=== maclin1 is now known as maclin [08:19] hello doko I'm not sure if openjdk8 in debian/ubuntu armhf is broken or not [08:19] Error: missing `server' JVM at `/usr/lib/jvm/java-8-openjdk-armhf/jre/lib/arm/server/libjvm.so'. [08:19] Please install or use the JRE or JDK that contains these missing components. [08:19] it is not even installable :( [08:19] https://buildd.debian.org/status/logs.php?pkg=gdal&arch=armhf [08:53] what's a $ prefixed rule in a makefile? "$./KeyboardNames.pl:" [08:58] and, well, how do I invoke it? :-) [09:15] LocutusOfBorg: I'll have a look [09:15] tjaalton: https://launchpad.net/ubuntu/+source/apitrace/7.1+git20170623.d38a69d6+repack-1build1 triggered by new glibc [09:15] infinity: ^^^ [09:17] thanks doko, ubuntu seems to be not affected, if this helps [09:17] tjaalton: mesa b-d on llvm-5? are you aware that it ftbfs? [09:19] I'm seriously thinking about using gcc-6 for llvm and arm64/ppc64el [09:41] xnox: What happened to apt 1.4.6~16.10.1 in zesty-proposed? Your last comment was: "And will redo zesty upgrade test again to make sure units are started for the upgrade timer." [09:42] 1.4.7 entered stretch a few weeks ago, I'd like to get this over into zesty eventually :) [10:44] jbicha, didrocks: there are glib2.0 autopkg test failures on ppc64el and s390x triggerd by glibc and elfutils [10:44] infinity, xnox: ^^^ [10:45] doko: yep, but rc2 is available [10:46] tjaalton: sure, if you can fix bugs in this version ... [10:48] Specifically 1691908 would be awesome to fix. [11:20] doko: apitrace has a bug for the ftbfs on debian too, i'll have a look [11:24] though it was due to gcc7 I think [11:43] hi cyphermox , Laney - I saw that the latest upload of ubiquity in artful is having a little bit of trouble building. Just wondering - any chance of my PR sneaking in with the next upload please? https://code.launchpad.net/~ubuntubudgie-dev/ubiquity/fix_lp_1713662/+merge/329911 [11:46] Probably. And I know. [11:46] I got 5 emails about it.... [11:46] No, 6 [12:42] hey doko, re bug #1638695, have you been doing any benchmarks of the artful builds after your changes? [12:42] bug 1638695 in python2.7 (Ubuntu Xenial) "Python 2.7.12 performance regression" [High,Confirmed] https://launchpad.net/bugs/1638695 [12:42] dgadomski: no [12:43] doko: ok, I'll check it out and share the results in the bug [12:46] dgadomski: sure, but remember that pie is turned on by default in artful [12:46] doko: ack, thanks [13:00] infinity, dpkg -S xlocale.h -> libc6-dev:s390x: /usr/include/xlocale.h is not true with new glibc in proposed. Is that intentional, or accidental? [13:01] infinity, autoconf, in prelude, fails to find perl.h because conftest includes xlocale.h [13:01] infinity, not sure if i'm supposed to fix autoconf/perl/prelude or if libc6-dev should continue to ship xlocale.h [13:04] infinity, doko - found changelog saying "Don't include xlocale.h. If necessary, include locale.h instead." hence intentional. [13:07] i think we need perl rebuild, to drop #define I_XLOCALE from its config.h [13:07] testing. [13:12] xnox, sync perl instead [13:13] LocutusOfBorg, right that could do it too. But will wait for my local test build to finish, to make sure that a rebuild of perl will resolve my issue [13:13] with perl we shoudl also sync... let me check [13:14] rename [13:14] huh? rename what? [13:14] I was going to sync, but autopkgtests are already fully [13:14] src:rename [13:14] if we have to rebuild, better sync instead [13:45] Laney: I'll fix ubiquity, unless you already started.. [13:46] rbasak, jamespage: there was no tomcat8 merge for a while ... [13:55] cyphermox: sure did [13:56] yeah I just noticed [13:57] pull, I merged fossfreedom's code [13:57] doko, I would say tomcat8 needs a sync, not a merge [13:57] LocutusOfBorg: no [13:58] tomcat8.5 breaks dogtag-pki [13:58] and tomcatjss [13:58] I mean, tomcat8 "can" be syncd, because the delta is useless [13:58] not "we should sync it right now" :) [13:59] you'd just get it stuck in proposed because of failing freeipa-server autopkgtest [13:59] so what's the point of syncing it then? [13:59] mmm I still just mean, that it can be syncd, not that we should sync it [14:00] okey dokey [14:02] I honestly think that tomcat-8 needs to be syncd on next release cycle, are the two above useful packages? [14:02] next cycle is when redhat should get dogtag/tomcatjss ported over [14:03] mmm patching security stuff on old tomcat might become painful soon :) [14:03] non-lts releases still [14:12] doko: uploaded new apitrace to debian, fixes build [14:14] LocutusOfBorg, rebuild results in I_XLOCLAE not defined; synced perl over. [14:21] tjaalton: would you mind noting that on https://merges.ubuntu.com/main.html please? [14:23] rbasak: sure [14:24] Thanks! As I read the scrollbackscroll I knew there was a reason but couldn't remember why, and I figure that's a good place to note it for reference the next time I'm asked :) [14:30] yeah, it's there now [14:30] juliank, regarding your gnutls patch, how can i test it? i have to add a repo, right? how can i pull just that one updated library? [14:31] LargePrime: It's all explained in the wiki page that was linked in the bug report https://wiki.ubuntu.com/Testing/EnableProposed [14:32] LargePrime: Just with zesty instead of xenial, of course [14:32] LargePrime: Summary: Add sources.list entry, add preferences entry, and then apt install libgnutls30/zesty-proposed [14:32] thanks again === mbiebl_ is now known as mbiebl [14:52] interesting xnox [14:52] # include [14:52] ^~~~~~~~~~~ [14:52] this happens in libguestfs [14:52] so, I presume until perl is rebuilt, mostly everything will FTBFS? [14:52] LocutusOfBorg, yes. [14:52] https://launchpadlibrarian.net/336044681/buildlog_ubuntu-artful-ppc64el.libguestfs_1%3A1.34.6-7ubuntu1_BUILDING.txt.gz [14:52] ack [14:53] /usr/lib/powerpc64le-linux-gnu/perl/5.26/CORE/perl.h:738:13: fatal error: xlocale.h: No such file or directory [14:53] yeah, that's the badger. [14:53] juliank, from that wiki 'sudo aptitude -t zesty-proposed' gives 'sudo: aptitude: command not found' [14:54] thanks for cathing it early! [14:55] i dont like pain when i cath [15:03] bdmurray, hey, I put some update-notifier fixes for wayland up for review on https://code.launchpad.net/~seb128/update-notifier/wayland-no-pkexec/+merge/330278 https://code.launchpad.net/~seb128/update-notifier/work-under-wayland/+merge/330271 , would be nice if you could review them today I would like to upload this week so we start getting more feedback about issues under wayland, atm users don't get prompted by apport to report those [15:03] seb128: I saw and will have a look today [15:04] bdmurray, thanks! [15:04] LargePrime: You don't need to do that [15:05] LargePrime: It would just show you what is available [15:05] LargePrime: Just apt install libgnutls30/zesty-proposed :D [15:27] xnox: perl ftbfs on i386. test failure. give back? [15:30] "06_timed.t" yes, I would say to give back [15:30] * LocutusOfBorg does it, I think this happened already in the past [15:32] tah for monitoring [16:00] xnox: have you seen bug 1713984? I just had a very long boot myself [16:00] bug 1713984 in systemd (Ubuntu) "started raise network interfaces - times out" [Undecided,Confirmed] https://launchpad.net/bugs/1713984 [16:02] bdmurray, yes and i think i know how to fix it. but not fixed yet. [16:03] xnox: Shall I target to artful? assign to you? [16:03] xnox: Definitely intentional. [16:03] infinity, tah. fix in perl is inflight - rebuild should make perl.h sane again. [16:03] bdmurray, yes please. [16:08] juliank, I have it working. patch works great! my question about the wiki command had to do with noting some sort of error on the wiki. [16:08] also thanks [16:17] infinity: https://bugs.launchpad.net/debian/+source/glibc/+bug/1715366 [16:17] Launchpad bug 1715366 in glibc (Ubuntu) "stage1 builds try to install gdb hooks, which are not built" [Undecided,New] [16:17] but the cross-build fails later with linker errors [16:27] Hi all. Are packages published on *.ubuntu.com on topic here? [16:28] It's really bugging me, that libgtest-dev is not being built in a post-install hook [16:28] this is not Gentoo [16:28] doko: ? [16:30] doko, tell that to the dkms maintainers :P [16:30] Anticom, we cant bild libgtest-dev [16:30] * control: Switch to cmake (upstream deprecated autoconf build). Build [16:30] only static library (remove libgtest0 package). Install full source [16:30] and example files. [16:31] http://metadata.ftp-master.debian.org/changelogs/main/g/googletest/unstable_changelog [16:31] packages depending on it should build it when needed locally [16:31] http://bugs.debian.org/662989 [16:31] Debian bug 662989 in gtest "gtest: libgtest.a needs to be built with -fPIC" [Normal,Fixed] [16:31] LocutusOfBorg: Well maybe i was not precise. Afaik you're supposed to invoke cmake after installing libgtest-dev yourself to produce the *.so's etc. However the package depends on cmake already iirc and debs do support post-install triggers don't they? so why not trigger the cmake 'build' after the package has been installed? [16:31] no, a package you maintain, that needs libgtest-dev should just include its cmake file, and live happy [16:32] LocutusOfBorg: So you mean adding gtest via cmake's add_subdirectory in order to build it as a target before linking against it? [16:33] that one, yes [16:33] hm, doesn't sound very convenient. However as i said i don't understand why everyone should build libgtest themselves anyway [16:33] because providing a system shared library failed [16:34] and google told us to remove it because it can't work [16:34] even the static one [16:34] you will find references to google/debian bugs if you search [16:34] Yea I get that, i've read that myself (in the google docs). However i'm suggesting to invoke cmake after the sources have been installed in a post-install trigger [16:35] Or wouldn't that work either? [16:35] nope, this is not Gentoo :) [16:35] the latter, it wouldn't work [16:35] Bc this is typically what i do. I just sudo cmake where it's installed [16:35] and with my projects i didn't have any issues [16:35] I suspect the concern here is that you need gtest built with the same compiler/linker options as your project. [16:36] I don't remember the exact issue, maybe some ABI [16:36] Hm well, too bad [16:36] Thank you guys for the elaboration :) [16:36] yes, probably, the ABI is not stable, building with different CFLAGS/LDFLAGS/GCC and so on changes the gtest library [16:36] Compiling it at install time would give the same results as us delivering a binary package to you -- it would have defaults that might be wrong. [16:36] exactly, but again, too much time has been since I looked at gtest [16:36] now it is googletest btw [16:37] LocutusOfBorg: nope, still libgtest-dev [16:37] the source package is src:googletest that provides the libgtest-dev libraru [16:37] ah i see [16:38] okay thanks guys [16:38] you are welcome [16:39] hello perl, welcome back on i386 [16:39] publisher run and FUN [17:24] infinity: LocutusOfBorg: I must say though that not shipping the library is a bit annoying. Other distributions do ship it, and the CMake plugin for gtest relies on the library, so you end up with ugly workarounds like the first 22 lines here: https://anonscm.debian.org/cgit/apt/apt.git/tree/test/libapt/CMakeLists.txt [17:24] (re gtest) [17:24] Although if everyone else would just switch over that would work as well [17:44] infinity: any idea about https://launchpadlibrarian.net/336065945/buildlog_ubuntu-artful-amd64.cross-toolchain-base_18ubuntu1_BUILDING.txt.gz ? [17:44] not architecture specific [17:50] argh, https://launchpad.net/ubuntu/+source/gcc-5/5.4.1-12ubuntu3/+build/13343450 ... SIGSEGV now undefined. header moved? [18:23] doko: Probably a side-effect of 9e78f6f6e7134a5f299cc8de77370218f8019237 [18:25] Though, how it manifests only in our cross builds is something to look into. [19:35] tsimonq2: I have added a debdiff file here: https://bugs.launchpad.net/ubuntu/+source/accountsservice/+bug/1699216 [19:35] Launchpad bug 1699216 in gnome-initial-setup (Ubuntu) "Encrypted home support" [Wishlist,Confirmed] [19:35] Let me know if that is what you need [20:10] juliank, http://sources.debian.net/src/pcl/1.8.1%2Bdfsg1-2/cmake/Modules/FindGtest.cmake/?hl=27#L27 [20:12] you are probably right: https://gist.github.com/oneamtu/3734295 [20:28] LocutusOfBorg: The offiical FindGTest.cmake shipped with CMake is different: https://github.com/Kitware/CMake/blob/master/Modules/FindGTest.cmake [20:32] maybe we should make it smarter and submit the patch upstream [20:34] LocutusOfBorg: Probably [20:35] but it's work and it takes time [20:35] the cmake module maintainer might want to do it [20:35] might have migrated to meson by the time it's widely available :) [21:08] jackpot51: ack, I'll look in a few hours. Thanks! [21:13] jsalisbury: Hi, could you please advise what other kernel to test for bug 1646277? [21:13] bug 1646277 in linux (Ubuntu Xenial) "Xenial server 16.04.x will have a black screen after PXE installation" [Medium,Confirmed] https://launchpad.net/bugs/1646277 [22:24] infinity, are you aware of missing xlocale.h ? https://launchpadlibrarian.net/336088033/buildlog_ubuntu-artful-ppc64el.boinc_7.9.0+dfsg~git20170906+r23843~r10~ubuntu17.10.1_BUILDING.txt.gz [22:24] this smells like a glibc regression [22:26] catbus, sorry for the delay on that bug. I'll review it and see what the next steps should be. [22:26] jsalisbury: thank you!