=== _salem is now known as salem_ === salem_ is now known as _salem === SAM113101 is now known as sam113101_afk === sam113101_afk is now known as sam113101 === jalcine is now known as jalcine_ === jalcine_ is now known as jalcine === jalcine is now known as jalcine_ [06:39] Trevinho: hey! can you please take a look (once you are around) to unity FTBFS in trusty? https://bugs.launchpad.net/unity/+bug/1243529 [06:39] Ubuntu bug 1243529 in Unity "unity FTBFS on trusty" [Critical,New] [06:39] jamesh__: your turn now ;) we have a FTBFS in trusty on mediascanner (seems one test timeout on all archs): https://bugs.launchpad.net/mediascanner/+bug/1243536 [06:39] Ubuntu bug 1243536 in mediascanner (Ubuntu) "mediascanner FTBFS in trusty" [Undecided,New] [06:40] jamesh__: can you take a look asap? === iahmad_ is now known as iahmad === jamesh__ is now known as jamesh [06:42] looking [06:46] thanks [06:51] didrocks: is there a new grilo or grilo-plugins in trusty? [06:51] jamesh: no, both are the same than in saucy from what I see [07:14] didrocks: I've just put up https://code.launchpad.net/~jamesh/mediascanner/bug-1243536/+merge/192292 for review. Is the CI likely to catch this problem, or would it still be testing against saucy? [07:14] jamesh: I don't really know about upstream-merger, seems they have issues to switch to trusty [07:14] jamesh: but I'll keep you posted once this lands, I have a ppa build then :) [07:15] didrocks: looking at the error, I think the changes in that branch should address it though. [07:17] jamesh: excellent! is someone going to review it? [07:17] like mhr3? [07:17] I'll get someone on my team to [07:17] thanks a lot jamesh ;) [07:26] hey mardy! [07:26] mardy: it seems that libaccounts-glib is FTBFS on saucy (https://launchpadlibrarian.net/154768960/buildlog_ubuntu-saucy-i386.libaccounts-glib_1.14%2B13.10.20131023-0ubuntu1_FAILEDTOBUILD.txt.gz) [07:27] mardy: but I guess first that the 13.10 just rather just be a bzr push from trunk to that branch [07:27] Mirv: why didn't you just bzr push trunk to https://code.launchpad.net/~online-accounts/libaccounts-glib/13.10? [07:35] sil2100: any opinion? [07:35] sil2100: otherwise, I guess we have a diff and that's why it tries to rerelease it with no change [07:38] larsu: sometimes the sound indicator on the desktop does not go to the "muted" state when i middle click, even if it indeed mutes the sound coming out from the speakers, is that something known? [07:42] didrocks: how does libaccounts-glib trunk look like? Is that all that we released into saucy already? [07:42] sil2100: can you check it, please? it seems so to me === jalcine_ is now known as jalcine [07:46] didrocks: it looks like so to me too - you think I can do a bzr push --overwrite to 13.10? [07:47] Or merging it in with conflict resolving? [07:48] sil2100: no, better to bzr push --overwrite IMHO [07:49] didrocks: hi! I'm not sure I understand what the problem is. That m4/gcov.m4 file is in trunk [07:49] mardy: yeah, same here :/ [07:50] mardy: anyway, let's fix the 13.10 branch, and we'll see [07:50] (we have enough other issues ;)) [07:50] didrocks: pushed the trunk, let's maybe re-spin it and see if it's the same? [07:51] Or actually [07:51] didrocks: what is wrong with the 13.10 branch? it looks fine to me [07:51] sil2100: to trunk? [07:52] mardy: now it's fine, it was a manual merge where a pushed should have been enough [07:52] No, pushed the trunk to 13.10 ;) [07:53] didrocks, sil2100: weird, I think that I pushed trunk to 13.10; not sure why it appeared as a merge [07:54] sil2100: looks good :) [07:54] mardy: it was… anyway, all fixed now :) [07:54] thanks sil2100, mardy [07:54] didrocks: ok seems now resolved with the branches identical? [07:55] but yes pushing the trunk is a lot better in that case [07:57] yeah, ok ;) [08:05] Well, who would have thought that those two can be pushed directly ;) [08:13] nic-doffay, ping [08:20] Cimi, in carousel-music-video, why do we need the XXX comments, can't we just test it? ;) [08:25] pstolowski, pong [08:29] didrocks: fyi, https://code.launchpad.net/~jamesh/mediascanner/bug-1243536/+merge/192292 is now merged [08:29] jamesh: excellent! will keep you posted at next build :) [08:29] it passed Jenkins' version of Trusty, so hopefully it will pass Soyuz's one [08:29] nic-doffay, I looked at the filters; it seems to me that UI always marks the 1st filter option as selected (be it 'All' or a real option) [08:30] jamesh: that should do it, indeed :) [08:30] nic-doffay, perhaps it has something to do with the index of selected option to you manage yourself? === jalcine is now known as jalcine_ [08:32] nic-doffay, and btw, home scope doesn't have 'all' option by design, in that case it means no filter option should be selected by default. and 'All' is not currenntly pre-selected by the bindings which is something I'll need to fix. but most important, UI shouldn't assume 1st item is selected by default [08:42] pstolowski, ah I see. [08:42] pstolowski, if you add that then I'll remove the selection on my side. [08:57] nic-doffay, I think you can remove it right now, because now it makes 'Applications' filter option pre-selected in Home, and this needs to be fixed anyway [08:58] mzanetti, what's qCDebug? [08:59] pstolowski, yeah I was planning to. [08:59] mzanetti, I mean, what's the difference between it and qDebug [08:59] dandrader: that's the new qDebug which allows you to define categories and turn on/off categories at runtime [08:59] dandrader: let me find the docs [08:59] mzanetti, does it append the category name as well to the print message? So it would replace the [DDA] prefix? [09:00] s/append/prepend [09:00] dandrader: its fully configurable. you can make it append function names, classes, line numbers and what not [09:01] (at least the preview I've been working with in Nokia was) [09:01] mzanetti, but we would still keep the "#if DIRECTIONALDRAGAREA_DEBUG" clauses, right? [09:01] I'd say no [09:01] Qt devs claim it's very fast if the category is disabled [09:01] mzanetti, hmmm, but won't it affect runtime performance as you would have at least to do a check on whether you should proceed with the debug [09:02] dandrader: according to their makers they say it won't affect performance really. I don't have any numbers though [09:02] mzanetti, and the code that produces the debug message costs something [09:03] dandrader: yeah true. if you do heavy calculations just for the debug message, then yes [09:04] dandrader: http://doc-snapshot.qt-project.org/qdoc/qloggingcategory.html [09:08] Saviq, sorry missed the notification - I don't have a nexus 10, when I tried oren one was crashing and was not able to test it [09:08] Cimi, no worries - I put more comments on the MP [09:11] sil2100, ping? [09:11] mhr3: pong! [09:11] mzanetti, what's the lock screen? [09:12] Cimi: do a "echo password=pin > ~/.unity8-greeter-demo" and restart unity8 [09:12] mzanetti, thx [09:12] sil2100, hey, got a question - should we branch S branches for our projects and continue pushing new stuff to trunk? or is everything setup to land in T now, and therefore enough to create S branches only if we need to sru something? [09:13] mhr3: which projects do you have in mind? [09:13] mhr3: since we basically did a split for all the projects that needed splitting, and trunks are now basically targetting T right now [09:14] mhr3: touch projects didn't get a split as we're not having maintenance branches for saucy there - but those that are shared with desktop have now a trunk branch (for T) and a saucy branch for saucy (usually /saucy or /13.10) [09:14] sil2100, online guys wanted us to push a small change to home-scope and it should be sru-ed to S === dandrader is now known as dandrader|afk [09:15] mhr3: ok, then push it to [09:15] ah, yea, i see the /saucy branch now [09:15] lp:unity-scope-home [09:15] and then backport to lp:unity-scope-home/saucy [09:15] ok, cool, thx [09:16] We'll then pick it up for SRUing when the time comes [09:16] sil2100, when will that time come btw? [09:16] roughly [09:16] mhr3: ...I have no idea right now [09:16] Sadly [09:17] well... let me know once you do :) [09:21] Sure ;p === dandrader|afk is now known as dandrader [10:10] why does messaging app run some odd indicator app on 101? [10:16] dandrader: but you don't reject the gesture now when a second touch point comes in, right? [10:17] not a problem for me... just noticed [10:21] Qt 5.2 Beta Available [10:23] mzanetti, not if that second touch point comes after the composition time [10:23] ah... [10:24] rephrasing: if a second touch point comes within the composition time window in the recognition phase (undecide status) the gesture gets rejected. otherwise it's ignored [10:25] /food [10:28] pstolowski, we have an issue with no filter option. [10:29] Just had a look at the branch again (was on something else) [10:29] Since an option always has to be selected if it's a drop down selector. [10:29] We'd need a None option in the backend. [10:29] For Home. [10:29] dandrader: I see traces of a hintingAnimation. where is that used? [10:31] mzanetti, you mean from DragHandle.qml? [10:31] mzanetti, in the panel [10:31] mzanetti, press on the indicators bar. it will come up a bit [10:31] dandrader: ah, that's in draghandle already. I assumed that's in the panel [10:32] mzanetti, same when you press on the bottom of an open indicators panel [10:32] dandrader: ah ok. [10:32] dandrader: I managed to break the right edge once. Not able to reproduce yet [10:33] again [10:33] ok... happens with trunk too. I finally have a way to reproduce the stuck greeter \o/ [10:34] mzanetti, what do you mean by "breaking the right edge"? [10:35] dandrader: it's not in the DDA. but try this: drag the greeter away but release it in a way that it would return back to locked state. while it returns, tap on it [10:36] nic-doffay, oh, that's not good. we would need to add 'All' button in home scope backend for phone only (based on form factor hint). But there is no nice way of giving it a 'None' label just in home [10:37] mzanetti, another funny thing is to slide in the launcher and in the middle of it start sliding down the panel with another finger :) [10:38] yeah. you can operate all edges simultaneously :) but it doesn't break, does it for you? [10:39] mzanetti, there's logic to hide the launcher if the panel is open, thus there's some fighting taking place in this scenario [10:40] dandrader: yeah, its worse the other way round. first the panel and then the launcher [10:40] pstolowski, yeah it's unavoidable. It's part of the components design. The only way to get around this would be to use an expanded selector, but that wouldn't be per the filter design. [10:41] how do I populate unity with my music/video on the desktop? [10:42] unity8 [10:42] nic-doffay, I see [10:42] Cimi, banshee or rhythmbox for music [10:42] Cimi, videos - just put stuff in your Videos folder [10:42] nic-doffay, don't say it's impossible, and instead try and think of how to solve that conundrum - maybe a property on OptionSelector to allow selecting none or something [10:43] nic-doffay, although I agree the design doesn't really fit with either none-selected or multiple-selected [10:43] nic-doffay, it's a classic single-selection drop down [10:43] * Saviq really food now [10:48] Saviq, still empty here, maybe I have to wait [10:51] Cimi, what did you do? added stuff in ~/Videos? [10:55] seems to work now [10:56] Cimi: just do a search for a video you know is in there same for music [10:56] Saviq, the only way this can be solved is by going against the design which I think is a bad idea. Adding another option in the backend seems like the best solution currently. [11:00] nic-doffay, for the record, adding new option in the backend will be against design as well ;), because they specifically requested 'all' and alike not to be shown in Home ;) [11:08] pstolowski, cool, I'll bring it up with them. [11:08] Saviq, which subtitle do we need to support in carousel video? [11:10] dandrader: why is TimeSource abstract? do we need different ones for different devices? [11:10] nic-doffay, awesome, thanks. keep in mind 'All' in Home would be misleading (just in case this idea is discussed), because it never shows all results [11:11] mzanetti, for the tests. they will implement a fake timer and pass it to the DirectionalDragArea under test [11:11] dandrader: ah, I see [11:12] s/fake timer/fake time source [11:13] pstolowski, will do [11:15] Cimi, comment, probably - same as in music [11:15] Saviq, but the video grid does not have them [11:15] Cimi, the video carousel does [11:16] Saviq, I can add one line for the comment [11:16] Cimi, yeah, I believe that's the idea https://lh3.googleusercontent.com/Y9l0nJDyaCA7fZ92jNb6dA5SjlPnvkDy6aFHsZxBwVd9tWVODvgVfvlQ0VwBHurq3fZn7ELxctwF-uh4WvZlnEVZv9eflCTSgCNPC3WgMRDMQ6THrjmRh1Vnqg [11:17] Saviq, that's cheesy title, who did this mockup? :D [11:17] Cimi, Rosie [11:17] hah [11:17] cute rosie <3 :D [11:18] every time I drop by her desk she wants gossip on my girls :D [11:20] dandrader: your branch looks really good. left some small comments [11:20] * dandrader reads [11:21] mzanetti, you have the 13 or 15 retina? [11:21] Cimi: 15 [11:21] mzanetti, I fancy the 13, but it has dual core [11:21] * Cimi upset at apple [11:21] why??? why????? [11:21] * mzanetti likes looking at htop displaying 8 cores :) [11:22] mzanetti, not sure 15 is portable [11:22] 4 + ht that is [11:22] mzanetti, if I have to travel to the office [11:22] Cimi: it is. it's so thin and light... [11:22] Cimi: I always use backpacks tho [11:22] I like messenger bags [11:23] Cimi: well, the weight is definitely not the issue [11:23] mzanetti, how about hand luggage on flights? [11:23] is it quite big inside? [11:23] I always keep it in my backpack [11:23] yes but when I fly I have a trolley [11:23] and it's really thing (ethernet doesn't even fit into it) [11:23] well, might not fly with the 15 [11:24] -g [11:24] iPad air seems pretty cool [11:24] I like a device to browse the bed on sofa/bed [11:24] can't install ubuntu on that... not for me [11:24] and one productivety monster [11:25] mzanetti, you can have two [11:25] mzanetti, one ubuntu one not [11:25] * mzanetti doesn't own a tablet... [11:25] mzanetti, it's good to have a look at competitors === dandrader is now known as dandrader|lunch [11:26] Cimi: sure. but I won't spend 500€ for a thing that can't do anything except displaying browser bookmarks and call it apps [11:26] mzanetti, it is expensive [11:26] mzanetti, I have a nexus 7 2012 and all browsers are shit [11:27] I don't mind expensive if it is of real use [11:27] but I don't feel like the iPad is [11:27] for me [11:27] can't wait to have my hands dirty on our tablet [11:27] so I can push to have what I want ;) [11:28] bregma: hi! I see updates on bug #1243529 - can you assign anyone from your team to modify unity accordingly to the new ABI? [11:28] bug 1243529 in Unity "unity FTBFS on trusty" [Critical,Triaged] https://launchpad.net/bugs/1243529 [11:28] in the meanwhile, 45 mins to finish downloading mavericks [11:28] bregma: thanks! [11:29] * Cimi really likes the ubuntu touch sounds, I put them on android === MacSlow is now known as MacSlow|lunch [11:37] Saviq, your last comment on the filters mo [11:37] *mp [11:37] What do you mean by the crop line? [11:49] nic-doffay, there's 2gu between the point where the overlay ends and where the content gets clipped [11:49] nic-doffay, and it's transparent to input - you can activate the stuff in the dash through it [11:51] Mirv, hey, any plans on building Qt 5.2 somewhere? :D [11:52] Mirv, we need to kick the crap out of the new js engine ;) [11:53] Saviq: I just heard the beta is out. I'm currently out of steam, but yes makes sense. earlier this week I did manage to push the first trusty 5.1 builds into qt5-beta-proper - it seems it's built for arm https://launchpad.net/~canonical-qt5-edgers/+archive/qt5-beta-proper/+packages?field.series_filter=trusty [11:53] Mirv, cool, no worries - get some rest :) [11:53] you've been busy the last few weeks - we had it easier the last days at least [11:54] Saviq, cool, gotcha] [11:54] Saviq: no rest, but other stuff like the first trusty landings of ~everything :) [11:55] Saviq: oh right, those 5.1.1 trusty builds are now debug builds [11:55] Mirv, meaning with CONFIG+=debug? === alan_g is now known as alan_g|lunch [11:55] tsdgeos, you'll be pleased ↑ ;) [11:55] Saviq: so the qtbase with -debug and the rest with CONFIG+=debug [11:56] Mirv, cool beanz [11:57] Mirv: so qtbase is still not debug? [11:57] I'll update the recipes to build for trusty too [11:57] tsdgeos: it has a configure switch -debug instead [11:57] tsdgeos: I'd trust it's the same thing as qtbase doesn't have a qmake to give the parameter to before it has built the qmake in the first place? [11:58] tsdgeos: did you get anywhere with running unity8 with 5.2? [11:58] I've not tried since early last week tbh [12:14] Saviq, I updated the branch, but still waiting jouni for a better asset for music [12:14] Cimi, k [12:15] Saviq, I have no idea where he is [12:15] Saviq, first day I stay home for working he is not online :) [12:15] dammit === dandrader|lunch is now known as dandrader [12:18] Cimi, have you gotten hold of him? [12:19] nic-doffay, nope [12:19] * Cimi install mavericks, see you soon (hopefully) === jhodapp|afk is now known as jhodapp [12:19] best use of lunch break I hope [12:37] greyback, hey, this should be relatively easy to track down https://bugs.launchpad.net/unity-mir/+bug/1243444/comments/5 [12:37] Ubuntu bug 1243444 in unity-mir "Unity (and sometimes maliit) crashes after maliit-server is restarted and is used." [Medium,New] [12:38] Saviq: ta [12:51] tsdgeos: yes, that's known and a fix is in -proposed (libindicator) [12:51] larsu: awesome [12:51] tsdgeos: thanks for the report! [12:52] Mirv: ah sure, i read "-debug" as "wihtout" debug, not as -debug passed to configure :D [12:52] greyback: didn't try because it wanted more stuff to be linked against 5.2 [12:53] tsdgeos: ah yes you mentioned that. Ok so [13:03] tsdgeos, here's one for you to look at in your KDE hat: bug #1240408 [13:03] bug 1240408 in Unity 8 "unity8 crashed with SIGSEGV in QCoreApplication::postEvent()" [High,Triaged] https://launchpad.net/bugs/1240408 [13:03] Saviq: ? [13:04] tsdgeos, it's all Qt in the stack trace ;) [13:04] ah :D [13:04] tsdgeos, I suspect either the album art provider fails, or the qpixmapcache [13:04] ok, will have a look [13:04] how urgent is it? [13:04] tsdgeos, and since the album art provider deals with QImages [13:05] tsdgeos, relatively urgent - it reliably crashes the shell if you expand Music in the home scope [13:05] tsdgeos, but not something that should throw you off of whatever you're doing now [13:05] ok [13:05] especially since we can't merge the fix anyway ;) [13:05] * tsdgeos never had any crash in the music scope .S [13:06] Saviq: desktop too? or phone only? [13:06] tsdgeos, not music scope - music in *home* scope [13:06] tsdgeos, or when opening a music preview, apparently [13:06] tsdgeos, didn't try on desktop, let me [13:07] larsu: ping [13:07] tsdgeos, doesn't seem to crash on desktop, no - but we don't get music in Home on desktop [13:08] dednick: hey [13:08] tsdgeos, no crash after searching for music in home and expanding, either [13:08] +on desktop [13:08] i see [13:08] so only crash is [13:09] tsdgeos, expand music in Home - that was reliable for me [13:09] expand music in home in phone before searching [13:09] or [13:09] tsdgeos, yeah [13:09] ok [13:10] Saviq: shall i get trusty on the phone? or the "old" images ? [13:10] tsdgeos, I had trusty [13:10] larsu: hi. I've been working on the datetime indicator and it gives me time + time-format (in strftime format). I'm going to need your TimeFormatter to parse it, but it's taking Qt time. Any way we can Separate the TZ part from the Formatter so we can use it like a plugin? [13:10] ook, so let's do trusty then [13:10] tsdgeos, so yeah, trusty-proposed to get the latest image [13:10] larsu: although i guess this could be a later mod [13:11] May I ask a question about Icons in the unity dock? [13:11] larsu: *taking Qt time format [13:11] mzanetti, updated [13:11] dandrader: ack [13:11] dednick: if you don't need the timezone changing functionality, can't you just call strftime? [13:12] larsu: it gives a time in us from epoc. so will need the tz change. [13:12] i'm guessing [13:13] * chelgrian has an QT executable which has no .desktop file and doesn't oviously call a setIcon or other method but never the less seems to show an icon in the dock rather than a ?. I was looking for pointers on how to work out where it is getting this image from. [13:16] dednick: strftime does this for you... [13:16] dednick: or does the service give you a timezone as well? [13:16] * larsu rereads your question [13:16] larsu: we dont get the change notification. :) [13:17] larsu: nevermind. I'll sort it out once the formatter is in. [13:17] mzanetti, is CI working at the moment or did it get confused by the saucy->trusty transition? [13:17] dednick: the service should notify you about tz changes, no? [13:18] dandrader: ci is still on saucy afaik. but there's an issue that we have a very dirty hack when installing unity8 on the phone which doesn't work on the read only image. so ci can't install the test packages there and fails [13:18] dednick: because I could have the clock set to a different tz than my system is running on [13:18] dandrader: that's why autolanding is disabled right now and we have 15 or so approved branches waiting to land [13:20] larsu: This is for the alarms/event menu items in the indicator. They are given in us since epoc, which wont change when the tz changes, so it wont update the UI. I'm saying i need to use your TimeFormatter so that we get the dbus notification that the timezone has changed so we can update the UI. [13:21] larsu: as far as i know, same as you did for message indicator [13:21] larsu: but the datetime indicator also provides the format which we want to show the alarm in. === MacSlow|lunch is now known as MacSlow [13:24] dednick: I know, my point was that the service has a timezone property as well. If you have a timezone and a time, you don't need and 3rd party notifications. But it just occured to me that the service doesn't need to export a tz, because it changes the system's tz as well. So your way would work as well. [13:25] dednick: if we're going that route, I'm totally fine with extending TimeFormatter [13:26] larsu: ok. [13:27] larsu: is g_bus_get and g_dbus_connection_signal_subscribe lightweight calls? [13:27] larsu: still not sure about doing it for every TimeFormatter :) [13:28] dednick: signal_subscribe adds an entry to a hash table. [13:28] dednick: g_bus_get adds an event to the end of the message queue in the usual case (which is that we already have a connection to the system bus) [13:29] larsu: yeah, it's the queue that i'm concerned with. [13:30] dednick: it's a mainloop, we have tons of events on there already. I wouldn't worry about that tbh [13:30] larsu: meh. we can singleton'ize it later i guess. [13:30] dednick: ya, if there's a problem. Also, having a singleton means doing atomic ref counting... [13:30] I just didn't think it was worth it [13:31] larsu: ok. i'll review and test the branch now. [13:31] eh? [13:32] how did i quit the hangout === _salem is now known as salem_ [13:45] dednick: thanks === alan_g|lunch is now known as alan_g === dandrader is now known as dandrader|afk === dandrader_ is now known as dandrader [14:37] Mirv: i'd say you apply only the patch i gave you (the fix) instead of fix+test [14:37] the test is upstream and proves it works [14:37] not sure we need it in our packages [14:38] Mirv: if you really want the test in the patch too i can try to port it against 5.1.1 [14:49] Saviq: just had a chat with jhodapp about our edges. and he has a valid point that we're triggering them too easily. and I think shell.edgeSize should be rather 1 gu instead of 2 (or even less). Would you agree? I could create a branch trying to tune it a bit [14:58] mzanetti: was it you that did the narrowMode thing for the PageHeader/Search ? [15:02] tsdgeos: nic-doffay [15:02] mzanetti, yep [15:05] nic-doffay: i may have to change that code a bit with the upcoming "tabbar as header", what's the logic in there? the idea is that on the tablet you get the search in a different place? right? [15:07] tsdgeos, it's still in the same place. [15:07] ok [15:07] tsdgeos, it's all governed by internal states in the component. [15:08] yes [15:08] i have to change the component [15:08] that's why i am asking [15:08] tsdgeos, what sort of changes? Chances that it will create big conflicts? [15:09] probably [15:09] as said [15:09] the label goes away [15:09] we have a tabbar tere [15:09] tere -> there [15:09] like the one in indicators [15:09] wonder how the searchbar has to play with it [15:09] i need a designer :D [15:39] mzanetti, problem with 1gu is that the touch hardware is crap [15:40] Saviq: oh really? *that* bad? [15:40] mzanetti, and if you swipe too fast it will just go unnoticed, 'cause the first touch you get is already beyond the 1gu threshold [15:40] mzanetti, yes, dandrader knows all about it [15:40] mzanetti, just run the qml edge test under surfaceflinger [15:41] mzanetti, and you can experiment there [15:41] Saviq: ah ok. but we do have some improvements on our roadmap, right? [15:41] or is the general opinion that this is good enough? [15:41] mzanetti, not in the strictness with which we detect whether something's an edge drag or not [15:41] mzanetti, we've loosened the rules after the first iteration [15:41] mzanetti, 'cause there were too many false negatives [15:42] Saviq: but right now we a lot of false positives [15:42] mzanetti, what we are looking at is allowing apps to get the events on the edges if they're *not* edge swipes [15:42] mzanetti, that's not really for us to make that determination [15:43] tsdgeos, were you chatting about this new bar that's obscuring the current search field earlier? [15:43] nic-doffay: which new bar? [15:45] mzanetti, the situation we have now was an improvement according to people that have complained about too many false positives [15:45] mzanetti, so it's really a he-said/she-said - maybe we didn't catch the right balance yet [15:45] mzanetti, but still, none of us are the right judges I'm afraid [15:45] Saviq: yes. I agree that at some point we were too strict [15:46] mzanetti, feel free to file an ubuntu-ux bug, try and explain in what situations you're getting the false positives [15:47] Saviq: ok, I will [15:47] mzanetti, I talked to jhodapp yesterday or so and IIUC his beef was mostly with the bottom edge [15:47] Saviq: jhodapp made a good example. try tapping the time in the video player [15:47] mzanetti, which does not have the fancy edge detection that the shell has [15:47] mzanetti, that's different [15:47] Saviq: no [15:47] Saviq: when holding the player in landscape [15:47] mzanetti, nothing should happen on tap [15:47] the right edge (which at that time is the bottom one) kicks in [15:48] Saviq: exactly [15:48] Saviq: but it does [15:48] mzanetti, so you're not tapping ;) [15:48] Saviq: well... we could argue on that :D [15:48] mzanetti, not sure what the minimal movement is - maybe that's just too small [15:48] Saviq: I think our rules are too loose [15:49] mzanetti, tbh whatever I do, when I tap, the right edge does not kick in... [15:49] Saviq: I guess our first iteration had the angle way too strict. but beause of that we also dropped minimum speed etc [15:49] Saviq, you're trying that from the mediaplayer in landscape? [15:49] jhodapp, yes [15:50] mzanetti, it wasn't random, though - people sat down and tried to find a balance - and that's what we have now [15:50] Saviq, using your thumb? [15:50] jhodapp, no, you can't tap with a thumb :D [15:50] jhodapp, still [15:50] lol [15:50] jhodapp, I can make a video ;D [15:50] I can tap on the N9 with a thumb [15:50] :P [15:50] jhodapp, mzanetti, tapping with a thumb does not bring the right edge for me, no [15:51] does it toggle the time display? [15:51] interesting...maybe you have an improved image that I don't? [15:51] jhodapp, it wasn't touched [15:51] mzanetti, that's a different thing [15:51] mzanetti, the fact that it doesn't toggle is indeed a problem - and one that we're solving [15:52] mzanetti, but that's not the same as the edge being triggered [15:52] Saviq, yes, I see what you're saying now and the same happens for me [15:52] * mzanetti tries again [15:52] I need to do an explicit edge gesture to get the right edge [15:52] I don't trigger the gesture most of the time, but it is very hard to get the time to toggle [15:52] jhodapp, ok then - that's two different things :) [15:52] jhodapp, and yes - that we're working on (dandrader is) [15:52] Saviq, but I can tap to get the edge gesture just below the time [15:53] jhodapp, that doesn't happen here [15:53] interesting...no matter how close to the edge you get? [15:54] jhodapp, no - it's not about how close to the edge you get [15:54]