[00:37] slangasek: Could you axe kdesudo? bug 1757682 [00:37] bug 1757682 in kdesudo (Ubuntu) "Please port your package away from Qt 4" [Medium,Confirmed] https://launchpad.net/bugs/1757682 [00:59] -queuebot:#ubuntu-release- New source: grub-legacy-ec2 (bionic-proposed/primary) [1:1] [01:25] slangasek: https://bugs.launchpad.net/ubuntu/+source/plymouth/+bug/1592405/comments/1 Do you happen to remember what version clobbered it? [01:25] Ubuntu bug 1592405 in plymouth (Ubuntu Bionic) "plymouth hook in initramfs needs font but doesn't Depend on it" [High,Triaged] [01:35] tsimonq2, it may have been the merge I have done [01:36] tsimonq2, to get the new upstream release in. [01:36] tsimonq2, should i poke and try to look into this too? [01:36] xnox: You're more than welcome to; I'd certainly like to learn from this but if it's a quick solution, by all means, JFDI. :) [01:41] xnox: Hum, might it be that debian/initramfs-tools/hooks/plymouth is missing? [01:42] tsimonq2, well, i'm trying to dig history, i'm pretty sure I was the one who _introduced_ using UbuntuMono font.... [01:43] xnox: Heh, OK. [01:52] tsimonq2, now I am confused. I think, this may only be using ubuntu font, post initramfs, and never in the initramfs. [01:52] tsimonq2, and I do not see any attempts to have UbuntuFont in the initramfs. [01:53] tsimonq2, but I do recall proposing to do that circa 2013/2014, but possibly there were concerns about "too large" file size. [01:53] xnox: I think it could be readdressed. [01:54] xnox: I wonder what the filesize difference might be. [01:55] Ubuntu regular is 357k, we may want the Ubuntu Mono regular which is 209k [01:56] And right now, DejaVu is used? [01:56] I mean, for a 209k difference, I personally think it's a good idea. [01:56] DeajuSerif + Sans are 381k + 758k [01:56] Ok, so if we *just* use Ubuntu Mono, then that's less size? [01:56] Or am I not seeing this right? [01:57] supports less languages, but i'm not sure we support translations in the initramfs [01:57] and we need to customize fonconfig conf file, in the initramfs [01:57] but easy enough to special case that. [01:57] cause, Ubuntu font is not "metric" compatible with Arial, unlike DejaVuSans [01:57] OK [01:57] then again... this is plymouth and plymouth-text, not thesis typesetting [01:58] Hehe, right. [01:58] xnox: So is this something that you could JFDI pretty easily? [01:58] tsimonq2, do you know if derivates plymouth themes also specify and use 'Ubuntu 11'? or just the Ubuntu one? [01:58] aka lubuntu-logo, etc? [01:58] * tsimonq2 looks at what Lubuntu does [01:59] just need to write a "better" /etc/fonts/conf.d/60-latin.conf [01:59] to be all ubuntu font family [02:01] xnox: Yeah no, this just specifies a module of "ubuntu-text" [02:01] (in Lubuntu's) [02:01] lubuntu-logo should be the one splashy [02:01] ubuntu-text, is the "serial non-quiet non-splash" version [02:02] plymouth-theme-lubuntu-logo or like plymouth-theme-lubuntu-next-logo [02:02] Right. [02:02] But we also have just text packages. [02:03] yeah, for text-only we don't do anything - no fancy fonts, no pango, no truetype, because we asume the user wants just the kernel font rendering. [02:03] because the user wants it shmall [02:04] tsimonq2, so it looks like at least lubuntu-font got forked, before 'Ubuntu 11' was added in the "ubuntu-logo" theme. [02:04] so does not use that. [02:04] A quick fix, is to stop using 'Ubuntu 11' in the ubuntu-logo theme =) [02:05] or ship the ubuntu font, and force it to be used by default, and then there would be no need for it to be specified. [02:05] If it doesn't use too much resources, I'd argue for the latter. [02:06] I mean, if it introduces little to no size, even in cases where the user wants it small, if it's just 1/5 of a MB, it shouldn't be a big deal. [02:06] when kernel modules are huge. [02:06] Right. [02:06] The kernel is kinda huge. :P [02:06] (In comparison.) [02:07] xnox: Oh, jbicha made a good point on the bug report. [02:07] tsimonq2, this will need testing. [02:07] hm? [02:07] tsimonq2, yes, i did notice the transitional package dep already [02:07] OK, cool. [02:08] xnox: Would you like to take care of this (because I assume you know the codebase better) or should I? [02:08] I'm fine either way. [02:08] And indeed, testing will be needed. [02:08] the transitional font dep has been annoying me for a while, but I was not annoyed enough to touch plymouth 🙈 [02:09] This might be a good occasion. :) [02:10] jbicha, hahahhahaha [02:10] tsimonq2, i'll do it all; but will need to do boot tests; so not right now, but soon. [02:11] xnox: Alright; I'll assign the bug to you. Thanks, and please keep me updated. :) [02:12] Er, you did already. Cool. [04:38] xnox: ah; I guess I was wrong about the initramfs hook having used the Ubuntu font previously? I just checked the last version before the Debian merge (0.9.0-0ubuntu9 in vivid) had a dep on both fonts-dejavu-core and ttf-ubuntu-font-family, and only copied fonts-dejavu-core into the initramfs [04:39] xnox: so, it would be nice to get it using the Ubuntu font consistently (and dropping the need to pull dejavu in in this context), but re-adding the dejavu dep would apparently suffice to fix the regression [04:58] slangasek: Would you happen to know what's going on with armhf autopkgtest builders? [04:58] (I'm asking you because I hope you saw that the build you've triggered is also queued but not running. :P) [04:59] tsimonq2: nope, haven't looked; looking now [04:59] ack [05:06] lol @ someone uploading zfs-linux to Debian that was likely meant for Ubuntu [05:06] * tsimonq2 shrugs, it happens I guess [05:06] tsimonq2: armhf sorted; looks like the lxd runners aren't entirely happy with the daily maintenance job, they likely would've started right about now on their own if I had done nothing [05:07] but that means they were down for 2 hours, unhelpfully [05:07] slangasek: ah ok [05:07] cool, thanks [05:18] slangasek: It'd be good to get a second opinion on node-cross-spawn [05:19] Its tests fail with the new nodejs, but the version we ship doesn't have upstream tests for nodejs 8.x. [05:20] We also ship a release that's a major version behind. [05:21] Ultimately the arm{hf,64} test failures are just timeouts. [05:22] I've been trying to set the timeout integers to be larger values in my PPA testing but have ultimately been unsuccessful so far. [05:24] I'll try some more, but if the situation doesn't improve, then I'm not sure what to do from there. [05:27] In fact, that sort of thing seems almost commonplace; we're shipping node modules with out-of-date versions which have failing tests that have been completely revamped upstream. [05:28] (I'm making a generalization, but still.) [07:22] tsimonq2: there are enough nodejs-triggered autopkgtests that fail only on arm64 and armhf that it makes it suspicious that there may be arch-specific regressions in our nodejs [07:23] tsimonq2: node-block-stream, node-commander, node-cross-spawn, node-liftoff - are these related? [07:31] slangasek: I haven't taken a close look but it's possible. [07:35] slangasek: node-liftoff seems to need timeout bumps like node-cross-spawn. [07:36] Doing, although it's a bit of a PITA because turnaround time for testing is 20 minutes minimum, and these tests stop as soon as there's one error. Sigh. [07:38] slangasek: I don't quite know what to make of node-block-stream. [07:38] slangasek: node-cross-spawn *should* be fixed now. [07:39] slangasek: And node-commander is kind of weird: AssertionError: expected '' to be 'SIGHUP\n' [07:39] slangasek: So while these are unrelated, it's weird that these only happen on *some* arches. [08:17] -queuebot:#ubuntu-release- New: accepted gcc-defaults-ports [amd64] (bionic-proposed) [1.175ubuntu1] [08:17] -queuebot:#ubuntu-release- New: accepted ubuntu-wallpapers [amd64] (bionic-proposed) [18.04.0-0ubuntu1] [08:17] -queuebot:#ubuntu-release- New: accepted virtualbox-hwe [i386] (bionic-proposed) [5.2.8-dfsg-5ubuntu18.04.1] [08:17] -queuebot:#ubuntu-release- New: accepted gcc-defaults-ports [i386] (bionic-proposed) [1.175ubuntu1] [08:17] -queuebot:#ubuntu-release- New: accepted virtualbox-hwe [amd64] (bionic-proposed) [5.2.8-dfsg-5ubuntu18.04.1] [08:17] -queuebot:#ubuntu-release- New: accepted mir [amd64] (bionic-proposed) [0.31.0.1-0ubuntu1] [08:17] -queuebot:#ubuntu-release- New: accepted mir [armhf] (bionic-proposed) [0.31.0.1-0ubuntu1] [08:17] -queuebot:#ubuntu-release- New: accepted mir [ppc64el] (bionic-proposed) [0.31.0.1-0ubuntu1] [08:17] -queuebot:#ubuntu-release- New: accepted mir [arm64] (bionic-proposed) [0.31.0.1-0ubuntu1] [08:17] -queuebot:#ubuntu-release- New: accepted mir [s390x] (bionic-proposed) [0.31.0.1-0ubuntu1] [08:17] -queuebot:#ubuntu-release- New: accepted mir [i386] (bionic-proposed) [0.31.0.1-0ubuntu1] [09:13] slangasek: Actually, note-liftoff can be badtested. [12:04] -queuebot:#ubuntu-release- Unapproved: snapd (xenial-proposed/main) [2.31.2 => 2.32] (desktop-core, ubuntu-server) [12:14] ginggs: ping. bcolz 1.2.0+ds1-1 now ready to be synced from debian/incoming. should fix the autopkgtest failure of c-blosc 1.14.2+ds1-1. [12:15] ahem s/of c-blosc/with c-blosc/ [12:15] estan: waiting for it to be picked up by launchpad https://launchpad.net/debian/+source/bcolz [12:15] ginggs: aha. [14:21] -queuebot:#ubuntu-release- New sync: deepin-movie-reborn (bionic-proposed/primary) [3.2.3-2] [16:43] ginggs: looks like the eagle landed ^ :) [18:24] hello jbicha, I'm uploading a gtk+2.0 merge, because the current one is FTBFS in release pocket, probably your glib2 upload broke it? [18:25] LocutusOfBorg: sync meson [18:25] I mean unless I need yet another FFe for that [18:26] the Ubuntu diff can be dropped now [18:27] estan: infinity beat me to it [18:27] LocutusOfBorg: I mean you could merge glib2.0 if you want too… [18:29] LocutusOfBorg: I'm all confused. Let's start over. Do you have a build log for the gtk2 build problem? [18:29] ok glib2.0 done [18:29] jbicha, https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa/+packages [18:30] you synced gtk2, why? [18:30] this is a no change rebuild of the version in release [18:30] jbicha, I didn't :) or better, I did just to avoid uploading the source tarball for the diff [18:30] I have a connection that sucks a bit, I didn't make the sync build at all [18:33] jbicha, do you see the symbols that have been dropped in the merge? http://launchpadlibrarian.net/361892471/gtk+2.0_2.24.32-1_2.24.32-1ubuntu1.diff.gz [18:33] should I readd them in your opinion? g_cclosure_marshal_VOID__BOXED [18:33] they are removed in the last patch that ubuntu applies [18:33] they are coming from glib, and they shouldn't be there... [18:33] not sure if this is a problem or not [18:42] somebody from desktop team today told me to just remove them [18:43] because such symbols are inside glib2.0, and they shouldn't be in gtk2 (and gtk2 links glib2, so there is no runtime issue) [18:45] yes, they can be removed as I think that was a glib bug [18:46] y'all probably shouldn't have changed gtk_print_backend_set_password@Base 2.24.25-0ubuntu2 but since a newer version is in xenial maybe it doesn't matter so much [18:46] please push your changes to bzr lp:~ubuntu-desktop/gtk/ubuntu [18:47] I believe we'll be switching to git for Chaotic Camel ( web link for bzr is https://code.launchpad.net/~ubuntu-desktop/gtk/ubuntu ) [18:48] if y'all are really eager, you can merge gtk3 too :) [18:51] "y'all probably shouldn't have changed gtk_print_backend_set_password@Base 2.24.25-0ubuntu2 but since a newer version is in xenial maybe it doesn't matter so much" [18:51] this is *exactly* what I changed, but after I changed it back, because "better safe than sorry" [18:51] sure, I'll push them [18:52] that symbol version was changed in the version pushed to bionic [18:52] I know, but maybe reverse-deps complains if we don't rebuild? [18:52] hmm? [18:53] I mean, ok the lower bound is already satisfied [18:55] I'll probably have to do a gtk2 upload in unstable to drop the extra glib symbols [18:56] thanks jbicha :) [18:56] btw, I sync'd meson, just bugfixes, new tests and a bug I was hitting with libinput [18:57] thanks [19:04] ginggs: you guys have been my dream team through this :) [19:18] LocutusOfBorg: gtk2 builds fine in unstable. I guess I could mark those closure symbols as optional… not really sure why Debian & Ubuntu are different there now [20:09] tsimonq2: what's the rationale for considering node-liftoff badtest? [20:38] LocutusOfBorg: gtk2 is one of the last Debian GNOME packages to use cdbs, I wonder if that helps explain the g_cclosure symbols? none of our other libraries have those symbols [20:43] LocutusOfBorg: https://launchpad.net/ubuntu/+source/meson/0.45.1-1/+build/14489818 (it was a binary upload in Debian :| ) [22:39] -queuebot:#ubuntu-release- New binary: javatools [amd64] (bionic-proposed/universe) [0.63] (no packageset) [22:50] jbicha, gtk2, the latest patch (ubuntu specific) is changing the file "./gtk/gtkmarshalers.list" [22:50] so removing them manually I would say [22:50] Debian has not this patch, so they appears there? [22:52] ok, the use-secrets-service patch? [22:56] yes, this: +-VOID:POINTER,POINTER,POINTER,POINTER,STRING [22:56] btw, doko any idea if you are building libphobos.a without fPIC? [22:56] /usr/bin/ld: /usr/lib/gcc/x86_64-linux-gnu/8/libgphobos.a(stdio.o): relocation R_X86_64_TPOFF32 against symbol `_D3std5stdio10readlnImplFPOS4core4stdc5stdio8_IO_FILEKAawE3std5stdio4File11OrientationZ1nm' can not be used when making a shared object; recompile with -fPIC [22:57] this happens with meson [23:04] LocutusOfBorg: we have a packaging branch for glib2.0 if you'd be so kind as to push there please [23:05] link please? I'll be happy to push [23:06] also, gtk3 uploaded [23:06] LocutusOfBorg: it's in the VCS fields in debian/control :) [23:08] ok, lets do it tomorrow :) [23:08] too late here [23:18] I think I did also glib2.0, please don't shoot at me if I did import it wrong :) [23:18] * LocutusOfBorg goes to sleep