/srv/irclogs.ubuntu.com/2013/11/07/#ubuntu-desktop.txt

=== thumper is now known as thumper-afk
Mirvmorning04:56
cyphermoxMirv: morning05:21
Mirvcyphermox: hullo05:21
cyphermoxso, so far I haven't been able to find out what's wrong with phonesim -- if I disable it though, the lab devices seem to work properly05:22
cyphermoxso I think we'll need to involve pitti to look at what's up with phonesim or the related pieces05:23
pitticyphermox: define "wrong"?05:25
* Mirv switches computers to dedicate this one to test libunity05:25
cyphermoxpitti: unclear, sadly05:25
cyphermoxpitti: seems that now, the phones in the QA lab tend to get stuck with indicator-network/dbus/ofono in high CPU05:26
cyphermoxit happens about 30% of the time, apparently05:26
cyphermoxbut you quickly notice indicator-network hiking up to 60%+ CPU for a while when phonesim gets restarted, and sometimes (that 30% I was telling about) on boot05:27
pitticyphermox: oh, is that still the "NetworkManager crashes in a loop" problem?05:27
cyphermoxI looked, but I can't make sense of the reason -- it's hard to figure out what's going on05:27
pittibug 1231964 ?05:27
ubot2Launchpad bug 1231964 in network-manager (Ubuntu) "NetworkManager crashes in ofono plugin with phonesim [crashed with SIGSEGV in <unavailable> in ??()]" [High,Triaged] https://launchpad.net/bugs/123196405:27
cyphermoxpitti: NM doesn't seem to be crashed, but hold on05:27
cyphermoxthat could be it, maybe05:27
pittiI didn't notice this NM crash any more on my recent phone test runs, though05:28
pittiwhen I reported the bug, I had to kill NM to make it not go crazy05:28
pittibut I haven't done this any more for weeks05:28
cyphermoxthis isn't a new bug though, so yeah05:28
pittibut as NM didn't change, I guess we are missing something there05:28
cyphermoxwell05:28
pittiso I figure this only happens in some cases05:28
cyphermoxNM did change -- we're getting 0.9.8.4 now in the images05:28
pitticyphermox: oh, it's only happening since that update?05:29
pittiI haven't tried .4 on the phones yet indeed05:29
cyphermoxbut the ofono code certainly hasn't change between then and now05:29
pitticyphermox: is that on the latest trusty image? flashing now05:29
cyphermoxin theory yes05:29
cyphermoxI don' think you'd see indicator-network in high CPU though, in that case05:30
cyphermoxhold on, I can check if NM crashed on the mako used by psivaa as an example05:30
cyphermoxhmm05:31
cyphermoxno crash in /var/crash for NM05:31
cyphermoxand no segfault apparent in /var/log/syslog.05:31
pittiok, so it's not that then05:31
cyphermoxpitti: if you want to play with this stuff, you can use adb  -s 01ade38b552014d4 shell05:32
cyphermoxin the lab05:32
pittiso 9 hours ago we got https://code.launchpad.net/~boiko/dialer-app/fix_ap_1.4/+merge/194205/comments/447936 which apparently at least succeeded05:32
cyphermoxah?05:32
cyphermoxpitti: didrocks mentioned it was happening since Tuesday.05:33
pittihttps://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-mako/3174/? has two crash reports, but they've always happened (and are known bugs)05:33
cyphermoxyeah, looks consistent with what I have on the device right now05:34
cyphermoxpitti: as one thing I noticed in recent proposed images was that I didn't have write access to /var/lib/dpkg even with a writable image, maybe this is breaking dbus in some way that upsets phonesim05:37
cyphermoxor phonesim itself05:37
cyphermoxnevermind, both are explicitly listed in mounts05:38
=== duflu_ is now known as duflu
cyphermoxpitti: so here's the full list of changes from 1105, which would be relevant for what we're looking at, possibly: http://people.canonical.com/~ogra/touch-image-stats/20131105.changes05:46
* pitti taps foot waiting for flashing to finish05:49
cyphermoxhehe05:50
pitticyphermox: ok, flashed05:54
pitticyphermox: so, do you already see this happening when merely starting phonesim, or just when the tests run?05:55
cyphermoxyeah even just starting with-ofono-phonesim I see it happening05:55
pittii. e. should I do "sudo apt-get install ofono-phonesim-autostart"?05:55
pittiack05:55
cyphermoxyup05:55
pitti2394 B/s05:58
pittioh c'mon, ports.u.c.!05:59
pitticyphermox: so this image ("trusty" channel) still has NM 0.9.8.0-0ubuntu2206:00
pittiI wonder where ogra_ takes his image from :)06:01
cyphermoxtursty-proposed :)06:01
cyphermoxtrusty, I mean06:01
pittithat's what I used06:01
cyphermoxthe proposed one?06:01
pitticyphermox: started "sudo with-ofono-phonesim", indicator-network spins CPU at 70% indeed06:02
cyphermoxwell at least now we're confirmed without doubt that it's not NM06:02
pittiand on the phone, when I open it I see "Empty!"06:02
cyphermoxbut it makes it that much less clear what's wrong06:02
pittistrace says its spinning on a poll()/write() loop06:03
cyphermoxyeah, I was seeing the same06:03
abhishek_I want to port Ubuntu Desktop on my Development board which is having Android06:03
cyphermoxisn't write failing?06:03
abhishek_Can someone please helpme06:03
pittifd 11 is anon_inode:[eventfd], whatever that is06:03
cyphermoxor maybe it was a read what was failing with EAGAIN before06:03
pitticyphermox: no, write succeeds, read() is EAGAIN06:04
cyphermoxpitti: what process is that?06:04
pittiand there is a really fast poll which times out06:04
pitticyphermox: indicator-network-service06:04
cyphermoxyeah06:04
cyphermoxok06:04
pitti$ pkill -f indicator-network-service; /usr/lib/arm-linux-gnueabihf/indicator-network/indicator-network-service06:05
pitti** (process:4256): ERROR **: network-menu.vala:202: Unable to get D-Bus bus name06:05
pittimeh, it respawns too fast06:05
pittiand that's why useful D-BUS services have a "--replace" option :)06:05
cyphermoxso hey, I couldn't see the ofono-phonesim.conf dbus config file in /etc/ofono, that's because of the mount in with-ofono-phonesim, right/06:06
abhishek_ogra told me to create kernel zImage and initrd.img images ....and pack both these using abootimg utility .......I have created the zImage ....but I don't know how to create initrd.img. Please tell me how to create06:06
pitticyphermox: right, it's an unshared tmpfs overlay, to avoid modifying the system06:06
pitticyphermox: hm, I managed to start it in the foreground now, but its quiet now06:07
cyphermoxabhishek_: phone?06:07
abhishek_I have IFC6410 development board which is having android06:07
abhishek_I have android BSP available to me06:08
cyphermoxabhishek_: ogra_ should be around again soon, I'd think, you may want to ask him directly, and probably more in #ubuntu-touch if it's for touch06:08
abhishek_I want to run Ubuntu on this board06:08
cyphermoxok06:08
abhishek_It's for Desktop06:08
abhishek_cyphermox: Can you please give me some guidlines ....to create initrd.img ...06:09
abhishek_cyphermox: I am using Kernel source from Android source .....06:10
cyphermoxabhishek_: that usually gets built along with teh zImage06:10
abhishek_cyphermox: Please give me some guidelines ...this will help me alot06:10
pitticyphermox: got it to spin now in the foreground, but no output at all (except for a single "CRITICAL **: nm_device_get_iface: assertion 'NM_IS_DEVICE (device)' failed" which happened when I started phonesim)06:10
cyphermoxhrm06:11
cyphermoxpitti: spin as in do nothing useful?06:11
pitticyphermox: do you see polkitd spinning as well? (and apport coming up often)06:13
cyphermoxI had not noticed it, no06:13
pittiERROR: apport (pid 8040) Thu Nov  7 07:13:38 2013: called for pid 8036, signal 5, core limit 006:13
pittiERROR: apport (pid 8040) Thu Nov  7 07:13:38 2013: executable: /usr/lib/arm-linux-gnueabihf/indicator-network/indicator-network-service (command line "/usr/lib/arm-linux-gnueabihf/indicator-network/indicator-network-service")06:13
pittievery 2 seconds06:13
cyphermoxlet me look it up06:13
cyphermoxhey, that would explain a lot06:14
pittibut that might be because I'm running it in the foreground now, and D-BUS activation repeatedly fails because the name is taken06:14
pitti/usr/share/upstart/sessions/indicator-network.conf06:14
pittioh dear, why do we have to upstartify everything?06:14
* pitti yells "D-BUS activation!"06:14
abhishek_cyphermox: I didn't found that .....I used update_initramfs to create initrd.img for my Laptop kernel (when I upgraded my laptop kernel) .......But, I don't know how to create the initrd.img for the kernel which is meant for some other board?06:15
pitticyphermox: yep, that stopped the apport/polkit loop :)06:15
pitticyphermox: ok, running under gdb in the foreground now06:15
pittibut now it's quiet, of course06:16
cyphermoxabhishek_: I don't know. I think you should ask ogra_ directly06:16
cyphermoxpitti: yeah :/06:16
abhishek_cyphermox: Ok ....thanks06:16
pitticyphermox: ok, reboot, and attaching gdb directly when it happens06:16
abhishek_cyphermox: Is orgra around ?06:17
pittiunlikely at this hour06:17
abhishek_*ogra06:17
pittigive it another hour or two06:17
abhishek_ok06:17
cyphermoxoh!06:17
cyphermoxpitti: I got a g_source_destroy_internal() in gdb now06:18
cyphermoxah, I get all kinds of things if I'm lucky enough06:19
pittiyeah, all main loop/poll stuff06:22
cyphermoxactually, not so much06:22
pittiaah06:24
pitti$ pkill -f indicator-network-service; G_DBUS_DEBUG=all G_MESSAGES_DEBUG=all /usr/lib/arm-linux-gnueabihf/indicator-network/indicator-network-service06:24
cyphermoxfor instance: http://paste.ubuntu.com/6374719/06:24
pitticyphermox: ^ that's the magic06:24
cyphermoxah, wise!06:24
cyphermoxI've exceeded my efficiency time for tonight :)06:25
pitticyphermox: so you apparenlty can stop the spinning by merely killing the indicator, it'll respawn and behave06:25
cyphermoxno06:25
cyphermoxmaybe sometimes it works but I did kill it a few times and got it back i nthe same state06:26
pitticyphermox: http://people.canonical.com/~pitti/tmp/indicator-network-service.log.xz06:28
cyphermoxliek http://paste.ubuntu.com/6374736/ ?06:29
pittiit loops in StartServiceByName()06:29
pitticyphermox: yes06:30
cyphermoxyeah06:30
pittiThe name org.ofono was not provided by any .service files06:30
pitti!?06:30
cyphermoxwell eventually it does find it though06:32
pittiI let it spin for 20 seconds or so, ofono was well up and running at that time06:34
cyphermoxhttp://paste.ubuntu.com/6374745/06:35
cyphermoxyes, it was06:35
cyphermoxsomething's wrong06:35
cyphermox^ ofono is up, indicator talks to it06:35
cyphermoxin the end it's still waiting for something though06:36
cyphermoxurgh06:37
cyphermoxhttp://paste.ubuntu.com/6374764/06:37
pitticyphermox: do you see anything interesting there?06:40
cyphermoxnevermind no06:41
cyphermoxcan I let you keep looking into it?06:44
cyphermoxI really should get to bed by now, it's nearing up on 2am06:45
pitticyphermox: yeah, better go to bed; I've got some other fires to put out, but I guess it's not that desperately urgent?06:45
cyphermoxit's pretty crippling for the testing of the phone images, apparently06:46
cyphermoxbut I have really no way to tell how urgent, better ask didrocks about that06:46
Mirvcyphermox: would you be still up to pkg ack? http://pastebin.ubuntu.com/6374808/06:51
cyphermoxsure06:52
Mirv(basically bzr 306 + 307 from https://code.launchpad.net/~unity-team/libunity/trunk)06:52
Mirvwe need it to unblock unity + ap + everything, I've now tested it06:52
cyphermoxyeah, it's fine. I just wish they made commit messages a bit more meaningful for changelogs than just "Make tracing easy"06:53
pittitest_extras_LDFLAGS = -static !?06:53
pittithe rest looks okay to me06:54
pittiMirv: ^06:54
Mirvyes, I guess it's related to "Making tracing easy" but explanation would be nice06:54
MirvI wonder if mhr3 would be up soon06:54
Mirvfiled bug #124884006:57
ubot2Launchpad bug 1248840 in libunity "Make libunity test extras not link static or add a comment" [High,New] https://launchpad.net/bugs/124884006:57
Mirvpitti: does it hurt if it's in tests only, so not in published binary packages anywhere?06:59
pittiMirv: not sure, but why would you want it in the first place?06:59
pittiit seems better to me to test the actual .so libs that you built06:59
Mirvpitti: beyond me, my only guess it's related to that tracing one way or another. but I filed the bug to get an explanation committed.07:00
Mirvpitti: it seems all the other tests already had the -static, this just adds it to the _extras too07:02
Mirvpitti: since it's just harmonizing the Makefile to what was already being used, do you think the bug report is enough for now? I'd like to release it, and all tests ran fine.07:14
Mirvit's blocking the autopilot 1.4 transition at the moment07:14
pittiMirv: oh yes, I do; thanks for filing07:15
Mirvok, thanks07:15
=== slomo_ is now known as slomo
Mirvit happened again! drwx------ 2 root root  80 marra  7 08:46 pulse (in /run/user/1000)07:45
MirvI wonder what's causing that, I need to chown it to make sound work again07:45
mlankhorstmorning btw :P08:03
Mirvmorning mlankhorst08:06
Mirvmornign sil210008:18
sil2100Morning!08:26
sil2100Mirv: hmmm, strange, but stuff didn't leave proposed yet from what I see?08:27
Mirvsil2100: nothing strange about that, see landing plan08:27
sil2100Mirv: Ken checked in my evening and he said that the last missing link blocking  stuff was unity-autopilot08:28
sil2100Mirv: ah, libunity08:28
sil2100Mirv: this Ken didn't notice, oh well08:28
Mirvsil2100: yep, but it was not really08:28
sil2100Mirv: thanks for publishing that o/08:29
Mirvtesting was hard, that's why it's already 3.5h since I started08:29
Mirvcu2d was of no use etc08:29
didrocksbonjour pitti! ça va?09:05
pittibonjour didrocks ! ça va bien, et toi ?09:05
didrockspitti: on fait aller :)09:06
pittididrocks: c'est tard pour toi, c'est encore le jetlag ?09:06
didrockspitti: cyphermox told me you are looking at what is causing the network indicator to become crazy in the QA lab after boot?09:06
didrockspitti: oui, encore du jetlag :/09:06
didrockspitti: like, it would be an ofono-phonesim issue?09:06
pittididrocks: cyphermox and I took a stab at debugging this this morning, but we didn't yet get that far; have to put out some fires first09:06
pittididrocks: it seems indicator-network-service is spinning on trying to contact ofono over D-BUS09:07
didrockspitti: I'm still puzzled as http://people.canonical.com/~ogra/touch-image-stats/20131105.changes doesn't have any ofono/indicator changes09:07
didrocks(and as told in my email, we try to revert network-manager)09:08
didrockspitti: do you think you'll have time to have a look later on? It's basically what is going to block us to publish some results09:08
pittididrocks: not sure where ogra_ took his image from, but I flashed from the "trusty" channel this morning, that got me the old network-manager for example, so I guess it's an older image09:08
pittididrocks: don't bother, it's not network-manager09:08
didrockspitti: trusty-proposed09:08
pittierr, devel, devel-proposed, trusty, trusty-proposed..09:09
pittididrocks: but as I said, it's nothing to do with http://people.canonical.com/~ogra/touch-image-stats/20131105.changes09:09
pittididrocks: where do you see this?09:09
didrockspitti: on the image containing those changes (in the QA lab only)09:09
pittididrocks: we ran a phonesim dialer-app test just about 10 hours ago on mako and maguro which succeeded09:09
didrocksso image 12 and 1309:10
pittididrocks: I mean, which test does now fail which didn't before?09:10
didrockswell, tests are impacted because indicator-network is taking > 63% of CPU09:10
didrocksso a lot of AP fails09:10
didrocksthe system isn't idling09:11
pittididrocks: do we run all autopilot tests at the same time now, i. e. does the recent addition of ofono-phonesim-autostart to the dialer-app/messaging-app tests now cause all the other autopilot tests to run with that as well?09:11
pittididrocks: ah, we don't run the AP tests in separate runs, but all in one session?09:11
didrockspitti: yeah, there is one boot09:11
didrocksand running all AP tests for this app09:11
didrockspitti: http://reports.qa.ubuntu.com/smokeng/trusty/touch/mako/13:20131105.1:20131031.1/4910/ubuntu-rssreader-app-autopilot/09:11
pittididrocks: that sounds a lot like that D-BUS spinning has been there all the time, it just didn't get exposed until now because ofonod wasn't restarted09:11
didrocksfor instance, you can see 3 AP failing09:11
didrockspitti: look at multiple top snapshot at boot: https://jenkins.qa.ubuntu.com/job/trusty-touch-mako-smoke-ubuntu-rssreader-app-autopilot/16/artifact/clientlogs/top_before.log/*view*/09:12
didrocks2197 phablet 20 0 253920 220080 3796 S 70.8 11.5 3:27.03 /usr/lib/arm-linux-gnueabihf/indicator-network/indicator-network-service09:12
didrocks(last run, 5 minutes after the phone booted up)09:12
pittididrocks: I need some ~ 10 more minutes to finish fixing the flood of autopkgtest failures, then I can go back to this09:12
didrockspitti: \o/ thanks a lot :)09:12
didrockspitti: the additiona of ofono-phonesim-autostart to some tests can explain that09:13
pittiyes, as these restart ofono09:13
pittiwhich probably is the thing that breaks the network-indicator09:13
didrocksbut once dialer-app tests are over09:13
didrockswe reboot the device09:13
didrockswait… doing nothing09:13
didrocksfor 5 minutes09:13
didrocksand then start rssreader tests09:13
pittiyes, but ofono-phonesim-autostart is still installed I guess09:13
didrockswill your mock influence?09:13
didrocksright09:14
pittiit's not a mock09:14
pittiit's an actual ofono driver09:14
didrocksah ok ;)09:14
didrocksand so… it restarts ofono at each boot09:14
pittiso after boot, ofono will still restart, possibly after indicator-network09:14
didrocksso, depending in which order we do the tests09:14
didrocksas soon as we start one depending on ofono-phonesim-autostart and we install it, all the following tests are screwed09:15
didrocksat least it *starts* to make sense to me09:15
* didrocks was soooo puzzled yesterday09:15
didrockswe tried to reflash, only rerun some tests…09:15
pittiI didn't see this kind of failure in the MPs09:15
didrockssomething all went fine and suddenly, 100% strike09:15
pittiand trust me, mako/maguro ran these tests dozens of  times :)09:16
pittibut that was all in isolation, i. e. just the dialer or just the messaging tests09:16
didrockspitti: yeah, it's weird you didn't get it in the upstream merger09:16
didrocksright09:16
pittiso those didn't exhibit the indicator going crazy09:16
pittididrocks: well, possibly I did (even locally), but I didn't check top09:16
pittididrocks: it's perfectly reproducible on my phone, too09:17
didrockspitti: let me try to install it and confirm ;)09:17
* didrocks hugs pitti. So happy to have a start of explanation09:17
pittididrocks: install ofono-phonesim and xvfb, run "sudo start with-ofono-phonesim", wait a few secs09:17
didrockswas really depressing yesterday09:17
* pitti hugs you back09:17
pittididrocks: but yes, these tests also expose lots of Mir and other crashes :)09:18
didrockspitti: nice stress test! ;)09:18
didrocksok, I'll let you go on the autopkgtests, just ping me if you need access to anything09:18
Laneymorning09:18
pitti(which is why it took me so long to land them in the first place -- Mir crashes caused the MP to get rejected)09:18
didrocksargh "fun" :)09:18
pittididrocks: nope, can debug on my device09:18
didrockspitti: ok, just keep us in touch. I think there is no real value that we run the tests before you get a fix09:19
pittididrocks: just uploaded systemd, postponing gvfs autopkgtest, other tests are green again, so looking now09:19
didrocks\o/09:19
seb128good morning desktopers09:20
didrockshey Laney, seb12809:20
seb128hey Laney pitti didrocks09:20
pittibonjour seb12809:20
=== Rodge is now known as Rodge_
pittididrocks: actually, much simpler than that09:22
pittididrocks: "sudo stop ofono; sudo start ofono", that makes it go crazy09:22
pittididrocks: which is actually a nice test, as it very much does need to cope with htat09:22
pittididrocks: so, nothing to do with my 13117 unshare -m tricks or so :)09:22
didrockspitti: confirmed!09:22
pittididrocks: want me to file a bug, or is there one already?09:23
didrockspitti: I don't think there is any09:23
seb128didrocks, did you manage to resolve the tests issue yesterday?09:23
didrockspitti: hum, after a reboot, I can't get the indicator going crazy09:23
didrocksseb128: we are discussing about it, there is a lead already09:24
seb128didrocks, so it's not ido at the end?09:24
didrocksseb128: can you try "sudo stop ofono; sudo start ofono"09:24
seb128didrocks, on my desktop or on the device?09:24
didrocksseb128: device09:24
seb128sure, give me a min09:24
didrockspitti: grrrr, it's not that easy, worked the first time, but now, can't retrigger it. Is it reliable for you?09:25
seb128didrocks, that seems to be fine ... but my device is not uptodate ... /me updates09:26
* pitti files bug 124888009:27
ubot2Launchpad bug 1248880 in indicator-network (Ubuntu) "spends lots of time spinning on D-BUS when ofono restarts" [Undecided,New] https://launchpad.net/bugs/124888009:27
pittididrocks: fairly, yes09:27
pittididrocks: I'll add some extra information with d-bus debugging, hold on09:28
didrockspitti: tried to reboot twice, can't reproduce :/09:29
pittihm, I can reproduce it every time09:31
didrockspitti: do you have ofono-phonesim installed?09:31
pittididrocks: but sometimes it only lasts some 10 seconds until it reconnects, sometimes 3 minutes09:31
pittididrocks: yes, but I didn't run it (and I don't have -autostart installed)09:31
didrocksstill no luck :/09:32
seb128pitti, do you have a sim in the device? does that make a difference?09:32
pittiseb128: I do, and I don't know09:32
seb128didrocks, do you have one?09:32
pitti$ /usr/share/ofono/scripts/list-modems09:33
pitti[ /ril_0 ]09:33
pitti    Features = net sms gprs sim09:33
pittiwithout a SIM, it could indeed make a difference09:33
pittithat's perhaps why you see it better with phonesim09:33
pittias that always exports a device09:33
pittididrocks: updated the description accordingly09:35
pittiI now have a case where it has spun for > 1 minute already09:36
didrockspitti: confirmed with popey09:36
pittiit seems it's rolling dice and has an 1% chance of reconnecting or so09:36
didrocksa SIM is needed09:36
pittididrocks: or phonesim :)09:36
pittididrocks: I updated the description to show how to reproduce without a SIM09:36
didrockspitti: ah, will try after that meeting, thanks!09:39
pittididrocks: also posted a link to the g_bus_watch_name() API docs which I suggest using09:43
mfischseb128: that dbus api for adding launcher icons works great, but in some circumstances the tooltip says "waiting to install" which is odd09:44
seb128mfisch, seems like an unity bug, maybe check with bregma/Trevinho09:45
mfischseb128: yeah I found an old one in LP that I'm looking at09:45
pittididrocks: should I do anythign else with this bug? assign it to someone, etc?09:45
didrockspitti: no, I'll hunt for people :)09:48
didrockspitti: back from the meeting, let me try installing the various pieces09:48
didrockspitti: confirmed that with-ofono-phonesim, the indicator goes crazy09:49
didrocksthanks a lot pitti!09:49
didrocks:)09:49
* didrocks targets a victim now :)09:50
* seb128 is happy to be off didrocks' enemy list09:50
didrocksseb128: until next time! :-)09:50
* seb128 is also happy that didrocks let him land the ido segfault fix yesterday09:50
seb128didrocks, yeah ;-)09:50
didrocksseb128: in fact, it's why it was so puzzling09:50
didrocksseb128: depending on the test order09:51
didrocksyou install that package earlier or later09:51
didrocksand then, the package stays installed09:51
didrockswhen you run other tests09:51
didrocksthat's why it was a "wth"09:51
seb128right, that's making for fun debugging09:51
didrocksand we didn't find any culpurit in the diff list09:51
didrocksright09:51
pittiseb128, didrocks: ido> oh yes, so am I, thanks for the fast fix!09:51
pittithis was so annoying09:51
didrockspitti: well, thanks seb128 to fight for it :p09:52
seb128pitti, thanks to larsu ;-)09:52
* seb128 hugs didrocks09:52
* pitti donne une accolade à larsu09:52
* didrocks hugs seb12809:52
* pitti hugs seb128 and didrocks09:52
seb128group hug ;-)09:52
* didrocks rehugs pitti09:52
larsu:)09:54
pittididrocks: quick, assign larsu to the network indicator issue, too! he's good with gdbus!09:55
* pitti runs09:55
didrockslarsu: I think you are the perfect target (the first to answer on IRC though)09:56
didrockslarsu: seriously, interested in a somewhat-really-critical indicator-network-gdbus issue? ;)09:56
didrocksit's a nice story, with a lot of praise for fixing it :)09:56
=== thumper-afk is now known as thumper
larsuerm...09:57
larsuno?!09:57
* larsu is afraid09:57
didrockslarsu: bug #1248880, do you have any idea on how to debug it?09:57
ubot2Launchpad bug 1248880 in indicator-network (Ubuntu) "spends lots of time spinning on D-BUS when ofono restarts" [Critical,New] https://launchpad.net/bugs/124888009:57
larsu(a) I always want pitti-level bug reports from now on10:02
larsu(b) this looks really interesting :)10:03
* larsu goes for tea and has a more dtailed look later10:03
seb128larsu, would you describe pitti's bugs the equivalent of mpt bugs, just applied to technical issues? ;-)10:03
didrocksoh no, we spoiled larsu with high quality bug report :)10:03
didrocksseb128: quick, let's edit it :)10:03
larsuseb128: exactly10:05
pitti* description changed to: *10:05
pittiItz broken! fix it!10:05
larsuhaha10:05
didrocks;)10:05
* pitti removes reproducer, analysis, and pointer to "you should use this" too, way too little fun otherwise!10:06
* larsu wonders how ServiceUnknown is send sometimes but then the service gets started sometimes10:06
pittilarsu: it spins on ServiceUnknown for 5 to 500 seconds and eventually seems to hit the sweet spot and reconnect10:07
pittior these are just the 10 gazillion queued requests that it sends up in a tight loop10:08
pittig_bus_watch_name()!!10:08
larsuya, that sounds more like it10:08
larsuted is known for "oh, didn't work, let's try again immediately"10:08
larsuso... indicator-network doesn't even build for me :-(10:17
seb128larsu, wrong vala version?10:21
larsuargh, I guess I need to be on trusty to get vala 0.22?10:21
seb128larsu, no10:22
larsuseb128: yes, they've just updated it in r31210:22
larsu_without_ bumping the vala dep10:22
* larsu shakes fist at kenvandine!10:22
seb128larsu, install valac-0.2210:22
seb128larsu, then "update-alternatives --config valac"10:22
larsuno such package10:23
seb128larsu, that let you pick the default one10:23
seb128of course10:23
seb128larsu, https://launchpad.net/ubuntu/+source/vala-0.22/0.22.0-2/+build/515728210:23
seb128get the debs from there and dpkg -i10:23
seb128they should install fine on saucy10:24
larsumakes sense, thanks10:24
seb128yw10:24
larsuI guess I should jjust update this week10:24
larsualso needs valac-0.22-vapi10:30
larsu*sigh*, vala10:30
seb128larsu, sorry10:32
seb128larsu:10:32
seb128https://launchpad.net/ubuntu/+source/vala-0.22/0.22.0-2/+build/515728510:32
seb128larsu, ^ that has the arch all binaries10:33
larsuthanks :)10:33
larsugot it to build now10:33
LaneyI'd have just added trusty to sources.list and installed it from there :P10:33
seb128I was going to suggest that next10:33
Laneythen probably forgotten to remove it and accidently upgraded to trusty10:33
* Laney did that at client sprint10:33
larsuhaha10:33
seb128but that's usually more work that wget & dpkg in a deb10:33
larsuLaney: is it stable enough to run?10:33
Laneyyeah10:34
Laneyand you get click packages in the dash apparently10:34
LaneyMADNESS10:34
seb128larsu, yeah, there was an annoying ido segfault but I got told it's fixed :p10:34
* larsu whistles10:34
Laneyinfact it seems to prefer them over regular applications10:34
Laneyactually wtf10:35
LaneyI guess "Browser" is webbrowser-app which we recently seeded10:36
larsu`sudo restart ofono` gives me "restart: Unknown instance:"10:36
pittioha, not running?10:36
pittilarsu: try stop ofono; start ofono?10:36
LaneyI'm only getting new apps on the home lens now before typing anything10:37
larsupitti: stopping gives me the same error10:37
* larsu kills it10:37
larsuah, that fixes it10:37
larsuapparently upstart didn't know ofono was running10:37
seb128I guess the running one was not upstart managed10:38
seb128upstart doesn't take ownership of random processes you run manually10:38
larsuthat makes sense, but I wonder how that could have happened10:38
seb128is ofono dbus activated?10:38
larsuoh! Maybe it wasn't running before and indicator-network dbus-activated it10:38
seb128right10:38
larsuand upstart doesn't know about this10:38
larsumy favorite problem! :P10:38
seb128I was going to say :p10:39
larsupitti: I must dissapoint you. I can't reproduce :-/10:40
pittilarsu: do you have a SIM card?10:40
pitti*cough* systemd! *cough*10:41
pittisorry10:41
pittilarsu: if you don't have a SIM card, use the second reproducer in the description, which will fake one for yuou10:41
larsupitti: yeah sorry I just read your description in full :)10:42
Laneyjibel: did / can you forward your new autopkgtests to debian?10:42
larsuoh yeah here we go10:42
pittiLaney: oh, for what package?10:42
Laneyparamiko python-jsonschema10:43
Laneyare the two I see atm10:43
Laneymaybe more10:43
pittiparamiko is collab-maint, happy to commit for you, jibel10:43
pittijsonschema is "openstack", that needs a bug10:43
Laneyoh, is that allowed?10:44
pittiwell, it's called "collab-maint"..10:44
pittiI've been doing it many times, didn't get a complaint10:44
LaneyInteresting, I wouldn't complain myself10:44
larsuhm, ofono doesn't even ship with a .service file10:45
larsuStartServiceByName is just plain wrong10:45
pittiand Debian's release team also annoucned that they want to enable autopkgtests, and provide quicker testing propagations for packages which have them, and encourage people to add them10:45
pittilarsu: yeah, it's only handled by upstart ATM10:45
larsuokay I'll remove that then10:46
pittilarsu: hence, watching for dbus names to STFU while it's not running10:46
Laneythere's also https://bugs.launchpad.net/ubuntu/+source/python-dateutil/+bug/1247877 https://code.launchpad.net/~jibel/ubuntu/trusty/python-mock/lp1248066_enable_autopkgtest/+merge/193888 https://code.launchpad.net/~jibel/ubuntu/trusty/python-imaging/lp1248743_enable_autopkgtest/+merge/19424910:46
ubot2Launchpad bug 1247877 in python-dateutil (Ubuntu) "Enable autopkgtest" [Medium,Triaged]10:46
Laney-mock is DPMT, others are maintained by people (last one is doko)10:46
didrocksjust a long time I haven't done propaganda, but Dart 1.0 is coming! :)10:47
jibelLaney, sure thing, will do.11:23
Laneyrocking11:24
jibelLaney, is it preferable to submit to debian directly then request a sync or propose a patch in both distro?11:24
Laneyjibel: as you wish11:24
LaneyIt might take a while for your patches to be uploaded11:25
pittiLaney, jibel: if I commit stuff to collab-maint for an otherwise ubuntu-unmodified package, I usually do a -2git1 upload to ubuntu; that keeps autosync, but immediately gets us the change11:29
Laneyworks for me11:29
larsupitti, didrocks: I've put up a branch that should fix your issue for now. Please let me know if it works.11:35
larsuwe'd need a lot more substantial fixes, but I don't feel comfortable making them right now11:35
larsuas I'm not the maintainer of that project and have never read the code before11:35
didrockslarsu: \o/11:35
didrockslarsu: do you have an url?11:36
pittilarsu: I don't see a branch on the bug?11:36
larsupitti: ya I was waiting for the push to complete while typing11:36
larsudidrocks: https://code.launchpad.net/~larsu/indicator-network/dont-activate-ofono11:36
larsuit's the simplest-possible solution for now11:36
larsuWellark promised to look into it more deeply11:37
didrockslarsu: do you think Wellark review that branch as well?11:37
pittilarsu: oh, that won't retry in a tight loop?11:37
didrocksto get it landed, image respin and so on?11:37
pittilarsu: or it will, and just make it less CPU intense?11:38
larsupitti: it won't11:38
larsuwell, it doesn't for me11:38
pittinice11:38
* pitti hugs larsu11:38
seb128larsu, yes, please, don't spend days fixing their codebase, just let that to them11:38
larsulet's see and wait if this actually works :)11:38
larsuseb128: ;)11:38
pittiin any case, this patch is obviously correct as we don't have a .service; as for whether it is sufficient, I don't know11:39
larsudidrocks: no need, this patch is definitely needed11:39
didrocksnice work larsu!11:39
pittieventually it should still use g_bus_watch_name and connect/disconnect asynchronously without polling, but I guess that's for someone else to do :)11:39
larsupitti: yes. I was starting out doing that, but it was getting complicated quickly11:40
larsumr is here: https://code.launchpad.net/~larsu/indicator-network/dont-activate-ofono/+merge/19432311:41
GunnarHjLaney: Saw your question about seeding of ttf-wqy-microhei. Guess it depends on why it's seeded. As you know, we have pkg_depends in any case.11:47
LaneyTo have a chance of CJK working by default I guess11:47
GunnarHjLaney: How does pkg_depends not make CJK work by default?11:48
LaneyYou have to run l-s, right?11:50
LaneyIt's for everyone to have a CJK font all the time11:50
GunnarHjLaney: l-s is used for installing languages anyway. Not sure off-hand is you select a CJK language at installation, but have a feeling that the installer makes use of pgk_depends as well.11:52
LaneyI wouldn't get them on my en_GB installation though, even if it did11:53
GunnarHjLaney: Aha, you mean when e.g. displaying a Chinese web page?11:54
LaneySomething like that11:54
LaneyGuessing as to why it's installed by default, it's obviously from way before my time11:55
GunnarHjLaney: Possibly we should then start seeding fonts-droid instead. But I suppose we'd better let some time go to evaluate fonts-droid before messing with the seed. happyaron indicated yesterday that there might be a problem with libreoffice...11:56
LaneyAnd it's not in main, see my comment on the bug11:57
LaneyNot sure what that means for pkg_depends.11:57
LaneyIt's probably alright, I bet other stuff there is in universe11:57
LaneyStill means that Chinese users will have microhei installed, which might be bad for them, not sure11:58
LaneyYou can probably work with happyaron on this stuff now. Wasn't them most enjoyable project for me. :P11:58
Laneys/them/the/11:58
GunnarHjLaney: fonts-android is in main, and that source package carries fonts-droid nowadays.11:58
LaneyGunnarHj: run this: rmadison -S -s trusty fonts-android11:59
GunnarHjLaney: Run it. :( Then, what does main mean at https://launchpad.net/ubuntu/+source/fonts-android? That it's in main in Debian?12:02
LaneyYeah12:03
LaneyLook at the table below12:03
LaneyProbably not so hard to get it MIRed, mind you12:03
LaneyIt's on touch and there's a plan to get all of that stuff in main eventually ...12:04
GunnarHjLaney: See it; my mistake.12:04
LaneySo it'll happen at some point, but you could help make it happen sooner if you care to12:04
GunnarHjLaney: No, it shouldn't be hard to get it MIRed, especially since fonts-droid is pulled when you install u-s-s. ;-)12:05
Laneyyeah, that's the touch thing12:05
Laneyit builds from universe currently12:05
GunnarHjLaney: Will file a MIR application.12:05
Laneynice, thanks12:05
happyaron" Sorry, there was a problem connecting to the Launchpad server. "12:06
Laneyf5 f5 f512:06
=== MacSlow is now known as MacSlow|lunch
seb128Sweetshark, can you tell me if my comment on https://bugs.launchpad.net/ubuntu/raring/+source/libreoffice/+bug/1204449 is correct (e.g no sponsoring needed, that got uploaded just without listing the bug in the changelog)?12:12
ubot2Launchpad bug 1204449 in libreoffice (Ubuntu Raring) "[SRU] LibreOffice 4.0.4 for Ubuntu 13.04 (raring)" [Low,Fix committed]12:12
didrockslarsu: confirmed to work for me! nice ;)12:13
didrocksthanks again12:14
larsudidrocks: pleasure. Wellark is on it to fix this more robustly12:15
seb128once again larsu is saving the day12:15
didrocksyep ;)12:15
=== tjaalton_ is now known as tjaalton
Sweetsharkseb128: checking12:25
Sweetsharkseb128: yes, comment is correct. Seems to be still in proposed though, so not setting to 'fix released' ...12:27
seb128Sweetshark, right12:29
seb128speaking of SRU verification12:29
seb128Sweetshark, could you verify https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1207057? or should we ping jmp&co about it?12:30
ubot2Launchpad bug 1207057 in libreoffice (Ubuntu Precise) "presentation causes system to hang" [Undecided,Fix committed]12:30
seb128Sweetshark, the quantal SRU is waiting on 3 bugfix verifications as well12:31
Sweetsharkseb128: On my machine that wasnt completely reproducable from the start. I got a delay, no hang. The bugfix reduced the delay.12:32
seb128Sweetshark, did we have anybody able to reproduce the hang?12:32
seb128dobey, hey, could you review/sponsor the precise patch for https://bugs.launchpad.net/ubuntu/+source/software-center/+bug/926763 ?12:43
ubot2Launchpad bug 926763 in software-center (Ubuntu Precise) "Cannot install local packages (.deb files) without network connection (offline)" [Medium,Confirmed]12:43
Sweetsharkseb128: bug 1246583 seems serious, but as per comment 9 seems to be not a LibreOffice issue. Any hints on whom to get involved?12:49
ubot2Launchpad bug 1246583 in libreoffice (Ubuntu) "All hotkeys of LibreOffice don't work in non-English keyboard layout [ubuntu 13.10]" [Medium,Confirmed] https://launchpad.net/bugs/124658312:49
=== alan_g is now known as alan_g|lunch
seb128Sweetshark, attente12:49
seb128hum13:02
seb128attente, the good news is that this issue is not due to you ;-)13:02
seb128same problem on f20 (/me just tested in a vm)13:02
seb128Sweetshark, ^13:02
seb128so not an appmenu/unity issue13:03
Sweetsharkseb128: urgh.13:03
Sweetshark"also an issue with KingSoft" sounds like its "somewhere in the stack" not LO itself though ...13:04
seb128yeah, could be gnome-settings-daemon or gtk213:04
Sweetsharktime to arm the needle in haystack detector ...13:04
seb128Sweetshark, https://bugzilla.redhat.com/show_bug.cgi?id=95830013:05
ubot2bugzilla.redhat.com bug 958300 in libreoffice "Keyboard shortcuts doesn't work in Gnome if a non-US keyboard layout is active" [Unspecified,New]13:05
seb128no useful info there though13:05
Sweetsharkseb128: if there are any quantal LibreOffice SRUs, we should just drop them unless there is a customer on them. IIRC quantal is EOL in two month and everyone caring about those fixes has switched to the LO ppa a long time ago anyway.13:08
Sweetsharkseb128: copy paste work on a german layout here ...13:09
seb128Sweetshark, well the current one is in the queue, we could as well get it moved to updates13:09
seb128Sweetshark, german is an ascii layout, C,V are on the same position than the english layout13:10
Laneyquantal has until april13:10
SweetsharkLaney: ah, right. It was the last 'longer' one.13:14
Sweetsharkseb128: iteresting *cough*13:15
Sweetsharkseb128: I just added Bengali -- an yep: Copypaste doesnt work there. Then I tried to switch to a 'Guest session' ....13:16
seb128Sweetshark, I tried with Russian13:16
seb128Sweetshark, I can confirm it doesn't work13:16
Sweetshark... which repeatedly popped up 'IBus-Update' message on an otherwise black screen with a classic 1990 X11 cursor. I only got out of that with ctrl-alt-f2 and killing the second lightdm process.13:17
seb128Sweetshark, weird13:18
seb128Sweetshark, https://bugs.freedesktop.org/show_bug.cgi?id=5558513:20
ubot2Freedesktop bug 55585 in UI "Should check all XKB group indexes when matching key events for accelerators and mnemonics" [Normal,Needinfo]13:20
Sweetsharkseb128: so a XFCE Guest session with Bengali does copy-paste as expected, so this is somewhere in the gnome3 stack ...13:23
seb128Sweetshark, right, see the libreoffice bug report13:24
seb128Sweetshark, it describes the issue13:24
=== MacSlow|lunch is now known as MacSlow
Sweetsharkseb128: yep13:42
seb128Sweetshark, upstream GNOME doesn't plan to change their stuff, their position is that libreoffice needs to be fixed13:42
seb128Sweetshark, I'm still pondering if we should rollback to the old way to handle keyboards or if we should try to push thing forward before the LTS13:43
attenteseb128, yeah, i'm not sure what in the gnome stack could have caused that issue13:43
seb128attente, not your fault in any case13:44
seb128attente, good morning btw ;-)13:44
seb128attente, the first comment on https://bugs.freedesktop.org/show_bug.cgi?id=55585 describes it well13:44
ubot2Freedesktop bug 55585 in UI "Should check all XKB group indexes when matching key events for accelerators and mnemonics" [Normal,Resolved: duplicate]13:44
attenteSweetshark, i wonder how it was working in libreoffice before, do you know?13:46
=== alan_g|lunch is now known as alan_g
attenteseb128, good afternoon :)13:47
seb128attente, read the first comment on that bug13:50
seb128attente, in the old GNOME world xkb groups were handled differently13:50
attenteoh... right13:51
dobeyseb128: i'll try to look at it today14:11
seb128dobey, thanks14:12
Laneyseb128: do you think we need a session on gnome plans to assign work items or do people generally know what to do?14:29
Laney(or did you make one already?)14:29
seb128Laney, I'm not sure there is much to discuss14:30
Laneyturning it into work items14:30
seb128we would need everybody that is going to do work to join at the same time14:31
seb128my bet is that e.g robert_ancell is not going to wake up at 4am for that14:32
seb128we should have a blueprint in any case14:32
seb128I'm just not sure it's worth a session14:32
Laneyokay, if it is going to be assigned another way14:33
seb128Laney, I would suggest "create a blueprint -> dump a list of everything we know about that needs done in it -> chase people to grab workitems -> get me to assign the remaining one"14:36
seb128or something along those lines14:36
seb128Laney, though there is not so much desktop work, stuff like "bring back menus" are already tracked through bugs14:36
Laneyis that enough?14:39
Laneyif so, fine, maybe we shouldn't bother14:39
seb128do you have example of work you want lined up?14:40
seb128it feels we already discussed what to update and strategy etc14:40
seb128cf my email to the desktop list14:40
LaneyI think we know what the plan is, but aren't tracking the work to achieve it14:40
seb128it feels like a session would turn up into a "not a lot to discuss" or in "everybody lists their pet bugs"14:41
seb128hum, right14:42
seb128Laney, hum, assuming that times are UTC ones, it means sessions are from 2am to 8am for .au guys14:47
seb128Laney, I need to check with robert_ancell if he can/want to me online for a desktop workitems session14:47
LaneyCan just fill in a BP outside of UDS14:47
seb128right14:48
=== Maple__ is now known as Guest23765
=== slomo_ is now known as slomo
=== Guest23765 is now known as Mapley
seb128chrisccoulson, hey, do you still look at firefox issues? ;-)14:54
seb128chrisccoulson, do you know if https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1200272 is being worked/if it's an hunspell or firefox issue?14:54
ubot2Launchpad bug 1200272 in firefox (Ubuntu) "firefox crashed with SIGSEGV" [Medium,Confirmed]14:54
seb128jdstrand, mdeslaur, there are some apparmor changes in the sponsoring queue, do you guys plan to handle those? (some are debdiffs from Tyler)15:05
chrisccoulsonseb128, i haven't looked at a firefox issue for ages ;)15:06
mdeslaurseb128: yeah, we'll handle those, thanks15:06
seb128mdeslaur, should I unsubscribe sponsors so others don't spend time reviewing?15:06
mdeslaurseb128: sure15:07
mdeslaurseb128: could you give me the bug # so we don't drop them though?15:07
seb128mdeslaur, sure15:07
seb128mdeslaur, https://code.launchpad.net/~cmiller/ubuntu/trusty/apparmor/chromium-new-sandbox-name/+merge/193657 is one15:07
seb128mdeslaur, seems like we should get that one in, the current chromium already has the rename ... does it mean the apparmor profile is unactive?15:08
seb128mdeslaur, bug #1231778 is on the with the diffs from Tyler15:09
ubot2Launchpad bug 1231778 in AppArmor "wifi not working on Saucy Salamander" [Medium,In progress] https://launchpad.net/bugs/123177815:09
Laneybah15:10
Laneywayland testsuite fails15:10
Laneytjaalton: you like wayland? fancy looking at the updates? ;-)15:10
seb128timchen119, hey, thanks for the notify-osd work!15:13
mdeslaurseb128: thanks15:14
seb128mdeslaur, thank you for looking at those ;-)15:15
sil2100kenvandine: ping :)15:17
kenvandinesil2100, pong15:18
tjaaltonLaney: alrighty15:18
=== alan_g is now known as alan_g|tea
Laneytjaalton: cool (blocks new gtk)15:18
Laneythanks15:18
tjaaltonLaney: I guess wayland could be synced15:23
tjaaltonpushed the Conflicts/Replaces diff to debian earlier, 1.3.0-1 has it15:24
Laneytjaalton: oh, in cursor0 too?15:24
seb128happyaron, hey15:25
tjaaltonLaney: yup15:25
=== alan_g|tea is now known as alan_g
Laneytjaalton: whatever you think is right15:49
tjaaltonLaney: synced15:52
Laneywill need weston too to migrate to trusty15:53
LaneyIIRC15:53
tjaaltonweston needs cairo-gl, which then explodes on nvidia15:54
tjaaltonprobably could just disable building weston-screensaver instead15:54
tjaaltonbut wayland should be fine without weston15:55
Laneylibwayland-server0 Breaks: weston (<< 1.2.0)15:56
tjaaltonah yes15:57
=== MalcontentMatt is now known as MattJ
=== MattJ is now known as Guest51257
tjaaltonstupid unity doesn't know meld is open16:00
=== Guest51257 is now known as Johnson
=== Johnson is now known as mjohnson15
=== gatox is now known as gatox_lunch
tjaaltondoesn't look like it's easy to disable wscreensaver, it's everywhere16:04
seb128drop weston from the archive?16:04
tjaaltonuntil cairo-gl works, maybe16:04
Laneyhow come debian can enable it?16:05
dpmtedg, there's a discussion going on on #ubuntu-app-devel on how upstart could deal with fat packages containing binaries for different architectures. If you're around, would you mind joining in? thanks!16:07
tjaaltonLaney: we've disabled cairo-gl in cairo16:07
Laneytjaalton: remoing/changing the --with-cairo= line isn't enough?16:08
tjaaltonbecause the nvidia blob doesn't like it16:08
tjaaltonoh..16:08
tjaaltonI'll try16:08
tjaaltonwas grepping the wrong thing16:08
=== mjohnson15 is now known as register
=== register is now known as mjohnson15
Laneyhow can you tell if it's bad?16:11
Laneyjust try it on nvidia?16:11
tjaaltonyes, every gl client will use a lot of extra memory16:13
Laneyseems to build without that and removing -screensaver from .install16:13
Laneygo for it if that's good enough for you16:13
happyaronseb128: hi16:13
seb128happyaron, hey16:14
happyaronseb128: what's up?16:14
tjaaltonLaney: yeah good enough for me16:14
seb128happyaron, I'm trying to figure what to do with the keyboard situation for the LTS16:14
Laneytjaalton: right, feel free to upload then16:14
Laneyshould unblock gtk16:14
seb128happyaron, do you know how happy were IMs users with the old way of doing things, e.g having layouts and IMs being separate concepts rather than grouped like GNOME did (and we us now)?16:15
happyaronseb128: actually we don't like the current way in any extent, and very few people like what Windows8 do (super+space)16:16
seb128happyaron, I'm trying to decide on whetever we should push forward or rollback for the LTS and decide the old way is the best we can do with xorg16:16
seb128happyaron, then have a better go at solving the issue with unity8/mir16:16
seb128happyaron, how often do you guys change layouts? e.g in the old world, did you interact with the layouts or only directly with ibus/fcitx?16:17
happyaronseb128: only ibus/fcitx, and it handles layout16:18
seb128happyaron, I'm not even sure at this point what the GNOME guys are trying to solve with the new design...16:18
seb128if the old ibus way was handling layouts as well16:18
happyaronseb128: ibus's xkb integration isn't good, and they wish to have the feature at that time, so they come up with current solution16:19
happyaronibus-xkbc is almost abandoned upstream, and there are still compatibility issues with pure xkb settings16:19
seb128happyaron, what about fcitx?16:22
happyaronfcitx has the support in the main upstream project, and has some compatibility issues too.16:22
seb128happyaron, but do you need xkb integration? or do you just use e.g qwerty-english or qwerty+pinyin16:22
happyaronno, I don't.16:22
seb128I mean typically16:22
seb128so those are basically 2 different usecases16:23
seb128layouts are for people who e.g type russian or english and change between those16:23
happyaronI think few people need it, and when they really need it they could already have their own xkb configs.16:23
seb128IMs are for people who use 1 layout but need to change the way composition works?16:23
happyaronyes16:24
happyaronfor example I need to input Chinese characters, but I have many ways inputing it, use PinYin or Wubi, I need switch between them as I wish16:25
happyaronand I also need to switch to English keyboard at times to input alphabets.16:26
seb128happyaron, well, you are always on qwerty in all those cases though?16:27
happyaronyes16:27
seb128happyaron, e.g typically you would only need to super-space with ibus16:27
seb128ok16:27
seb128happyaron, so it seems you +1 for the old way, e.g reverting would be an issue for you16:27
seb128wouldn't*16:27
seb128(sorry)16:27
happyarontypically ctrl+space, super-space is wired...16:27
seb128yeah, I just mean $favorite_keys16:28
happyaron:)16:28
seb128happyaron, great16:28
happyaronthanks16:28
seb128I'm going to start some list discussions on reverting or going forward and put a vUDS session about it16:28
happyarongreat16:28
=== gatox_lunch is now known as gatox
colonelqubittkamppeter_: Bjoern (Sweeetshark) mentioned that you might be able to help LibreOffice triage a couple of printing bugs17:23
colonelqubit(We don't have access to particular printing hardware)17:25
* didrocks waves good evening, see you on Tuesday!17:27
seb128sil2100, cyphermox: can we get a 13.10 SRU for indicator-application (there is 1 commit in there and it fixes a frequently reported issued)?17:49
cyphermoxseb128: sure, I'll kick it off now17:49
seb128cyphermox, thanks17:49
cyphermoxI'm not staying around though, I'm just on IRC by accident :)17:49
seb128cyphermox, have fun!17:50
=== alan_g is now known as alan_g|EOD
cyphermoxseb128: where you editing the bug?17:59
cyphermoxseb128: build is running now, I'll check back on it later to push it to proposed18:01
seb128cyphermox, no I didn't, sorry was away for exercise, I can do it if you want, testcase is basically "watch if e.u.c reports stop coming"18:29
bschaeferattente, hey, were you able to get any callbacks from compiz?18:45
attentebschaefer, still no, but it's down to a modifier checking problem18:46
bschaeferattente, well hopefully thats good, so compiz is getting the action and everything? Just failing on a modifier not being set?18:47
attentecompiz gets the action, compiz does the grab18:48
attentewhen time comes to check for the action which matches the event, it flops18:48
attenteit seems to be dependent on what keybinding i provide too18:48
bschaeferhmm interesting, i've never actually had to follow that code before...is under one of the trigger* functions?18:49
attenteyeah, i added some fprintfs to PrivateScreen::triggerKeyPressBindings18:49
attentein some cases like <Shift>F8, it gets a matching keycode, but the modifiers don't match18:50
bschaeferwhen you say it seems to be dependent on what keybinding you provide, what do you mean by that i suppose?18:50
attentein other cases, like <Control>p, it doesn't seem to match keycodes properly18:51
bschaeferhmm strange18:51
bschaeferwhat if you took an already working hot key18:51
bschaeferand tried use that?18:51
bschaefersuch as Alt+Tab18:51
bschaeferhmm18:51
bschaeferattente, so its never returning true here:       if (match && eventManager.triggerPress (action, state, arguments))18:52
bschaefer?18:52
bschaeferattente, also, try again Shift+F8, cause you added in the init state after you tried that18:53
attentebschaefer, basically match is never even true18:53
* bschaefer looks for how it parses the key bindings18:53
bschaeferattente, looking at how its parsed in...i think you may be adding the hot key in, incorrectly18:56
bschaefer    binding.fromString(impl::CreateActionString(key, shortcut, flag));18:57
bschaefersomething like that, is whats done that i see in unityshell.cpp18:57
bschaeferattente, and looking at your code, you just use: action->keyFromString(accelerator);18:57
bschaefermy guess is something is not producing the correct tokens for compiz to check18:58
bschaeferattente, a place to check would be in compiz/src/action.cpp::CompAction::KeyBinding::fromString (const CompString &str)18:59
attentebschaefer, looking at that function, it seems to be going by '<Modifier><Modifier>x_keysym_name'19:16
bschaeferattente, hmm well possibly try to use impl::CreateActionString(key, shortcut, flag)19:18
bschaeferto generate a string then pass it to binding.fromString()19:18
bschaeferattente, as it seems add that its getting wrong keycodes/keysyms...19:19
attentebschaefer, for it to make the grabs properly, it must've parsed them correctly, no?19:20
bschaeferattente, hmm how were you testing if the grabs were set up properly?19:21
bschaeferattente, and yes that would hmmm19:21
bschaeferattente, so the keycode from the current event isnt matching up with the keygrab?19:21
attentePrivateScreen::triggerKeyPressBindings was getting called only after i added the action to the screen19:22
bschaeferattente, hmm IIRC triggerKeyPressBindings is done on all keypresses...from XNextEvent...19:23
bschaeferline 846 in event.cpp19:23
bschaeferbut if theres not keygrab set it i don't think the event would even get there hmm19:25
attentebschaefer, doesn't seem to be the case19:28
attentei can press most key combinations, and that doesn't get called19:29
bschaeferyeah, the grab has to be there, so yeah you're right that the grab is being set hmm19:29
attenteso, on line 420 in event.cpp19:30
attentebindMods is 819:31
attentesorry, 0x819:31
attenteevent->state is 0x1119:32
attentemodMask is 0x20000ed19:32
attentethis is for '<Shift>F8'19:32
attenteas far as i can tell, event->state is correct19:33
bschaeferhmm19:33
bschaeferattente, well19:33
bschaeferbindMods wont be equal19:34
bschaeferattente, what happens when you press say, alt+tab?19:34
bschaeferwhat is modMods, bindMods?19:34
bschaeferor something similar that compiz already has set up19:34
attentefor alt+tab, i get 0x8 0x18 and 0x20000ed respectively, which matches because of the mask19:35
bschaeferyeah19:35
bschaeferattente, possibly change a hot key in unity shell plugin in ccsm to but like shift+F719:36
bschaeferthat way you can get the mod and mask is getting to through there, our it'll fail19:36
bschaeferand then its not your fault but compizs19:36
bschaeferbut I just set up Shift+F1, and it seems to work hmm19:37
attenteif i try '<Control>n', then it doesn't even seem to match the keycode19:38
attenteok, i might be being stupid here, let me check something19:39
bschaeferattente, well i mean change a hotkey in ccsm to that of a similar hot key you are trying to match up (Such as changing Alt+Tab to Shift+F7)19:39
attentebschaefer, where's the Alt+Tab option?19:41
attentei can't seem to find it19:41
bschaeferattente, under ccsm-> Unity Shell Plugin -> Switcher19:41
attentebschaefer, ok, so19:44
bschaeferthat way you can see what the correct values should be19:44
attentethat worked19:44
attentei set Shift+F12 for the switcher19:44
attentethe correct mods are the same19:44
bschaeferreally?19:44
attente8 18 20000ed19:44
bschaefero that 18 part is different though19:44
attenteyeah, so something is clearly wrong on my side19:44
bschaeferyou said19:44
bschaefer event->state is 0x11, that should 0x1819:44
attentethe 18 is fine since the 1 gets masked out19:44
bschaeferso the state is getting set to an incorrect value19:45
attenteright, when ccsm does it, it works19:45
attentewhen i try to do it manually, it doesn't19:45
bschaeferyeah, hmm i wonder where that state gets set19:45
attentethe state is direct from the XEvent19:45
attenteoh..19:45
attenteweird..19:45
attentebecause i thought 0x11 was correct19:45
bschaeferyeah...why is it 0x11 for you?19:46
attenteand 0x8 was wrong19:46
bschaeferand 0x18 for ccsm?19:46
attentewhoops, sorry19:46
attentei read the wrong line19:46
attenteShift+F12 via ccsm is 1 11 20000ed19:46
bschaeferoo19:46
attentethat makes more sense19:46
bschaeferyeah19:46
bschaeferso the bindMods for you is incorrect19:47
attenteyep19:47
attentei'll see what ccsm is using for the string19:47
bschaefertake a look at this: action->key ().modifiers ()19:47
bschaeferprint that out for your shift+f8 and ccsms shift+f1219:48
bschaeferas im  guessing the value of that is whats inconsistent between the 2 cases19:48
attentebschaefer, sorry i'm taking so long, i'm getting overwhelmed with the debug output19:56
bschaeferattente, o no worries19:57
bschaeferattente, compiz is never easy to dig through, same with the x event loop19:57
bschaeferplus i've never actually dug through this part of compiz, so im just talking a bunch19:58
attentebschaefer, one thing i can say is that ccsm is not doing it by parsing a CompString19:59
bschaeferhmm let me take a look at the generated code20:00
attenteoh man. i have no idea what's going on here, lol20:00
bschaeferhaha, yeah neither do i :)20:00
bschaeferthey seem to be:     action.keyFromString ("<Alt>Tab");20:00
bschaeferthose are the default values though20:00
bschaeferattente, did you set this state:     state = CompAction::StateAutoGrab;?20:01
attentebschaefer, no, what does it do?20:01
bschaefernot sure, but all those hotkeys have it being set20:01
bschaeferthey are setting StateAutoGrab | StateInitKey20:01
bschaefersoo possibly that changes how its being grabbed, which could cause a different bindMod?20:02
attenteaction->key ().modifiers () is 0x1 for the ccsm case, 0x10000 for our case20:10
attentegoing to see what that StateAutoGrab does now20:10
bschaeferhmm i wonder why then its getting a different bindMod20:13
attenteyeah, no effect there20:14
bschaeferattente, well a funny thing to do would be, detect 0x10000, and change it to 0x120:15
bschaeferto confirm thats the only problem, and you should (hopefully) get a callback20:15
bschaefernow i wonder why .modifers is 1 for ccsm and 65536 for our case20:16
bschaefer65536 seems like a normal modifier...20:17
attentefunny thing is, if i do an alt+`, the action->key ().modifiers () is also 0x1000020:17
bschaeferhaha, try alt+f820:17
bschaeferi wonder if its getting alt?20:18
bschaeferinstead of a shift20:18
attenteoh. that already binds to something20:18
attentewindow resizing20:19
bschaeferbut you get the same value?20:19
attenteyes20:19
GunnarHjattente: Hi!20:19
attenteGunnarHj, hello!20:19
bschaefervery strange...now why is it turning your <Shift>F8 to <Alt>F820:19
GunnarHjattente: Do you think that the suggestion in bug 1248349 can help fix the language list issue in u-s-s?20:19
ubot2Launchpad bug 1248349 in ubuntu-system-settings (Ubuntu) "[language] Confusing list of display language options" [Low,Triaged] https://launchpad.net/bugs/124834920:19
attentebschaefer, can't seem to grab alt+f9, very weird problem indeed20:21
attenteGunnarHj, sorry, taking a look now20:22
attenteGunnarHj, ah20:23
seb128GunnarHj, didn't Laney said it should already be fixed in trunk?20:23
attenteactually we had this discussion before20:23
Laneyit's different in trunk20:23
Laneydon't know if you'd call it fixed20:23
Laneypitti and attente discussed how to get the list some weeks ago20:23
seb128well, we reduced the variants20:24
attentethe problem is u-s-s also uses that Display Languages for setting the formats locale as well20:24
seb128it was crazy before20:24
seb128we still have some though20:24
GunnarHjseb128: When testing a build of the trunk, it got worse (on the desktop).20:24
seb128GunnarHj, you clearly didn't test before20:25
seb128we had like every language and variant existing in the world earlier20:25
seb128it got better for sure ;-)20:25
GunnarHjseb128: True, I didn't. But it still seems to be pretty bad.20:26
seb128it's quite ok20:26
seb128not great, but good enough (well at least usable)20:26
GunnarHjseb128: What's wrong with doing it right?20:26
bschaeferattente, indeed very strange...20:26
GunnarHjseb128: You triaged the bug, btw. ;-)20:27
seb128GunnarHj, nothing, I agree it could be better, I just disagree with the way you undermine the work which has been currently done20:27
LaneyI think the way l-s displays the options is nicer20:27
seb128it is20:27
Laneybut our way isn't that bad20:27
GunnarHjseb128: I really don't try to undermine anything! Just trying to help improve.20:28
Laneythat's Language (Variant) [don't know the proper terms]20:28
seb128GunnarHj, so focus on what should be done rather than trying to show how buggy you find the current solution ;-)20:28
GunnarHjseb128: Of course.20:28
GunnarHjattente: Do you mean that languages and regional formats are not set separately?20:31
attenteGunnarHj, yes20:31
Laneyhttps://wiki.ubuntu.com/LanguageAndText#phone20:32
GunnarHjLaney: tnx20:32
LaneyIt doesn't say that, but there's only one option for language so it's used to set both20:33
seb128I don't think it makes much sense to have different options20:34
GunnarHjattente: Well, in that case a simple 'locale -a' based list seems to be appropriate. I thought its only purpose was to select the display language, which is why I proposed the language-tools script.20:34
seb128the current list seems a bit too verbose20:35
LaneyDisplay language might not be the best title in that case20:35
attenteGunnarHj, thanks20:35
seb128like it has English-Ireland20:35
seb128how is that different from English?20:35
GunnarHjLaney: Agree that the title should be changed.20:36
LaneyProbably in formatting20:36
attentedoes ireland use different formats?20:36
seb128I don't know20:36
seb128Laney might know20:36
Laneyhaha20:36
LaneyI do not20:36
seb128:/20:36
seb128we need cjwatson :p20:36
seb128or Riddell20:37
attentei guess they don't like being grouped with the UK...20:37
LaneySome people might have a problem being force to set en_GB though20:37
Laney...20:37
seb128Laney, you UK guys should know about the different between your countries :p20:37
LaneyIreland isn't in the UK20:37
attenteyeah. i know :P20:37
LaneySo yeah, it does make sense20:38
LaneyI think it'd look a bit less jumbly if we used the same scheme as l-s20:38
seb128Laney, North Ireland is no?20:38
Laneythat's probably not en_IE20:38
GunnarHjseb128: Just to clarify: Considering that it's considered sensible to distinguish between display language and formats on the desktop, why isn't that true for the phone?20:38
seb128GunnarHj, what are win8, iOS, android doing in that regard?20:39
seb128GunnarHj, phone tend to be less "geeky" than computer and we can do with some simplification there20:39
GunnarHjseb128: I don't own a phone, so I have no idea. ;-)20:40
seb128my android (samsung) seems similar to what we have20:40
seb128english20:40
seb128- australia20:40
seb128- ireland20:40
seb128- new zealand20:40
seb128- south africa20:40
seb128etc20:40
GunnarHjseb128: Ok.20:40
GunnarHjGuess I'm a little biased on this matter...20:41
Laneyyeah that's pretty alright20:41
LaneyIt's basically saying that you can't choose them separately20:41
LaneyI spy a ubuntu-tweak-tool on the horizon20:42
seb128Laney, sorry, I mean all those are all options20:42
Laneyyep20:42
LaneyI mean formats and language20:42
seb128right20:42
seb128well, "format" always seemed weird to me20:42
seb128I'm not sure what format people tweak20:42
GunnarHjLaney: And if that's the design decision, I'd better close the bug.20:43
seb128I never felt the need to tweak how numbers are displayed or such20:43
attentewe could've just said locale i suppose20:43
LaneyI dunno20:43
LaneyMaybe if you live in another country?20:43
LaneyHard to guess what people use things for20:43
seb128GunnarHj, I'm not sure we have enough input from device to device20:43
seb128I'm not even clear what we are missing with the current approach20:43
seb128or what it gives us over the "list only countries"20:43
GunnarHjseb128: What we miss is an opportunity for people who lives in one country, but prefer the display language of some other country, to set date/time, number, currency etc. formats of their site.20:45
seb128GunnarHj, there is always a balance between standard needs and UI complexity20:47
seb128not sure how common those options are on a phone20:48
GunnarHjseb128: Indeed. Not saying that there is a right solution here.20:48
seb128they are useful on a desktop, but that's a more powerful/work oriented environment20:48
GunnarHjseb128: Agreed.20:49
seb128I think we should keep the bug and get input from mpt/design20:49
GunnarHjseb128: Ok. Then I won't close it.20:49
seb128thanks20:49
=== thumper is now known as thumper-hacking
=== tkamppeter_ is now known as tkamppeter
tkamppetercolonelqubit, yes, I could help you with testing. Please give me links to the bug reports.22:36
=== thumper-hacking is now known as thumper-gym

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!