[02:13] hi [02:15] I am trying to set a msg count number in hexchat with python launcher.set_property ('count', count), and it works the first time, but if I reload the script is breaks like this: http://pastebin.com/raw.php?i=eudXgAcG [02:16] I was wondering if there was something I can do to clear out whatever is preventing the script from working upon second launch [02:16] I just think it's kinda weird that I can load test.py and it works, but if I unload/load it a second time, it breaks === jamesh_ is now known as jamesh [07:49] Saviq: yep. its networking issues in the lab. veebers is filing a ticket about it. I fear that's all we can do for now [07:49] mzanetti, yup, thanks [07:53] greyback: ping [07:55] mzanetti: sent, please fill in any details if I've missed any [07:55] veebers: ack. thanks [07:55] right, I'm off for the night. Beer and movie time :-) [07:58] have fun [08:05] mzanetti: pong [08:05] greyback: hey. on friday I wanted to schedule a meeting for today but managed to put it to last monday [08:06] greyback: I've just moved it to today but it would be in half an hour. is that ok for you? [08:06] mzanetti: yep [08:32] Trevinho, bregma, hey, could one of your review/ack https://code.launchpad.net/~larsu/unity/call-ido-init ? [08:33] Trevinho, bregma: I backported the ido init part to saucy on friday to unbreak custom widget in the gmenumodel based indicator that landed, having the other part about the ordering in would be nice as well [08:45] greyback, mzanetti, ping if you need me, btw [08:46] Saviq: yeah... actually... if you have a minute [08:46] would be useful [09:04] Small builds script fix: https://code.launchpad.net/~gerboland/unity8/fix-build-script/+merge/172290 if anyone has a minute [09:09] greyback, ping [09:09] tvoss: pong [09:09] greyback, hey there, I'm looking for libhybris-common, any idea where I can find it? [09:10] Saviq: do we have automerging in lp:unity/8.0 ? [09:10] you approved https://code.launchpad.net/~mhall119/unity/fixes-1196322/+merge/172251 9 hours ago and it's still there [09:11] tvoss: never heard of it [09:12] greyback, weird, platform api does not build without it [09:14] mzanetti: can you try to repro https://bugs.launchpad.net/unity8/+bug/1172212 on a new unity8? [09:14] Launchpad bug 1172212 in Unity 8 "HUD button gets stuck if application switches while dragging" [Medium,Incomplete] [09:14] tvoss: cmake should disbable hybris support if that lib is not available. You need platform-api with hybris? [09:14] tsdgeos: ack [09:14] greyback, yeah [09:15] greyback, we need it by default for the touch images as we need to pull over the sensor functionality iirc [09:15] tvoss: I see. Am hunting [09:15] greyback, thx [09:16] btw, i am the only one that gets unity7 panel notifications seconds later the real notification happens? [09:16] i.e. i get notify-osd to tell me "mzanetti's ack" and then like 10 seconds later the panel icon of the irc client vibrates [09:17] didrocks, any eta on next unity landing to s? [09:17] mhr3: well, we have some system issues in saucy making lxc not working fine [09:17] mhr3: and then, indicators did some changes that needs to be merged [09:18] mhr3: and if we don't have this hang, and all those are transitionned ^ we'll be fine :p [09:18] mhr3: you can review https://code.launchpad.net/~larsu/unity/call-ido-init meanwhile :) [09:18] tsdgeos: nope. can't repro any more either [09:18] good thing :-) [09:19] I'm getting pinged in all the channels... [09:19] oh, that's just a mention [09:20] didrocks: you're married! Congrats! [09:20] larsu: thanks! :) [09:20] didrocks: how are you? [09:20] didrocks, say what now??? [09:20] larsu: I'm excellent thanks! still a little bit tired by the week-end for sure ;) [09:20] and you? [09:21] didrocks, married? holy cow [09:21] didrocks: I'm great! Except for the insomnia :) [09:21] mzanetti: can you mark it as fixed then or soemthing? [09:21] mhr3: heh, fortunately I disconnected on Friday before the indicator issue was discovered, so mind-free :) [09:22] larsu: argh, I hope this will get better throughout the week (yeah, it's really early for you…) [09:23] didrocks, i can totally see this scenario "and you didier roche, do you take julie as your wife?" - "just a sec i'm being pinged here" [09:23] mhr3: ahah, I was afraid ted would try something like that! :p [09:25] tsdgeos, there's network issues in the lab [09:25] Saviq: ahhhh [09:25] tvoss: well libhybris package contains the /usr/lib/arm-linux-gnueabihf/libhybris-common.so.1 file [09:26] that's why the CI for the LVWPH didn't run yet either [09:34] greyback, okay, thx [09:35] tvoss: that wasn't much help. I'm rather confused how to build platform-api with hybris, there's something missing [09:35] greyback, yeah, I'm confused, too === rachelliu_ is now known as rachelliu [10:15] mhr3, hey [10:16] seb128, hello [10:16] mhr3, sometime my dash home stop listing apps ... is that a known issue? what sort of debug infos would you need in a bug report about it? [10:16] seb128, desktop/phone? just apps or everything? [10:17] mhr3: same issue here, haven't figured out yet when it happens exactly. Desktop, only apps. [10:17] mhr3, desktop, saucy, unity7, only apps [10:18] it lists files, music, et [10:18] etc [10:18] hmm, odd [10:18] so, good debug info would be whether the apps daemon is running [10:19] and ideally a bustle log trying to do a search in the dash [10:19] mhr3, the app lens is working [10:19] if that's an indication [10:19] but the home dash is sort of screwed it [10:19] when I open the dash it's empty, sometime with the "no result" string [10:20] let me do a screencast [10:20] but just in home? everything works fine in apps lens view? [10:20] yes [10:20] weird getting weirder :) [10:22] mhr3, you like screencasts? ;-) [10:22] of course! :) [10:22] mhr3, http://people.canonical.com/~seb128/dash.ogv [10:24] eeeek, why is everything so super slow there [10:24] welcome to smart scopes! [10:25] to be honest it's not always that slow [10:25] but when the bug happens it is [10:25] that box is freshly booted this morning [10:25] let me try to get you a bustle log [10:26] mhr3, hum, app lens doesn't have "recently used" either ... that's supposed to be first no? [10:27] seb128, yep, looks like something wrong with zg [10:28] apps lens doesn't handle that well [10:28] although i had a patch somewhere [10:29] mhr3, http://people.canonical.com/~seb128/bustle.bustle [10:29] ignore the indicator-service spam [10:29] I need that fix from larsu :p [10:31] fortunately bustle can do filtering :) [10:31] I had no idea my dbus was that spammed :p [10:31] seb128, ok, i think i know what this might be, can you try killing unity-applications-daemon and see if things start working normally? [10:32] mhr3, do you need any state/info from it before I do that? [10:32] nope, should be good enough [10:33] that didn't fix it [10:33] I guess "sorry there is nothing that match your search" now when I open the dash [10:33] with nothing in the text entry [10:33] did you uninstall zeitgeist by any chance? [10:34] no [10:34] but we got the new version last week [10:34] then it's broken :) [10:34] 0.9.14 [10:34] seb128, terminal `zeitgeist-daemon --replace`, any warnings? [10:34] ** (zeitgeist-datahub:8190): WARNING **: zeitgeist-datahub.vala:229: Unable to get name "org.gnome.zeitgeist.datahub" on the bus! [10:34] oh [10:35] I saw an update on -changes earlier to re-enable that I think [10:35] * seb128 goes to check [10:35] oh, no [10:35] mhr3, https://bugs.launchpad.net/ubuntu/+source/zeitgeist/+bug/1195841 [10:35] Launchpad bug 1195841 in zeitgeist (Ubuntu) "Zeitgeist should be built with fts support" [Undecided,New] [10:35] that wouldn't matter, anything else? [10:36] seems like it's currently built without fts [10:36] should I break things? [10:36] no [10:36] yea, i know about the fts thing, that makes the files search super slow [10:36] shouldn't affect apps [10:36] ok [10:36] no other warning [10:36] neither from zg nor from the applens [10:37] seb128, not even when you try doing a search? [10:37] oh, it's working after a search [10:37] mhr3, doing a search fixed it [10:37] so restarting zeitgeist actually fixed it [10:37] I guess [10:37] should I try that again next time it happens? [10:38] sure, i think that will be 100% effective fix [10:38] "workaround" I guess :p [10:38] right, yea [10:38] i'll try to think about what the root cause of it could be [10:38] let me know if you need debug info [10:39] I run into it quite often [10:39] so I can run a local patched version that logs infos or stuff [10:39] i'll update my vm and see whether i see that too [10:39] seb128, btw did it start happening with the new zg version? [10:39] not running saucy? ;-) [10:39] no, I'm pretty sure it was there before the zg update [10:39] 95% sure [10:40] heh, ok [10:40] seb128, and no, i find it a bit unstable :P [10:40] tssss [10:40] Saviq, I might give up with the looping using the listmodel [10:40] you have a responsability there! ;-) [10:41] Saviq, if I am at index 0, I add an item at index 0 (like a prepend) [10:41] Saviq, the stupid list view switches to the new item at 0 [10:41] seb128, heh, booted up vm and i'm seeing it too [10:41] so if I say when index = 0 insert 0… it kinda loops until it reaches the minimum date [10:42] mhr3, \o/ [10:42] index 0, insert 0, instead of being at index 1 it switches to 0 :-\ so it goes back again and it insert another one at 0 and switches to it [10:45] seb128, ok my bad, apps needs fts as well [10:45] mhr3, ok, so that's the issue? [10:45] mhr3, why does it work after restarting zg? [10:46] seb128, ehm, it shouldn't really :) [10:46] seb128, but one step at a time, let's get fts back first [10:46] Saviq: what happened to https://code.launchpad.net/~unity-team/dee-qt/deevarianttext-and-tests/+merge/153530 ? [10:47] mhr3, right, I will upload jbicha's sponsoring request for it [10:47] seb128, cool [10:49] but looking at my log, it's vastly different from yours, in your case zg wasn't working at all / wasn't queried by the apps lens, in my case it is and it's returning errors about missing fts [10:49] well, I'm pretty sure the breakage/slowness I saw was there before the zg update as said [10:49] the dash/lens are not really robust to zg issues it seems [10:49] which might be another bug ;-) [10:50] yea, maybe you have slow startup, some initial dbus request times out and things break [10:52] didrocks, hi. I fixed another important Compiz regression, the fix is tested, the breakage is critical, because Compiz does not know how to render text in trunk anymore - this branch fixes it: [10:52] https://code.launchpad.net/~mc-return/compiz/compiz.merge-fix1196493-no-more-text-anywhere/+merge/172302 [10:52] didrocks, please approve it, it's free karma ;) [10:53] MCR_: I would prefer sam to look at it TBH, Not sure how the vtable is supposed to be used in compiz [10:53] well, it is the same kind of breakage we had found and fixed on Friday... [10:53] it is tested [10:53] and it works [10:54] I trust you it works, I still prefer upstream taking care of their components as most as possible, when there is no urgency for releasing :) [10:54] ok === MacSlow is now known as MacSlow|lunch [11:26] dednick: want me to unapprove https://code.launchpad.net/~dandrader/unity8/dragHandleStretchAndHint/+merge/171908 until your comments are sorted? [11:28] didrocks: ping [11:31] tsdgeos: yeah, if you dont mind [11:32] tsdgeos: i just did [11:32] dednick: good [11:32] sorry i did not pick up that while reviewing [11:33] Saviq, ping [11:43] Cimi: what if you make sure you never reach index 0? [11:44] i.e. prepend already when you're at 1 [11:45] tsdgeos, we're not using it (the people preview was using it [11:45] ) [11:45] tsdgeos, s/preview/lens view/ [11:45] Saviq: that is the answer for the dee-qt changes? [11:45] tsdgeos, so, we could salvage some of the tests from that potentially [11:45] tsdgeos, ye [11:45] s [11:45] Cimi, if you're at 0 and prepend, and the list switches, I'd say that's a bug [11:46] IMO appending / prepending items to ListView's model should never change what's on screen [11:47] Saviq: it does if it's item 0 [11:47] mzanetti, but sometimes you're at 0 :( [11:47] it's a special case [11:47] mzanetti, unless I make that 0 is a greyed out month [11:47] Saviq, I need to write a testcase [11:47] i.e. if you are at list beginning and prepend, it still gets you at list beginning [11:47] afair [11:48] tsdgeos, well, that depends if you want to be at list beginning or item 0 [11:48] tsdgeos, will it switch if you're at item (count-1) and append? [11:48] no [11:48] IMO it's a broken special case [11:48] tsdgeos, yeah, exactly - it's inconsistent [11:48] tsdgeos, being at item 0 doesn't necessarily mean wanting to be on the beginning [11:49] true [11:49] so you can append and them put yourself where you want :D [11:50] tsdgeos, yeah, I would think that if you did it close in JavaScript, the result should be invisible [11:50] Cimi, did you try prepending and setting currentIndex to the correct value just after? [11:51] Saviq, yeah [11:51] tsdgeos, OTOH ListView is difficult like this - if the new item isn't created yet [11:51] Saviq, but kinda doesn't work [11:51] Saviq, I did [11:51] tsdgeos, it won't switch :/ [11:51] lunch [11:51] back later [11:51] if currentIndex == 0; insert (0..); currentIndex = 1 [11:51] Saviq, ^ but kinda sucks and probably doesn't solve the problem [11:52] Cimi, why? [11:54] Saviq, I think that if I set currentIndex in onCurrentItemChanged [11:54] Saviq, it does proceed with a new thread [11:55] Cimi, right, it might get confused [11:55] Saviq, I think I have two calls of onCurrentItem at the same time [11:56] Cimi, you should probably use a separate property to set the date selected [11:56] Cimi, and only modify the model when that changes [11:57] Saviq, the selected date works [11:57] Cimi, yeah, I know, but [11:57] Saviq, problem is when I flick/scroll between months [11:57] Cimi, if you use a separate property [11:57] I have 1 to 3 elements in my list model [11:57] Cimi, that will only change if the date _really_ changed [11:57] Cimi, regardless if currentIndex or currentItem change [11:57] when I move from index 0 or index 1 I want to insert a new item [11:57] when I move from 1 to 2 I want to append [11:57] Cimi, I know what you want to do [11:58] I don't understand what you mean with date [11:58] you are able to scroll without touching or selecting a new date [11:58] it's just flipping/browsing the calendar [11:58] what I could do might be storing the previous index [11:58] Cimi, then "displayedMonth" or something [11:59] Cimi, don't rely on currentItem / currentIndex to find out whether you need to modify the model [11:59] mmm [11:59] Cimi, but "proxy" those signals through a "currentMonth" property [12:00] Cimi, only when it really needs to change [12:00] and when it does? [12:00] Cimi, this way you won't get the loop of currentItemChanged → insert() → currentItemChanged [12:01] Cimi, because even if it does, the selected month won't change, so the loop will get broken [12:02] maybe just a boolean [12:02] an internal boolean I set true when I am modifying the model [12:03] then will be set false during next onCurrentItemChanged [12:03] and returning [12:03] Cimi, that's artificial [12:03] Cimi, the actual selected month is real [12:04] Cimi, you'd have to manage the boolean much more carefully than the selected month [12:13] tsdgeos: \o/ lvwph, nice work! [12:14] pstolowski: pong [12:14] didrocks: hi [12:18] didrocks: wrt https://bugs.launchpad.net/unity-lens-applications/+bug/1192081 need to double-check with you - is the problem mostly visible on the 1st app preview request? Is it faster later even when software-center-dbus process is not running anymore (i.e. does it help if it's in system cache and i/o is reduced to start s-c-dbus again)? [12:18] Launchpad bug 1192081 in unity-lens-applications "Previews for applications are really slow to display anything" [Critical,In progress] [12:18] pstolowski: no, if I kill the software-center-dbus process, it's slow again [12:19] so system cache doesn't help [12:19] didrocks: ok, I see. could you give the branch I linked to the bug a shot? [12:20] pstolowski: not right now, but I'll :) === jhodapp|afk is now known as jhodapp [12:20] pstolowski: apart if just restarting the daemon is enough? [12:20] (the application scope) [12:20] didrocks: sure, thanks... and take it with a grain of salt (i.e. read my comment).. but I'd be interested in your feedback [12:21] pstolowski: so, just building it and killing the scope is enough? (in that case, I can do it now) [12:21] didrocks: yes, kill software-center-dbus if running, restart apps [12:21] ok [12:22] didrocks: thanks [12:22] pstolowski: it's slightly better for sure, but I still have few seconds even before the spinner appears [12:23] (so animation, blank middle content, then spinner, then content) [12:23] didrocks: well, it's doing network query anyway to get the screenshot [12:24] pstolowski: shouldn't I see the spinner right away? [12:24] didrocks: or is it then fast on next previews? [12:24] pstolowski: yeah, then, it's fast [12:25] didrocks: also, if you are fast and request preview quickly after search results appear, it's may not be loaded yet [12:25] didrocks: (the s-c-dbus) [12:25] pstolowski: yeah, that's my guess [12:26] pstolowski: I think it's still an improvment, starting it a little bit earlier, but not sure we can ship with this [12:26] pstolowski: maybe it worthes a try already [12:28] didrocks: what do you mean? [12:29] pstolowski: like, it's already a small improvement in the end user experience, we can try getting this into distro and see for feedback === MacSlow|lunch is now known as MacSlow [12:51] kgunn: thanks :-) [12:52] Saviq: afair the rationale for inserts at 0 when list on top is "scrolling up" is because for example a move is sent as a remove + insert to the lsitview, so if you didn't do that, moving the item 0 to item 1 would end up with you at item 1 instead of item 0 [12:53] Saviq, shouldn't this end up in a file somewhere in lp:unity8? -> https://docs.google.com/a/canonical.com/document/d/1gd87Wo_CSB0DpFWLpTKIIXQfdmFncrq0PHSr9H2PTnk/edit [12:54] dandrader, yeah, not even unity8 but a a generic QML set of guides, it's on my TODO... [12:54] and about those guidelines [12:55] instead of having a __foo property. I instead should private.foo? [12:55] I mean s/__foo/private.foo [12:56] isn't this "private." prefix quite a mouthful? [12:56] C++ qt uses the d->foo idiom, for instance === _salem is now known as salem_ [12:59] dandrader, I don't think we meant to name them "private" [12:59] dandrader, "d" is fine [13:00] [...]" use 'private' to id these private objects to maintain consistency with C++ code"[...] [13:00] Saviq, cool [13:12] dandrader, yeah, that doc is incomplete and unmaintained for a bit, sorry [13:16] np [13:21] fginther: hey, around? :) [13:24] MacSlow: in case you're using this and it works fine for you, feel free to review https://code.launchpad.net/~mzanetti/unity8/improve-bzr-hook/+merge/172146 [13:25] mzanetti, I'll look into it after the stand-up [13:26] paulliu: afair you where working with pstolowski on https://code.launchpad.net/~stolowski/unity8/activation-and-previews/+merge/172102 ? are you reviewing it or need more people? (want to be sure i'm not stepping into someone elses toes) [13:27] tsdgeos: Yeah, I need more people. But I'll review it too. [13:27] ok [13:30] greyback_, mzanetti standup [13:30] ah, beat me again ;d [13:35] Saviq, FYI, https://wiki.ubuntu.com/AccountPrivileges#Phone [13:35] Saviq, We talked about that back on Oakland, we probably should finish it :-) [13:35] tedg, indeed [14:06] dednick, updated. https://code.launchpad.net/~dandrader/unity8/dragHandleStretchAndHint/+merge/171908 [14:06] didrocks, ping [14:06] davidcalle: pong [14:07] didrocks, hey! Could you please have a look at https://code.launchpad.net/~davidc3/cupstream2distro-config/new-server-scopes/+merge/172343 ? [14:07] davidcalle: simple enough, approved :) [14:07] didrocks, heh, thanks :) [14:08] yw! [14:24] yay jenkins is back [14:32] Saviq, all: Jenkins up'n'running again. build queue is half a meter long tho [14:32] mzanetti, yep, saw that [14:32] * Saviq is gonna have fun babysitting Jenkins tonight [14:34] dandrader: approved [14:34] dandrader, thanks! [14:44] Saviq, http://paste.ubuntu.com/5816922/ [14:45] Saviq, scroll right to increment [14:45] then scroll left and you'll see it loops [14:46] mzanetti, tsdgeos ↑ can you guys have a look? [14:46] didrocks, so, are we doing https://code.launchpad.net/~unity-team/unity/unity8-packaging-cleanup/+merge/171537 ? [14:46] Saviq: ack [14:47] didrocks, or did the jsons not land in distro yet? [14:47] Saviq: no, it didn't, let's wait for it to land [14:48] didrocks, right [14:48] Saviq: due to this jenkins issue, still catching up on merges [14:48] didrocks, yup [14:48] so probably tomorrow, if unity tests pass [14:48] * didrocks crosses fingers [14:50] mzanetti, tsdgeos basically I want to prepend an element [14:50] mzanetti, I expect the currentItem not changing [14:50] Cimi: yeah... I see the problem [14:50] oh well, if I am at index 0 and I append 0 [14:50] I expect to become 1 and stay [14:50] instead it goes to 0 for a sec [14:56] you should avoid changing the list on list signals, its easy to end up in some nasty recursive stuff === dandrader is now known as dandrader|afk [15:07] tsdgeos, but I don't have other ideas [15:07] tsdgeos, I would like to append/prepend while the user scrolls [15:07] Cimi: why do you need this? [15:08] tsdgeos, to have an infinite model [15:10] but why do you need a model? because you're using a listview? i'd say that simply having 2 items would be enough, but that means rewriting most of your code, so feel free to ignore me :D [15:11] i mean to me conceptually a calendar is not a listview [15:13] Cimi: here you go: http://paste.ubuntu.com/5817002/ [15:13] Cimi: only downside, the model currently grows unlimited. you should also remove the last from the other side when inserting [15:14] mzanetti, I know this works... [15:14] so whats the problem then? [15:14] mzanetti, I have your variation with removing the last [15:14] mzanetti, it loops [15:15] Cimi: ah... only the removing is the problem? [15:15] * mzanetti tries [15:15] mzanetti, I mean [15:15] mzanetti, sorry let me have a proper look at yours [15:15] maybe I quickly criticised but might do [15:16] Cimi: updated version that keeps the model down to 5 elements: http://paste.ubuntu.com/5817011/ [15:16] which is a cachebuffer of 2 items on each side [15:16] mzanetti, goal is 3 [15:17] mzanetti, because it has to work even with two items [15:17] Cimi: I expect it to be come less performant if you use only 3 items [15:17] imagine date from july to september [15:17] so the logic should work [15:17] Cimi: especially if you allow overshooting of the list [15:17] then you would see stuttering [15:17] because it quickly needs to create the item and the delegate which it should show it already [15:18] I think a cachebuffer of 2 items on each side sounds reasonable === dandrader|afk is now known as dandrader [15:22] Cimi: why would it need to work with only 2 items? [15:22] mzanetti, when you set a range [15:22] mzanetti, like only july and august [15:22] for example [15:23] imagine you're booking holidays for 2013 and you're in november [15:23] you can book only for november/december [15:23] Cimi: ok. understood [15:23] th calendar has an API to set max and minimum date [15:23] Cimi: but still not a problem [15:24] Cimi: add an if to alll the places that call insert or append. if its outside the range. don't do it [15:24] * mzanetti tries [15:24] indeed maybe the difference in yours is that you do onCurrentIndexChanged [15:24] and you are index-agnostic [15:25] so it doesn't care if the index is 0 or 1 or 2 [15:25] just looks at the value if it's in the range [15:25] sounds damn simple yet correct [15:25] I'm wondering why I didn't think of that before but was getting mental with the wrong approach [15:26] in any case I still believe there's a bug in qt somewhere :) [15:51] Cimi: there is one rule with QML: if it doesn't work, its probably too much code [15:51] mmm [15:52] Cimi: but I'm seeing the issues you've seen before [15:52] Cimi: its a bit of a pain to initialize the model more dynamically [16:09] Cimi: ok... I think this is it: http://paste.ubuntu.com/5817174/ [16:10] Cimi: you can set minValue and maxValue to undefined and it'll be infinite [16:10] Cimi: otherwise it'll respect those values [16:10] Cimi: one tricky part was to initialize it without triggering the logic [16:10]