't sure what calls had been made. [18:24] GunnarHj: Either way let's give it a bit more time to age though. [18:27] bdmurray: Sure, no problem. The new era started May 1, so it's late anyway.. TBH I would have expected some Japanese users to act as drivers to have it tested, so I'm a bit disappointed. [18:30] bdmurray: Maybe it would help if you added a comment to the bug about the decision to wait a bit due to few test reports? [22:25] LocutusOfBorg: Ah, you already commented on the opensips report, was just about to do that now. Very well then.