[09:42] sil2100: Thanks for the bionic upload for bug #1778041. Want to mention that it resulted in much fewer binaries than is present in bionic-release. Is that a problem? (I hope not.) [09:42] bug 1778041 in freshplayerplugin (Ubuntu Xenial) "browser-plugin-freshplayer-pepperflash broken" [High,In progress] https://launchpad.net/bugs/1778041 [09:46] GunnarHj: I don't think so, sometimes microreleases just drop binary packages so yeah, but I must say I didn't have experience with such cases before - accepted it as it seemed the logical way to go [09:46] Since it's either that or leaving it busted [10:03] sil2100: That's what I thought/hoped. There is an additional aspect with the xenial one - there is already stuff the old way in bionic-updates. [10:04] sil2100: s/bionic-updates/xenial-updates/ [10:28] sil2100, is it a good time for a review of https://github.com/CanonicalLtd/ubuntu-image/pull/155 ? === realitix_ is now known as realitix [13:01] Greetings. Can anyone help? Trying to update a chroot'd 18.04 base image on an 18.04 host. `Get:3 http://ports.ubuntu.com/ubuntu-ports xenial-backports InRelease [107 kB] Err:1 http://ports.ubuntu.com/ubuntu-ports xenial InRelease Couldn't create temporary file /tmp/apt.conf.vN6blL for passing config to apt-key` [13:02] filesystem is writable. `ls -als /tmp total 12 4 drwxr-xr-x 2 root root 4096 Aug 13 12:55 . 4 drwxr-xr-x 22 root root 4096 Aug 13 12:38 .. 4 -rw-r--r-- 1 root root 6 Aug 13 12:57 x` [13:02] g40: shouldn't /tmp be 1777? [13:04] @ahasenack boom! great catch, thank you. [13:04] Error: "ahasenack" is not a valid command. [13:40] mvo, any questions about command-not-found? [13:40] scientes: in a meeting right now, sorry. but excited about your C implementaton, I think its pretty cool to have that [13:41] hit me up [13:41] (later) [13:46] scientes: now :) so I think the C implementation is great, low latency/overhead and all that. the update-alternatives problem could be solved by making it a server side problem. I had this idea that we would store the commands in a new apt indexfile that is automatically downloaded by apt when c-n-f is installed. the apt snippet is super simple:https://paste.ubuntu.com/p/TW4gR7KnmF/ and here is an example what it would look like http://peop [13:46] le.canonical.com/~mvo/c-n-f/dists/bionic/main/cnf/ [13:46] scientes: this would then be parsed by c-n-f into whatever format it wants [13:47] scientes: the server side extractor is also easyhttps://git.launchpad.net/~mvo/command-not-found-extractor/tree/ [13:47] scientes: (unfortuantely I have some trouble getting the server side deployed but that is a different matter) [13:47] scientes: what do you think about this? [13:48] scientes: ideally this would be same for debian/ubuntu too, the format is pretty simple [13:53] yeah that is great [13:56] I'll code that up in the C version [13:58] scientes: thanks, lets coordinate with the proposal of the ideas to debian and please keep naging me about getting the extractor to the server side, it needs some admin work mostly I think [14:01] is gcc-8 the default one in cosmic? [14:02] ahasenack, $ ls -al /usr/bin/gcc [14:02] lrwxrwxrwx 1 root root 5 Jul 21 00:14 /usr/bin/gcc -> gcc-8 [14:02] trying to figure out this error in a ppc64el build: cc1plus: error: unrecognized command line option ‘-Wno-deprecated-register’ [-Werror] [14:02] same package builds in the other arches [14:02] well, is still building. so far amd64, s390x and i386 passed [14:03] thanks scientes === Eleventh_Doctor is now known as Pharaoh_Atem [14:04] is it known and expected that changing the upgrade-manager release-upgrade prompt to 'normal' enables starting an upgrade from 16.04 to 18.04 without -d ? (and apparently that fails? [14:05] that should always work even without -d [14:12] scientes: 16.04 -> 18.04 LTS upgrade is not enabled, yet, no. [14:17] oh my bad [15:34] seb128, based on our last week discussion, dgadomski_ did the MP for both LPs (dgadomski and tseliot) : https://code.launchpad.net/~dgadomski/ubuntu/+source/gdm3/+git/lp1782152/+merge/352976, it is set to "Need review". According to your process, should I wait for that MP to be approved/merged before uploading ? [16:06] seb128, basically what are the desktop team expectations according to your process before tseliot or I can upload gdm3 for Bionic (both LPs) which has a MP "Needs review" and Xenial (dariusz bug only) [16:38] hi, on a ppc64el lxd container (haven't tried elsewhere), dpkg-buildflags in cosmic returns -O3, instead of -O2 [16:39] I'm trying to track down where that is coming from [16:39] /etc/dpkg/* has nothing [16:39] any other ideas? [16:39] dpkg-buildflags --status output: https://pastebin.ubuntu.com/p/WXPwsbT5H9/ [16:40] in a lxd container in my laptop, amd64, I get -O2 in cosmic [16:41] ah [16:41] /usr/share/perl5/Dpkg/Vendor/Ubuntu.pm [16:41] if arch ppc64el, set -O3 [16:41] amazing [16:41] rbasak: ^ FYI [16:42] "-g -O3" actually [16:43] Yes, it was a specific requirement from the port's sponsor AIUI [17:06] hi guys, any idea why adding -O3 would enable -Werror=format=truncation? [17:07] https://gist.github.com/panlinux/4716e167c2e06612b28be4b9f8f2b52b scroll all the way to the right, the only difference between the two g++ command lines is the last -On I added [17:07] -O3 fails, -O2 works [17:07] (it was easier than editing -O2/-O3 inside the long command line, but has the same result) [17:10] slashd, no, it's fine, go ahead with the upload and I'm going to review/merge it [17:13] ahasenack: the GCC manual says that -Wformat-truncation sometimes estimates the number of bytes that will be written based on optimisation-level-dependent heuristics [17:13] "While enabling optimization will in most cases improve the accuracy of the warning, it may also result in false positives." [17:14] hm [17:14] You'll need to analyse the error manually and see whether it's a true overflow [17:15] I could try -O2 *with* -Wformat-truncation [17:15] If it's a false positive, then sometimes carefully-chosen assert() calls are enough to hint the compiler into what's going on [17:15] if that triggers again, then it's not the optimize, right [17:15] No, you misunderstand [17:16] -Wformat-truncation is enabled by default, but behaves differently depending on the optimisation level, because that influences the heuristics that it uses to analyse how many bytes will be written [17:17] Before going any further, you should analyse the code path that the compiler is complaining about and see whether it's actually a potential overflow [17:18] If it is, then fix it; if it isn't, then maybe there's an assert you can add to help the compiler prove that it isn't [17:26] ahasenack: I'm not sure what the specific issue is that you're working on, but it might be acceptable to override it down to -O2 for the time being and leave a bug open if you're trying to land something. [17:26] rbasak: it's the squid build I mentioned in the standup [17:26] that it only failed in ppc64, and at that time I saw that -O3 was used, and I thought it was something new in cosmic, not that it was ppc64 exclusive [17:27] Does it fail with -O3 on amd64? [17:28] it should, let me check [17:34] -O3 ppc64el reminds me of https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=905868 (probably just coincidence) [17:34] cjwatson: as far as I can see, the truncation can indeed happen. Target buffer is 64 bytes, source string can be up to 3*256+10 [17:34] Debian bug 905868 in gcc-8 "gcc-8: miscompiles vec_sl at -O3 with -fwrapv on ppc64el" [Normal,Open] [17:35] upstream seems aware: http://squid-web-proxy-cache.1019090.n4.nabble.com/compiler-Error-tp4686058p4686065.html [17:35] this error is a bit down in that page [17:35] I also got the first one, but that one was easy to fix [17:39] heh, look at what I found: https://bugs.squid-cache.org/show_bug.cgi?id=4875 [17:40] with a pr and a work-in-progress patch [17:40] goo [17:40] d [18:19] !dmb-ping [18:19] cyphermox, jbicha, micahg, rbasak, sil2100, slashd, tsimonq2: DMB ping. [18:19] 40 mins, but I'd really be happy to have quorum. ;) [18:19] tsimonq2, I'll be there [18:19] Thanks! [19:01] !dmb-ping For real this time [19:01] tsimonq2: I am only a bot, please don't think I'm intelligent :) [19:01] darnit [19:01] !dmb-ping | For real this time [19:01] For real this time: cyphermox, jbicha, micahg, rbasak, sil2100, slashd, tsimonq2: DMB ping. [19:01] I hate how ubottu is inconsistent in what it wants you to pipe and what it doesn't. [19:02] It isn't, you pipe !info as well as factoids. === grumble is now known as Guest76306 === gurmble is now known as grumble [19:11] Unit193: Oh? [19:11] Hmm. === simonquigley is now known as tsimonq2 === DrKranz is now known as DktrKranz === tomreyn_ is now known as tomreyn === geofft_ is now known as geofft === tedg_ is now known as ted__ === ted__ is now known as tedg__ === tedg__ is now known as tedg [20:33] infinity: So, I'm now a Core Developer: https://lists.ubuntu.com/archives/ubuntu-devel/2018-August/040436.html [20:34] infinity: I think you said you wanted to push a Core Developer-only DMB policy at one point, after I became a Core Developer. [20:34] infinity: So, there you go. :) [21:04] tsimonq2: \o/ [21:04] tsimonq2: Congrats. [21:04] infinity: Thanks!