| === Zic is now known as Guest19585 | ||
| === zumbi is now known as Guest49723 | ||
| slangasek | stgraber: so I have an answer (ish) on the rpc.gssd problem; there's an opaque pointer being passed between two libraries (libgssapi-krb5-2 and libgssglue1), and the two libraries don't agree about the pointer's internal structure. :-P | 00:21 |
|---|---|---|
| slangasek | (libgssglue has a gss_init_sec_context() that assumes an argument of type gss_cred_id_t can be cast to a gss_union_cred_t, which is not the case because the libgssapi-krb5 gss_cred_id_t has a leading loopback pointer at the front of the struct) | 00:24 |
| slangasek | stgraber: so! fixed with the addition of a single configure flag | 00:30 |
| slangasek | stgraber: I'll have a look at the Debian package; I suspect the bug is also present there | 00:31 |
| === psusi_ is now known as psusi | ||
| psusi | so the ppa build daemons get hung up trying to build grub2 and the job has some odd output and never finishes... is there someone who can debug the build daemons? | 00:36 |
| slangasek | stgraber: and that's Debian bug #707960. | 00:37 |
| ubottu | Debian bug 707960 in nfs-common "rpc.gssd segfaults when mounting a nfsv4 volume" [Important,Open] http://bugs.debian.org/707960 | 00:37 |
| infinity | psusi: Example? | 00:38 |
| psusi | infinity, https://launchpad.net/~psusi/+archive/ppa/+build/4631612 | 00:39 |
| psusi | I didn't even modify this build from the current raring sources | 00:39 |
| infinity | psusi: And this has consistently happened more than once? | 00:39 |
| psusi | yes | 00:39 |
| psusi | happens on both archs, tried rebuilding 3 times, and today uploaded an unmodified version to rule out my changes doing it | 00:40 |
| infinity | psusi: It could just that the testsuite hates running either under Xen, on kernel 2.6.32, or on LVM. None of those are things I can "fix". | 00:40 |
| infinity | s/32/24/ | 00:40 |
| psusi | what is different about how the main repo is built? | 00:41 |
| infinity | All of those things. | 00:41 |
| infinity | The main archive builds on bare metal, not Xen, isn't LVM-backed, and is on kernel 3.2.0 | 00:42 |
| infinity | psusi: Anyhow, I wish I could give you a more satisfying answer, but we're in the process of tearing out, redesigning and rolling out a new PPA infrastructure, so we are investing exactly zero resources into hunting weird behaviours in the old one. | 00:43 |
| infinity | psusi: Your best bet is probably to test your builds locally to make sure they pass the testsuite, then upload to your PPA with the suite disabled and call it good enough. | 00:44 |
| psusi | hrm... that was going to be my last resort... I was hoping to find a way to debug the issue thoguh | 00:44 |
| infinity | psusi: Well, the way to debug it would be to run an LVM-backed hardy Xen DOMU on a hardy DOM0. | 00:45 |
| infinity | psusi: Which, as you can imagine, isn't something most sane people want to do, since it's not 2008 anymore. ;) | 00:45 |
| psusi | I can't imagine how the volume ultimately being stored on lvm outside the vm would matter, but the tests do appear to use qemu so maybe it has a problem running under xen? my server I built it on locally is using lvm, but not xen currently... maybe I'll try booting it in xen mode | 00:45 |
| psusi | wait, hardy? | 00:46 |
| infinity | psusi: Ahh, if the tests use qemu, it could be a nesting issue. Xen on qemu on qemu on Xen on qemu on Xen on Xen works pretty well these days, but it was slightly less friendly back then. | 00:46 |
| psusi | wow, so the build daemons are still running hardy? | 00:46 |
| infinity | psusi: Only the virtual PPA ones, but yes. Long story. | 00:47 |
| psusi | any eta on the new build system? | 00:47 |
| infinity | psusi: Comes from us not having Xen support in lucid, so they never found a smooth upgrade path, and by the time we had Xen support in precise (which is quite good), we'd already decided to tear out our homebrew solution and replace it with openstack. | 00:47 |
| infinity | psusi: But that replacement isn't quite done yet. | 00:47 |
| === FlannelKing is now known as Flannel | ||
| infinity | psusi: I couldn't give you an ETA. I know some stuff has progressed to a bit of a testing stage, but how far away we are from rolling it out and calling it production-ready is anyone's guess. | 00:48 |
| infinity | psusi: (The hope is "soon", but certainly not soon enough to make you less sad about your grub builds :P) | 00:49 |
| psusi | so if I disable the test suite and get it to build in my ppa and have the user verify the fix, then when it gets uploaded to -proposed, it should build fine with the tests enabled? | 00:49 |
| infinity | psusi: That seems like a reasonable assumption, given that I haven't seen the grub2 testsuite fail on the archive buildds yet. | 00:51 |
| psusi | say, why do the ppa builds use virtual machines? | 00:51 |
| infinity | psusi: Because ANYONE can upload packages to build on them, which means we're giving the whole world root. | 00:52 |
| infinity | psusi: They're designed to wipe clean and reset after every build. | 00:52 |
| psusi | why not just build with fakeroot, like pbuilder does? | 00:52 |
| infinity | psusi: Err, no. You're missing the attack vector. A PPA can build-dep on packages in a PPA. | 00:52 |
| infinity | psusi: The initial phase of a build is installing build-deps, as root. | 00:53 |
| psusi | ohh | 00:53 |
| stgraber | slangasek: cool, thanks for tracking this down! | 03:31 |
| === ubott2 is now known as ubottu | ||
| === nhandler_ is now known as nhandler | ||
| === Logan_ is now known as Guest19720 | ||
| === Guest83934 is now known as Logan_ | ||
| === Logan_ is now known as Guest24881 | ||
| === Guest19720 is now known as Logan_ | ||
| === Logan_ is now known as Guest52577 | ||
| NikTh | Hello everyone | 05:49 |
| NikTh | Is here a good chanell to ask about Unity Next ? I've just downloaded per and builded per instructions here: http://unity.ubuntu.com/getinvolved/development/unitynext/ , but a problem occured.. | 05:51 |
| NikTh | I cannot running it because of this error : ./run: 56: ./run: ./builddir/qml-phone-shell: not found | 05:51 |
| NikTh | Tried to mkdir builddir and cp qml-phone-shell.desktop.in file inside but to no avail. (other bunch of errors occured). Also tried ./build_unity --clean (rebuilding) again to no avail. | 05:53 |
| === JanC_ is now known as JanC | ||
| === ubott2 is now known as ubottu | ||
| === mitya57_ is now known as mitya57 | ||
| === Guest19585 is now known as Zic | ||
| === liam_ is now known as Guest26501 | ||
| mitya57 | BenC: Please check "Note about PowerPC support changes" section in bug 1186577. It will be nice if you verified if we should keep or drop that delta. | 16:47 |
| ubottu | bug 1186577 in gccxml (Ubuntu) "Sync gccxml 0.9.0+git20130511-1 (main) from Debian unstable (main)" [Wishlist,New] https://launchpad.net/bugs/1186577 | 16:47 |
| mitya57 | (In other case we'll sync and re-add that if it still FTBFS) | 16:49 |
| mitya57 | jbicha: thanks for uploading metacity! | 16:52 |
| === hggdh_ is now known as hggdh | ||
| === jono is now known as Guest59264 | ||
| === txwikinger2 is now known as txwikinger | ||
| gotwig | where can I get support for ubuntu lts | 20:28 |
| penguin42 | gotwig: #ubuntu for free help | 20:29 |
| penguin42 | gotwig: You can submit bug reports using the ubuntu-bug command | 20:29 |
| penguin42 | gotwig: http://www.ubuntu.com/support/ | 20:32 |
| gotwig | penguin42: the thing is, there is no right click support for ubuntu 12.04 for multitouch clickpads | 20:33 |
| gotwig | I hope it gets fixed in the next LTS upgrade | 20:33 |
| penguin42 | I don't know anything about multitouch - sorry | 20:34 |
| === Ursinha is now known as Ursinha-afk | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!