/srv/irclogs.ubuntu.com/2015/08/27/#ubuntu-unity.txt

tsdgeosmzanetti: ota6 is out? can we prepare the landing of all the big branches?07:29
tsdgeosmzanetti: pstolowski: did you land unity8 in silo 44?07:33
pstolowskitsdgeos, i marked it ready for qa, it hasn't landed yet07:35
tsdgeosah ok07:35
tsdgeosmzanetti: is tagger you thing?07:40
tsdgeosseems it is07:41
mzanettitsdgeos, yes08:55
mzanettitsdgeos, can you check out this one when you have some time? https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/148930908:56
ubot5Ubuntu bug 1489309 in unity8 (Ubuntu) "Names in Favorite Contacts are cut at the bottom" [Undecided,New]08:56
mzanettitsdgeos, what up with tagger?08:56
tsdgeosso today i've found the first use case in history for it08:56
mzanettihahaha08:56
tsdgeosthough i guess our "security model" won't allow it08:56
mzanettiand it failed08:56
mzanetti:D08:56
mzanettiobviously08:56
tsdgeosmy router has a tag beind it08:56
mzanettitell me08:56
tsdgeosthat you scan it08:56
mzanettiright08:56
mzanettiwifi08:56
tsdgeosand if you have an android or iphone will connect to that wifi with the password08:56
tsdgeosi got the text that i had to copy&paste08:57
tsdgeoswas not horrible08:57
mzanettiI had that working on the N9[00|50]08:57
mzanettibut yes, apparmor08:57
tsdgeosyeah08:57
tsdgeosi said08:57
tsdgeossecure!08:57
tsdgeosit echo'ed trough the empty rooms of my apartment :D08:57
mzanetti:)08:58
mzanettiI'll start a mail thread... wanted to add this feature for a while already08:59
tsdgeosanyhow, do you think it' worth opening a bug against tagger+something else?08:59
mzanettiyes09:00
mzanettiI just need to highlight that the format how the wifi information is stored in the code was invented by google. that seems to confinve most people that it must be a good thing :D09:00
tsdgeoshmmm09:01
tsdgeosmzanetti: that's the today scope?09:01
mzanettitsdgeos, yes09:01
tsdgeosi don't see that carousel09:01
tsdgeoshome many contacts in favorites do i need?09:02
tsdgeosi've 409:02
tsdgeoslet's see if rebooting helps09:02
mzanettino clue... I've just one09:02
tsdgeosand you see it in the today scope?09:02
mzanettiyes, but not as carousel09:02
mzanettimaybe you need to enable it in the settings of the scope09:02
tsdgeosi don't have anyting09:02
tsdgeosi have it enabled09:02
mzanettiup-to-date?09:03
mzanettiaparently there was an update to the today scope the other day09:03
mzanettioh, I didn't mean to enable the scope, but the category in the scope. "Show entries from Contacts" or so09:03
tsdgeosyes yes09:04
tsdgeosah, i needed to press the "skip and setup later"09:05
tsdgeosright i don't have a carousel either09:05
tsdgeoslet me check the code i think we have some carousel auto conversion to list if < X09:05
tsdgeoshmmm not really, we fallback to grid09:07
tsdgeoswhich is weird09:07
tsdgeosmaybe it's the scope doing the fallback09:07
* tsdgeos adds more favourites09:07
dandradergreyback, this is the qtmir one-liner I need you to review: https://code.launchpad.net/~dandrader/qtmir/resizeSuspending/+merge/26920709:34
tsdgeospstolowski: do you know where the code for the today scope/contacts subsection is?09:34
tsdgeosor who could know?09:35
pstolowskitsdgeos, cwayne?09:35
greybackdandrader: ok, LP reporting merge conflict tho09:35
tsdgeospstolowski: k, tx09:36
dandradergreyback, saw that. seems false alarm. it's because of the mirSurface prerequisite09:36
greybackdandrader: okay09:36
tsdgeosmzanetti: i can't seem to get a carousel there09:37
mzanettitsdgeos, the reporter says he has 6 contacts09:37
mzanettiin favorites09:38
tsdgeosi've 16 :D09:41
tsdgeoslet me go back to 609:41
tsdgeosnah no carousel09:49
tsdgeosmaybe we killed it in ota609:49
tsdgeos?09:49
mzanettitsdgeos, nope. have the carousel here now09:52
mzanettitsdgeos, I just had to update the today scope tho09:52
tsdgeosupdate it on the store?09:52
tsdgeosor systemsettings :D09:53
tsdgeosouch09:54
tsdgeosi need my 2FA that i left at the other house09:54
mzanettitsdgeos, set up authenticator ;)09:54
mzanettitsdgeos, in systemsettings09:55
tsdgeosmzanetti: but i'll still need the 2FA to do that no?09:55
tsdgeosanyway i was planning to go back before lunch09:55
tsdgeosmay as well do it now09:55
mzanettitsdgeos, well, this is not totally critical09:55
tsdgeossure i understand09:57
tsdgeosstill i need to go back for food, so can do it now or in ~1 hour :D may as well time it with the bus09:58
mzanettidandrader, hey, saw your comment about the DeviceConfiguration10:47
dandradermzanetti, which of the two?10:47
mzanettidandrader, before he just had "property int ignoredMice: deviceConfiguration & deviceConfiguration.ignoredMice"10:48
mzanettidandrader, that was reaching out of context so I asked him to fix10:48
mzanettidandrader, the other option would be to pass a pointer to the deviceconfiguration (or an integer of the ignoredMice) around throughout the whole notifications system10:48
mzanettiwhat would be your favorite?10:49
mzanettiimo the singleton is quite nice for this... I also think we'll need deviceConfig stuff in more places in the near future10:50
mzanettibut you can obviously propose something nicer...10:50
mzanettiin regard to move vs delete/add, you're obviously right10:51
dandradermzanetti, maybe it could have the higher level info it wants, intead of going into the nasty details of checking how many mouse devices are available and how many should be ignored10:51
mzanettidandrader, yeah... thought about that too10:52
dandradermzanetti, I didn't properly review it. that's why it's just a comment withou a review decision10:52
mzanettibut then I'm not really sure what10:52
dandradermzanetti, but on a first look I didn't like this10:52
dandradermzanetti, is in it couples the Notificaions component to both DeviceInfo and DeviceConfiguration10:53
mzanettimhm10:53
mzanettibut passing those things through a chain of files kinda sucks too10:54
dandradermzanetti, but I would need some time to try to come up with a proposal10:54
mzanettiI'm not saying you're wrong... just that I'm not sure about something better right now10:54
mzanettiyeah..10:54
mzanettimaybe lukas has an idea10:54
tsdgeosmzanetti: so next silo has the big MRs of sdk13, quick24?10:55
mzanettitsdgeos, no :(10:56
tsdgeosoh :/10:56
tsdgeoswhy?10:56
mzanettitsdgeos, didn't get the "ok, let's move to 1.3 yet"10:57
mzanettibut will poke people about it again now10:57
tsdgeosgrrr10:57
tsdgeosok, correct, the updated today scope gives me a carousel10:57
tsdgeoswhich is weird since i remember design saying they didn't seem to like carousels anymore10:57
tsdgeosanyway off to debug11:00
dandradermzanetti, it's like sweeping the problem under the carpet11:03
dandradermzanetti, by accessing stuff over various singletons you're giving the illusion that your component is modular11:04
dandradermzanetti, "hey, I can just add Notifications {} to my shell and it just works!"11:04
mzanettidandrader, fair point, yes11:04
dandradermzanetti, whereas in reality it needs a bunch of info from the context that it's getting by other means11:05
dandradermzanetti, and if this context changes, it will break the component11:05
dandradermzanetti, not very diffent from accessing variables from some parent that are available in the context11:06
mzanettihmm... a bit better I'd say11:06
mzanettibecause if it's not there it'll fail to compile11:06
mzanettiwith the other it just breaks at runtime, maybe11:06
mzanettibut sure, I see your point11:07
dandradermzanetti, yes, it's better. that's why I said "not very different" instead of "the same as"11:07
dandradermzanetti, exposing the info in properties makes the dependencies obvious and exposes how modular you component truly is. hopefully making it easier to refactor APIs, so you could rearrange responsibilities to have a better, more modular, design11:11
=== MacSlow is now known as MacSlow|lunch
=== alan_g is now known as alan_g|lunch
tsdgeosfood!12:44
=== alan_g|lunch is now known as alan_g
mzanettidandrader, so... I've tested the mirSurface branch again, and the orientation thing is worse than with trunk. and it is one of the reasons why QA rejected it the first time.13:48
mzanettido you have a fix for that handy?13:48
mzanettigreyback, other than that, are you happy with the mirSurface branches?13:48
greybackmzanetti: yeah I am13:48
dandradermzanetti, lp:~dandrader/qtubuntu/resizeCatchUp and lp:~dandrader/qtmir/resizeSuspending13:49
mzanettigreyback, can you approve them ?13:49
greybackmzanetti: yep, going through them again now13:50
mzanettigreyback, please include those last two. I'll add them to the same silo13:51
greybackack13:51
=== MacSlow|lunch is now known as MacSlow
tsdgeosgreyback: ping14:12
greybacktsdgeos: pong14:17
tsdgeosgreyback: do you know what sets the mir host-socket to /run/mir_socket ?14:18
greybacktsdgeos: that is unity-system-compositor's socket, set in its upstart job14:18
greybacktsdgeos: the unity8 socket is $XDG_RUNTIME_DIR/mir_socket14:19
greybackalso set in its upstart file14:19
tsdgeosgreyback: can't find usc upstart job, know the name?14:21
Saviqtsdgeos, not upstarted, it's in the session14:27
Saviqtsdgeos, dpkg -L ubuntu-touch-session14:28
tsdgeosk tx14:28
Saviq/usr/share/ubuntu-touch-session/usc-wrapper is what launches usc14:28
greybacktsdgeos: what Saviq said, my apologies14:30
tsdgeossure, no worries14:30
greybacktsdgeos: if it's not set, mir does choose either /run/.. or /tmp/.. as the socket14:31
greybackdednick: ping!14:31
dednickgreyback: hey14:32
greybackdednick: hey, welcome back. When you're caught up, ping me14:32
greybackyou have a bunch of pre-holiday branches I've reviewed14:33
dednickgreyback: cool. i'll take a look and get back to you14:33
greybackdednick: cheers14:33
greybacktsdgeos: I've a theory on the unity8 restart problem. You're in the right direction with the socket being wrong.14:39
tsdgeosit is wrong14:40
tsdgeosi've verified that14:40
tsdgeosi *just* need to know where it gets wrong :D14:40
greybackI think it's possible for upstart to restart the unity8 job *before* the "stopping" phase of the previous unity8 job has run14:40
greybackas the stopping phase does the env var cleanup, it would mean the new unity8 gets old env vars14:41
tsdgeosgreyback: ted told me that's not possible, but might happen :D14:42
tedtsdgeos, I think I might have been wrong, it seems that post-stop isn't run on a restart: http://upstart.ubuntu.com/cookbook/#restart14:45
tsdgeosoh14:45
tsdgeosthat'd be bad14:45
tedtsdgeos, And probably more important to unity8, no pre-start or post-start14:45
dednickgreyback: hm. don't know what the hell happened with my touch tracing branch. looks like it didnt push properly...14:50
greybackdednick: uh oh14:50
dednickgreyback: i've re-pushed it. seems ok now.14:51
greybackcoolio14:51
dednickgreyback: what other branches did you review?14:51
greybackdednick: the polite app close stuff I think had open comments - and I suspect merging trunk will cause you pain14:52
dednickgreyback: oh hang on. i forgot i moved the touch tracing to unity-team. was looking at old one..14:52
dandradergreyback, could you please trigger a rebuild of qtmir in silo0?14:55
dandraderor can I do it myself?14:55
greybackdandrader: are you on the landing team?14:55
dandradergreyback, don't think so14:56
greybackI'd better do it then14:56
greybackdandrader: why does it need a rebuild? You've just added a debug output. If you're debugging, build it locally please14:57
greybackmzanetti: you tested the mirSurface stuff then? I just need to check the code again?14:59
dandradergreyback, because of this https://bugs.launchpad.net/ubuntu/+source/mir/+bug/1488417/comments/1315:02
ubot5Ubuntu bug 1488417 in qtmir (Ubuntu) "nexus4 window mode mouse motion bounded incorrectly" [Undecided,In progress]15:02
dandradergreyback, make it easier for others to reproduce15:03
dandradergreyback, and duflu has been denying it's a mir issue15:03
tsdgeosted: funnily that doc is wrong and they get executed15:04
tsdgeossince the echos i've added get printed15:04
tsdgeosconfusion!15:04
tedtsdgeos, Uhg, fun!15:05
greybackdandrader: reasons I hesitate are: (1) train will try to merge lp:qtmir into silo0, which fails as there's conflicts and I don't want to waste time on that (2) you're adding a debug flag to help you prove a point, when there should be a much easier way to reproduce the bug15:08
dandradergreyback, don't agree with the "there should be a much easier way to reproduce the bug" part15:10
dandradergreyback, but ok, let duflu build qtmir himself if he wants to follow those steps15:10
dandradergreyback, and this log should go to trunk at some point15:11
greybackdandrader: sure15:11
seb128ted, I opened mps for the indicators/unity8 https://code.launchpad.net/~seb128/indicator-session/unity8-system-settings https://code.launchpad.net/~seb128/indicator-bluetooth/unity8-system-settings https://code.launchpad.net/~seb128/indicator-power/unity8-system-settings https://code.launchpad.net/~seb128/indicator-datetime/unity8-system-settings https://code.launchpad.net/~seb128/indicator-sound/unity8-system-settings15:12
seb128ted, would be nice if you have a glance, would like review so I can fix today before being on vac if needed15:12
tedseb128, K, we should make sure charles is in the loop as well.15:13
charleswoo15:13
charlesin a meeting at the moment, will look at the mps rsn15:13
charlesseb128, LGTM, thanks for the patches15:22
seb128charles, yw! thanks for looking15:28
mzanettidandrader, greyback: this is tag-infected: https://code.launchpad.net/~dandrader/qtmir/resizeSuspending/+merge/26920715:30
mzanettiI cleaned qtmir trunk yesterday night15:31
mzanettistill not entirely clear how it is possible for those tags to jump repositories15:31
dandradermzanetti, it's a virus!15:32
mzanettiit's like the bird-virus starts infecting cats now...15:32
dandradermzanetti, cleaning up the tags now15:32
mzanettidandrader, tested the silo. IMO good. as soon as the last qtubuntu branch is approved I'll submit to QA15:32
tedMan, charles is too fast!15:36
dandradermzanetti, done the tag cleaning15:37
mzanettita15:37
dandradergreyback, updated https://code.launchpad.net/~dandrader/qtubuntu/resizeCatchUp/+merge/26919515:41
greybackdandrader: there's still lots of unnecessary changes. DASSERT/Q_ASSERT, removing DLOG.. yes they're good to fix up, but in a separate MR.15:43
greybackas this resizeCatchUp stuff is a workaround, ideally we'll just revert this branch when mir is fixed15:43
dandradergreyback, done15:50
tsdgeosgreyback: mzanetti: ted: don't ask me why, but this fixes it https://code.launchpad.net/~aacid/unity8/fix_upstart_restart/+merge/26939915:55
mzanettiinteresting15:56
tedNot sure I understand that.16:00
tedPerhaps a "sleep .1" too ;-)16:01
greybacktsdgeos: yeah I'm not clear on it either, but if it improves things, let's just take it16:01
tsdgeosted: ...16:02
ted16:02
greybackmzanetti: all MRs for silo23 are approved, qtubuntu needs a rebuild tho16:15
mzanettigreyback, ack, thanks16:16
=== alan_g is now known as alan_g|EOD

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