[04:58] <pitti> Good morning
[08:12] <Laney> hey hey
[08:14] <pitti> hey Laney, how are you? did all of you make it to London?
[08:15] <Laney> hey pitti, doing good thanks!
[08:16] <seb128> desktopers!
[08:16] <Laney> still a couple of others to arrive but mainly everybody seems here and still alive :-)
[08:16] <Laney> enjoying the charming ambiance of the bluefin office now
[08:16] <Laney> how are you?
[08:18] <pitti> hey seb128!
[08:18] <pitti> Laney: quite fine, thanks!
[08:18] <seb128> pitti, howdy, how are you?
[08:20] <Sweet5hark> moin!
[08:21] <Laney> first day breakfast misjudging
[08:21] <Laney> urghghgh
[08:37] <Trevinho> seb128: we need this in ups to get proper mouse hidpi/theme updates support in compiz: http://pastebin.ubuntu.com/12762083/
[08:37] <Trevinho> seb128: although.... I've just noticed that it has been proposed a similar thing upstream, but marked as "needs-work" because they want WMs that uses it
[08:37] <Trevinho> https://bugzilla.gnome.org/review?bug=747856&attachment=301556
[08:38] <seb128> oh ok
[08:38] <TheMuso> Hey guys, hope you all had safe travels.
[08:46] <seb128> hey TheMuso, seems we had ;-)
[08:46] <seb128> Trevinho, seems good to include, just mp the change for usd
[08:47] <seb128> pitti, hey, is https://bugzilla.gnome.org/show_bug.cgi?id=756275 (see comment #3) something known in udisk? do you have any hint to debug?
[08:47] <ubot5`> Gnome bug 756275 in libgdu "segfault in g_dbus_object_get_interface" [Normal,New]
[08:47] <seb128> pitti, that's quite a common segfault, seems to trigger often when trying to format devices
[08:49] <CardinalFang> good morning
[08:49] <seb128> CardinalFang, hey, having a new nickname?
[08:52] <pitti> seb128: hm, not off-hand; indeed there seems to be some races in between all the GObjectManager stuff; in the test suite I need to re-get the GObject sometimes, and at lesat call http://udisks.freedesktop.org/docs/latest/UDisksClient.html#udisks-client-settle after formatting
[08:53] <pitti> seb128: this never got debugged, though (or maybe it's the way it's supposed to be); so gdu tries udisks_client_get_object on the new partition immediately after format_sync()?
[08:53] <CardinalFang> seb128: Not intentionally!
[08:54] <seb128> pitti, should that bug moved to udisks/fdo then?
[08:54] <seb128> not sure what gdu does
[08:54] <seb128> but it doesn't do what is needed to avoid the segfault
[08:54] <pitti> seb128: maybe there is one there too -- but in either case it shouldn't crash of coruse
[08:55] <pitti> i. e. it might be that _get_object there should succeed, but this smells like missing error check
[08:55] <pitti> seb128: (sorry, just half a brain here, deep in debugging stuff ATM)
[08:56] <seb128> pitti, no worry/hurry, it just seems an annoying issue and it would be good to resolve next cycle if we can, it just that I don't know enough about udisks/gdu to be useful so I'm trying to add least triage/get the bug with details in the hands of the right people
[08:57] <pitti> seb128: one thing you can certainly try is to shove a client_settle() in between the format_sync() and the get_object(), and see if that helps
[08:57] <seb128> pitti, ok, thanks
[08:58] <pitti> seb128: at least the documentation says that it's meant for exactly this use case
[08:58] <seb128> pitti, otherwise the format/wipefs error, doesn't happen on my amd64 machine, but does every time on the 32 bit one, could be a 32 bits issue
[08:58] <pitti> seb128: ah, it's probably not using format_sync, but async, but please try it anyway to get one more piece of info
[09:02] <seb128> pitti, seems there is already a such call https://git.gnome.org/browse/gnome-disk-utility/tree/src/disks/gducreatepartitiondialog.c#n371
[09:04] <pitti> seb128: could you break on it in gdb, or add a sleep, to see if this is merely a timing issue, or whether the partition object indeed doesn't exist in udisks's brain?
[09:04] <seb128> pitti, sure can, thanks for the debugging hints
[09:04] <pitti> seb128: gdb might actually be nice -- if that still happens after simply waiting for 5 s or so, then you coudl run udisksctl dump to see if it exists in a separate d-bus connection
[09:05] <seb128> k
[09:05] <pitti> i. e. whether the problem is that it doesn't exist in udisks' brain, or there's something wrong with auto-updating the client-side object
[09:05] <pitti> my gut feeling is the latter
[09:15] <seb128> pitti, going to take a bit, can't debug on my machine because udisks_partition_table_call_create_partition_finish() fails
[09:15] <seb128> (the wipefs error issue)
[09:15] <seb128> or maybe I should start by debugging that :p
[09:16] <pitti> seb128: are you trying this on an USB stick?
[09:16] <seb128> yes
[09:16] <pitti> seb128: you can wipe that with dd, or you could even just use "modprobe scsi_debug" to get a virtual SCSI disk
[09:17] <seb128> oh, let me try the virtual one
[09:17] <pitti> seb128: that's massively faster than USB of course, so you might get different timing behaviour
[09:17] <pitti> but it's incredibly convenient to use ;)
[09:17] <seb128> yeah, I didn't know about it
[09:18] <pitti> seb128: by default it's 60 MB, if that's not enough (some file systems need more), you can try modprobe scsi_debug dev_size_mb=300
[09:18] <seb128> pitti, I can only format the partition in gdu with that
[09:19] <seb128> not create one
[09:19] <pitti> seb128: ah, right
[09:19] <pitti> seb128: it's possible, but requires some trickery with scsi_debug
[09:19] <seb128> pitti, don't worry, the usb key does it fine
[09:19] <seb128> and the wipefs issue is probably a real one on 32 bits
[09:19] <pitti> seb128: can you reproduce the wipefs one on the CLI?
[09:20] <seb128> no
[09:20] <pitti> (sudo wipefs -a /dev/foo)
[09:20] <seb128> no, that works
[09:20] <pitti> if it works, you have a way to wipe without gdu; if it doesn't work, you have an easier reproducer
[09:20] <pitti> note that running it twice might work (i. e. a failed one from gdu, and then a successful one)
[09:21] <seb128> gdu/udisks complains about
[09:21] <seb128> "Command-line `wipefs -a "/dev/sdb1"' exited with non-zero exit status 1: wipefs: erreur : /dev/sdb1 : échec d'initialisation de l'analyse: Aucun fichier ou dossier de ce type"
[09:21] <seb128> e.g no such device
[09:21] <seb128> but the device is there, so I guess maybe async work done in the wrong order or something
[11:34] <Trevinho> seb128: for your love https://code.launchpad.net/~3v1n0/unity-settings-daemon/export-xcursor-theme/+merge/274128
[11:38] <GunnarHj> happyaron: Hi Aron, saw that you opened the im-config task at bug #1481025. Are you about to change to "xim"? I think it's important to do so, because I just noticed that "cjkv" breaks X11 compose.
[11:38] <ubot5`> bug 1481025 in im-config (Ubuntu) "Keyboard shortcut for layout switching works in Unity but not in Gnome-Flashback" [Medium,In progress] https://launchpad.net/bugs/1481025
[13:02] <robert_ancell> anpok_, Do you expect bug 1504168 to be in wily or x-series?
[13:02] <ubot5`> bug 1504168 in libinput (Ubuntu) "fix handling of mx4 touch screen" [Undecided,New] https://launchpad.net/bugs/1504168
[13:11] <seb128> larsu, bug #408903
[13:11] <ubot5`> bug 408903 in gnome-settings-daemon (Ubuntu) "Does not handle microphone mute button (KEY_MICMUTE)" [Medium,Triaged] https://launchpad.net/bugs/408903
[14:08] <didrocks> pitti: I guess this is known, but
[14:08] <didrocks> q
[14:08] <didrocks> ok, thanks wall :p
[14:08] <didrocks> http://paste.ubuntu.com/12763398/
[14:08] <didrocks> apparently it does set the shutdown date contrary to previous bug
[14:08] <didrocks> ($wall foo works though)
[14:15] <pitti> didrocks: bug 1495178
[14:15] <ubot5`> bug 1495178 in systemd (Ubuntu) "Failed to set wall message, ignoring: Message recipient disconnected from message bus without replying" [Medium,In progress] https://launchpad.net/bugs/1495178
[14:16] <robert_ancell> larsu, it's only your opinion that the notification bubbles aren't caused by GTK+3.0? (bug 1473269)
[14:16] <ubot5`> bug 1473269 in notify-osd (Ubuntu) "Wily login screen shows an all black notification bubble" [High,In progress] https://launchpad.net/bugs/1473269
[14:18] <larsu> robert_ancell: welcome to me doing bugs on a sprint :D
[14:18] <larsu> (didrocks shouted "OPINION!!!")
[14:28] <didrocks> thanks pitti
[14:28] <didrocks> opinion?
[14:28] <didrocks> onion?
[14:33] <larsu> didrocks: oignon
[14:34] <Laney> zut alors
[14:34] <didrocks> j'ai pas ma baguette
[14:34] <larsu> bibliotheque
[14:34] <didrocks> bibliothèque
[14:34] <Laney> ordinateur
[14:34] <Laney> poisson
[14:34] <Laney> PISCINE
[14:35] <didrocks> Laney est sec maintenant
[14:35] <Laney> triple sec
[14:37] <pitti> c'est me rapelle : c'est l'heure de cours de français :)
[14:37]  * pitti s'en vais, à demain !
[14:38] <pitti> "pitti s'en va"
[14:47] <didrocks> bon courage pitti!
[16:05] <chrisccoulson> hey seb128, how's london? :)
[16:06] <seb128> chrisccoulson, hey, sunny! where are you? are you coming this week?
[16:06] <chrisccoulson> seb128, I'm catching the train over tomorrow afternoon
[16:06] <seb128> good
[16:09] <chrisccoulson> I arrive at kings cross at ~4pm. Not sure how long it takes to get over to southwark from there
[16:10] <chrisccoulson> oh, not long
[16:13] <seb128> chrisccoulson, it's like 4 stops on the distric line, so 15 minutes or so
[16:13] <seb128> well then 15 minutes walk to the hotel/office
[16:21] <seb128> dobey, mvo, do you have any idea how much work it would be to teach software-center about clicks or snaps?
[16:37] <Laney> chrisccoulson: thameslink from st pancras to blackfriars is the fastest way I think
[16:37] <Laney> but tube is also not long
[16:39] <seb128> shrug, forgot to write the actual changelog content to that file-roller but it got accepted anyway
[16:42] <Laney> no way to tell who did it
[16:43] <qengho> robert_ancell: I assigned a lightdm bug report to you and added a debdiff.
[17:03] <popey> BYE EVERYONE!
[17:03] <Laney> ttyl
[17:26] <seb128_> pitti,
[17:27] <seb128_>   udisks_client_settle (gdu_window_get_client (data->window));
[17:27] <seb128_>   sleep(3);
[17:27] <seb128_>   partition_object = udisks_client_get_object (gdu_window_get_client (data->window), created_partition_object_path);
[17:27] <seb128_>   system("udisksctl dump > /tmp/log");
[17:27] <seb128_>   g_print("%x\n", (unsigned int)partition_object);
[17:27] <seb128_> trying that ^ it still segfaults
[17:27] <seb128_> the "log" has sdb1
[17:27] <seb128_> but the partiion_object is null and the code segfaults
[17:28] <seb128_> just dumping bits of info for tomorrow and calling it a day
[17:28] <seb128_> bye
[21:02]  * desrt rides the train to the airport