* tsdgeos reports back from holiday week #1 | 07:04 | |
tsdgeos | bitten by mosquitos all over | 07:04 |
---|---|---|
tsdgeos | have to type slower because of swollen mosquito-bit finger | 07:04 |
tsdgeos | am i going to be killed if i mention we should move to Qt5.3.1 insteadd of Qt5.3.0? | 07:13 |
tsdgeos | mzanetti: Saviq: what's the status of https://code.launchpad.net/~aacid/unity8/swipe_down_close_no_remove_dash/+merge/221996 ? | 07:53 |
mzanetti | tsdgeos: Saviq found some issues and put it back to needs review | 07:56 |
mzanetti | tsdgeos: I have then done/fixed it on top of the QtComp branch | 07:56 |
tsdgeos | i see | 07:56 |
tsdgeos | mzanetti: can you comment in https://code.launchpad.net/~aacid/unity8/swipe_down_close_no_remove_dash/+merge/221996 saying it's done in the qtcomp branch? | 08:09 |
tsdgeos | so it's clearer it's "done-ish" | 08:09 |
mzanetti | tsdgeos: ack | 08:11 |
Cimi | mako improved so much over the last months | 08:21 |
Cimi | exciting | 08:21 |
mhr3 | tsdgeos, could you take a look at https://code.launchpad.net/~unity-team/unity8/caching-image-provider/+merge/224415 pls? it seems to be causing memory corruption and i have no idea why | 08:43 |
tsdgeos | mhr3: how does the memory corruption manifest? | 08:46 |
tsdgeos | i.e. what do i do to make it crash? | 08:46 |
mhr3 | Saviq, also, async image providers suck, qt is using a single thread to request them, so basically all requests to custom image providers get serialized | 08:47 |
mhr3 | tsdgeos, it's kinda random and i can make it happen only on armhf | 08:47 |
mhr3 | tsdgeos, basically play with it on phone and it crashes after a while | 08:47 |
mhr3 | tsdgeos, fwiw it's in 005 | 08:47 |
tsdgeos | ok | 08:48 |
Saviq | mhr3, requests, maybe, but the provider can return and fill the image later, no? | 08:48 |
Saviq | mhr3, if not, QTBUG | 08:48 |
mhr3 | Saviq, nope, the sync part of it (the actual ->requestImage() calls) is serialized cause it's being done by just one thread | 08:50 |
tsdgeos | mhr3: not the cause of the problem for sure but i'd say if (uri.search(/^http(s)?:/) == 0) { should be if (uri && uri.search(/^http(s)?:/) == 0) { since we're passing down stuff we previously were not checking was undefined or anything | 08:51 |
Saviq | mhr3, well, yeah, but I was under the impression that you can return an empty image and fill it later, isn't that how asyncing custom providers work? | 08:52 |
tsdgeos | Saviq: nope | 08:52 |
mhr3 | Saviq, not as far as i know, after you return the image, it will try to just upload it to the gpu, no? | 08:52 |
tsdgeos | async providers just promise they won't crash if running in a multithread env | 08:52 |
Saviq | mhr3, tsdgeos, :| | 08:53 |
Saviq | mhr3, OTOH, the provider is just a single object, right (as opposed to a factory)? | 08:53 |
Saviq | mhr3, so that kinda precludes multiple requests at the same time | 08:54 |
mhr3 | Saviq, not really, the docs do explicitely say that the methods image provider implements need to be reentrant | 08:54 |
Saviq | mhm | 08:54 |
Saviq | mhr3, in any case, that calls for a QTBUG to find out what's the deal and what can be done about it... like 90% of our SDK uses a custom image provider (via the image://scaled/ provider for @blah) :| | 08:55 |
mhr3 | well, the thread is per QQmlEngine, but i haven't seen there being multiple QmlEngine, when can that happen? | 08:57 |
tsdgeos | mhr3: i really can't find anything wrong in the code, but otoh i'm not really much into std::future (i should look at them more carefully) | 09:07 |
mhr3 | tsdgeos, i think it actually has something to do with the future/promise | 09:08 |
tsdgeos | mhr3: might as well be that the glibc/arm implementation of it is not totally perfect | 09:08 |
tsdgeos | it's not like stuff like that hasn't happened before | 09:09 |
mhr3 | dont even want to think about that | 09:09 |
mhr3 | tsdgeos, were you able to rep the crash? | 09:09 |
mhr3 | maybe it's just me in the end? :) | 09:10 |
tsdgeos | mhr3: still flashing the phone, just came back from a week of holidays today | 09:10 |
mhr3 | tsdgeos, right, sorry for jumping on you like that :) | 09:11 |
tsdgeos | it's ok :) | 09:11 |
mhr3 | guess i could try to reimplement it using simple mutex and condvar | 09:12 |
tsdgeos | may be one way to try to get it fixed | 09:23 |
tsdgeos | mhr3: if you give me soem time i'll try it on the phone and see if i can make it crash too | 09:23 |
Cimi | Saviq, I updated infographics apart moving infographics code to sdk, which I will start now | 09:41 |
Saviq | Cimi, ktx | 09:42 |
mhr3 | tsdgeos, think i fixed it, will rebuild the silo | 09:53 |
tsdgeos | mhr3: cool | 09:53 |
mhr3 | tsdgeos, oh, should add the js check.. how do i check if something is string? | 09:55 |
mhr3 | cause ints won't have .search() :) | 09:55 |
tsdgeos | mhr3: there's the tyeof thing | 09:55 |
mhr3 | - if (uri.search(/^http(s)?:/) == 0) { | 09:57 |
mhr3 | + if (typeof uri == "string" && uri.search(/^http(s)?:/) == 0) { | 09:57 |
mhr3 | ^^ k? | 09:57 |
=== davmor2_ is now known as davmor2 | ||
tsdgeos | triple equal to be more JS | 09:58 |
=== lool- is now known as lool | ||
mhr3 | Saviq, btw anyone trying u8 on the device with valgrind? it's *very* unhappy about things | 10:13 |
mhr3 | might have something to do with hybris, but still... at least having a suppressions file would be nice | 10:14 |
Saviq | mhr3, tsdgeos has experience | 10:23 |
tsdgeos | mhr3: i just ignore all that crap | 10:23 |
tsdgeos | :D | 10:23 |
tsdgeos | mhr3: agreed someone that knows about android should look at it and either fix it or provide a supressions file | 10:24 |
mhr3 | tsdgeos, do you also get the please use --workaround-gcc-bug thing? | 10:24 |
tsdgeos | hmmm | 10:24 |
tsdgeos | don't remember i get that no | 10:24 |
mhr3 | hm, when i tried it, it started to print hundreds of those, so after a while it just went - oh yea, too many error, i won't report any more | 10:25 |
tsdgeos | hmmm | 10:26 |
tsdgeos | mhr3: does adding --smc-check=all-non-file help? | 10:26 |
mhr3 | haven't tried that | 10:26 |
tsdgeos | give it a go | 10:27 |
tsdgeos | i think i'm using it | 10:27 |
mhr3 | yea, from reading about that, i guess it could help | 10:28 |
mhr3 | don't feel like trying to launch it under valgrind again, too much pain and takes ages :P | 10:29 |
tsdgeos | :D | 10:33 |
greyback | any chance we could get https://codereview.qt-project.org/#/c/84026/ landed in our Qt5.3? I really miss my scrollwheel in QtC | 10:40 |
tsdgeos | greyback: let's just update to 5.3.1 :D | 10:42 |
greyback | tsdgeos: or that :) | 10:42 |
tsdgeos | i'm confused :/ | 10:45 |
tsdgeos | http://paste.ubuntu.com/7725836/ | 10:45 |
tsdgeos | how can root be null :/ | 10:45 |
greyback | dat be f*cked up yo | 10:46 |
greyback | tsdgeos: does it crash? | 10:47 |
tsdgeos | greyback: i can make it assert or crash or work depending on the elements of that list i comment out | 10:47 |
greyback | tsdgeos: eeek | 10:48 |
tsdgeos | yep | 10:48 |
mhr3 | Saviq, ok, so i'm ready to land 005, seems like it's no longer crashing which is good :) the question is are we ok about loosing http pipelining? | 10:52 |
Saviq | mhr3, QNetworkAccessManager doesn't do it? | 10:53 |
mhr3 | Saviq, it does, but as i said earlier, requests from image provider will be serialized, so it will never ask for an image until the previous one finishes loading | 10:53 |
Saviq | mhr3, well, as it's cached... shouldn't it be better anyway? | 10:55 |
mhr3 | it should... once they fix the servers | 10:55 |
Saviq | mhr3, most times we shouldn't even hit the servers, right? | 10:55 |
Saviq | based on cache control? | 10:55 |
mhr3 | yea.. it's just they don't serve cache control yet :) | 10:55 |
Saviq | so it'll get worse before it gets better? ;) | 10:56 |
mhr3 | kinda... it will still use the cache, but it will have to validate it first | 10:56 |
Saviq | ah so just the HTTP header? | 10:56 |
mhr3 | yep | 10:57 |
mhr3 | fwiw that's like 90% of the latency :P | 10:57 |
Saviq | trueth | 10:57 |
mhr3 | but will still save quite some data | 10:57 |
Saviq | mhr3, what's the feel btw? | 10:57 |
mhr3 | Saviq, hm, let me reflash current image to be able to tell :) | 10:58 |
mhr3 | Saviq, or you can check 005 | 10:58 |
=== gatox_ultra is now known as gatox | ||
=== MacSlow is now known as MacSlow|lunch | ||
Saviq | dednick, hmm DashContent failure in move-indicator-qml now? | 11:18 |
dednick | Saviq: hm. i'll take a look | 11:18 |
sil2100 | Saviq: hi! On the last image we noticed a crash happening during unity8 tests, probably causing a test failure | 11:42 |
sil2100 | Saviq: psivaa filled in a bug for it: | 11:42 |
sil2100 | Saviq: https://bugs.launchpad.net/ubuntu-app-launch/+bug/1329141 <- | 11:42 |
ubot5 | Ubuntu bug 1329141 in qtdeclarative-opensource-src (Ubuntu) "qmlscene crashed while running test_can_launch_multiple_applications" [Undecided,Incomplete] | 11:42 |
sil2100 | Saviq: it might not be unity8 here at fault, but still I guess you would be able to judge the best | 11:43 |
Saviq | sil2100, if it's qmlscene crashing, then it's not unity8 at fault for sure ;) | 11:43 |
Saviq | sil2100, if you get some symbols, please let us know | 11:44 |
sil2100 | Sure :) | 11:47 |
mhr3 | Saviq, when you know what you're looking for, you can see that things take longer to load | 11:49 |
Saviq | mhr3, I'm starting to wonder... maybe we just need a caching proxy on the device :P | 11:50 |
Saviq | mhr3, what's the ETA on fixing the servers? | 11:53 |
mhr3 | Saviq, i did have the same idea... but shelved in the crazy folder :P | 11:53 |
Saviq | not so crazy now! | 11:53 |
mhr3 | Saviq, for click store it was on staging on friday, but had some issues | 11:54 |
mhr3 | Saviq, we could still go with what you did - set the cache on the global QmlEngine's network access manager | 11:55 |
mhr3 | then it will at least still do the pipelining | 11:56 |
mhr3 | but then we won't be able to cache scaled images | 11:57 |
Saviq | mhr3, I'm leaning towards that, we were just missing the per-request setting to use the cache on no network? | 11:57 |
mhr3 | that shouldn't even be necessary when the servers are fixed | 11:57 |
Saviq | well, I'd rather not rely on all servers being fixed ;) | 11:58 |
Saviq | mhr3, the scaling is most important for svgs, and we could put those through a different code path than the rest | 11:58 |
Cimi | Saviq, i am lost with the inline diffs | 12:05 |
Cimi | do you have a tip to see where you comment? | 12:05 |
Saviq | Cimi, click on "see inline diff" in the comment header | 12:05 |
Saviq | Cimi, and then scroll, unfortunately | 12:06 |
Cimi | Saviq, but sends me to your first comment | 12:06 |
Saviq | to find | 12:06 |
Cimi | not the new added | 12:06 |
Saviq | Cimi, right, if it's on the same revision, it's not great | 12:06 |
=== alan_g is now known as alan_g|lunch | ||
facundobatista | Holas | 12:08 |
Cimi | Saviq, I think it doesn't work great with the current implementation, can we stick to the old reviews until they fix the ui? | 12:09 |
Saviq | Cimi, no | 12:09 |
Saviq | Cimi, I will use it | 12:09 |
Cimi | I'm scrolling in this super long page trying to find your comments | 12:09 |
Saviq | it's not great, but it's still much better | 12:09 |
Cimi | and where you replied | 12:09 |
Saviq | Cimi, Ctrl+F | 12:10 |
Cimi | before was simply looking at the comment, easy | 12:10 |
Cimi | so I should ctrl f saviq? :) | 12:10 |
Saviq | Cimi, yeah, you know it was me that commented | 12:10 |
Cimi | Saviq, so basically what's wrong is the wrapper? | 12:11 |
Saviq | Cimi, its existence | 12:11 |
Cimi | I thought it was fine | 12:11 |
Cimi | abstract programming | 12:11 |
Saviq | Cimi, just for the purpose of having 2 props, there's no point in adding another object / hierarchy level | 12:12 |
Cimi | Saviq, that's because now we have two props | 12:12 |
Cimi | we might change implementation | 12:12 |
Cimi | anyway ok | 12:12 |
Saviq | Cimi, when we change the implementation so that it requires a wrapper, let's! | 12:13 |
Cimi | Saviq, so I should remove the first preload in tests/qmltests/CMakeLists.txt ? | 12:13 |
Saviq | Cimi, LD_ only matters for linking, for dynamically loaded libraries | 12:14 |
Cimi | I thought I needed it for test/mocks/Infographics | 12:14 |
Cimi | to avoid using thr system one | 12:14 |
Saviq | Cimi, ld doesn't know anything about QML, qmldir or whatnot | 12:14 |
Saviq | Cimi, that's what you have QML2_IMPORT_PATH for | 12:14 |
Cimi | ok | 12:14 |
Cimi | Saviq, although makes testing the infoghraphics harder | 12:17 |
Saviq | Cimi, what does? | 12:17 |
Cimi | Saviq, because you'd have to use the whole greeter then | 12:17 |
Saviq | Cimi, when? | 12:17 |
Cimi | Saviq, merging Greeter/Infographics.qml into Greeter/GreeterContent | 12:17 |
Saviq | Cimi, I'm not saying you should do that | 12:18 |
Cimi | Saviq, I'd have to move the tests into tstSingleGreeter | 12:18 |
Saviq | Cimi, I'm saying the Item in Infographics.qml is not needed | 12:18 |
Cimi | sh using CrossFadeImage directly? | 12:18 |
Saviq | yes | 12:18 |
Cimi | that was abvious | 12:18 |
Cimi | my bad | 12:18 |
=== MacSlow|lunch is now known as MacSlow | ||
Saviq | mhr3, so what do we do, what do we do? it's a rock and a hard place :| | 12:25 |
mhr3 | Saviq, i want proper async image providers :/ | 12:26 |
Saviq | mhr3, I *know*, and a way to override the default one... | 12:27 |
mhr3 | Saviq, if we're not going to do caching of scaled images, i think we should just override the network access manager | 12:28 |
mhr3 | Saviq, although you didn't like it's feel either when you tried it right? | 12:28 |
mhr3 | its* | 12:28 |
Saviq | mhr3, that was non-authoritative... | 12:28 |
Saviq | mhr3, I never tried on device | 12:28 |
Saviq | mhr3, and spent like 2 minutes on this | 12:29 |
mhr3 | Saviq, hm, let's revisit it, i'll build it in 005 and see if it's better | 12:32 |
Saviq | mhr3, yup, thanks | 12:32 |
mhr3 | Saviq, can you change owner of https://code.launchpad.net/~saviq/unity8/cache-network-data to unity-team? | 12:33 |
=== alan_g|lunch is now known as alan_g | ||
kgunn | tsdgeos: welcome back! | 13:11 |
tsdgeos | kgunn: hi | 13:12 |
kgunn | dednick: do you just need a volunteer for review ? | 13:12 |
kgunn | https://code.launchpad.net/~nick-dedekind/unity8/indicator.call-hint/+merge/218627 | 13:12 |
dednick | kgunn: hm. nearly. i'm busy fixing the tests. | 13:13 |
dednick | i'll put it back in wip | 13:13 |
kgunn | dednick: its fine... | 13:14 |
kgunn | pretty large code diff, i never realized it was that meaty | 13:14 |
paulliu | Saviq: if I want to test the logout Dialog stuff. I'll need to trigger a fake dbus event. Does that mean I need to go writing some code in tests/mocks/Unity/Session? Is tests/mocks for that? | 13:49 |
Saviq | paulliu, yeah, you want a mock object that doesn't actually talk DBus | 13:51 |
Saviq | paulliu, but you can control from the test | 13:51 |
paulliu | Saviq: ok. got it. | 13:51 |
Cimi | Saviq, branch updated | 13:51 |
Saviq | Cimi, k | 13:52 |
=== dandrader is now known as dandrader|afk | ||
=== alan_g is now known as alan_g|tea | ||
=== alan_g|tea is now known as alan_g | ||
=== dandrader|afk is now known as dandrader | ||
Saviq | dednick, could you go through https://code.launchpad.net/~unity-team/unity8/suru-switch/+merge/207991 please | 14:55 |
Saviq | dednick, stuff this depends on is built in silo 004 | 14:55 |
Saviq | dednick, ah crap, need to put the image thing in settings components, settings UI is broken with it, too :| | 14:57 |
dednick | Saviq: image thing? | 14:57 |
Saviq | dednick, I had to re-do SDK's Icon | 14:58 |
Saviq | dednick, to maintain aspect ratio of icons | 14:58 |
Saviq | dednick, bug #1284235 | 14:58 |
ubot5 | bug 1284235 in Ubuntu UI Toolkit "Icon has to be sized explicitly" [Undecided,Incomplete] https://launchpad.net/bugs/1284235 | 14:58 |
tsdgeos | dandrader: http://paste.ubuntu.com/7726801/ looks scary | 14:59 |
dednick | Saviq: property string iconPath: "/usr/share/icons/suru/status/scalable/%1.svg" - is that your fix for it picking up the lower prioirty icons? | 15:03 |
Saviq | dednick, not really, I need to load the images directly through Image | 15:03 |
Cimi | mterry, we have a crash on the wizard with the wifi | 15:03 |
Saviq | dednick, themed icons can only be square | 15:03 |
Saviq | dednick, side-effect is indeed that the priority problem is gone | 15:04 |
Saviq | dednick, it's a nasty hack, but we won't fix it properly in time | 15:04 |
Saviq | dednick, and we need to land this shit finally | 15:04 |
dednick | Saviq: uhuh :) | 15:05 |
Saviq | dednick, another solution welcome | 15:05 |
Cimi | mterry, https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1334203 | 15:05 |
ubot5 | Ubuntu bug 1334203 in ubuntu-system-settings (Ubuntu) "ubuntu-system-settings-wizard crashes/freezes when using the back button" [High,Incomplete] | 15:05 |
Cimi | mterry, I had a look, and I will have a look now | 15:05 |
Cimi | mterry, if you have ideas... throw | 15:06 |
dednick | Saviq: write a c++ image component? | 15:06 |
Cimi | mterry, from what I have seen, it's the property http://paste.ubuntu.com/7710506/ | 15:06 |
mterry | Cimi, I haven't looked into that yet. Not sure off the top of my head | 15:06 |
Saviq | dednick, that doesn't solve the problem that themed icons can only be square by definition of the icon theme | 15:07 |
Saviq | dednick, so would be a workaround just as the one I'm proposing | 15:07 |
mterry | Cimi, ah maybe destroying that property is buggy | 15:07 |
Saviq | dednick, the index.theme files only deal with one-dimensional "size" of the icon, resulting in the assumption that the icons are square | 15:07 |
Saviq | dednick, which means you'd need to load the image file to find the aspect ratio out, at which point the whole icon theme system becomes kinda useless | 15:08 |
mterry | Cimi, maybe also related to bug 1335298 | 15:08 |
ubot5 | bug 1335298 in ubuntu-system-settings (Ubuntu) "ubuntu-system-settings-wizard crashes on stop" [Undecided,New] https://launchpad.net/bugs/1335298 | 15:08 |
dednick | Saviq: hm. I'm sure the bluetooth icon was rectangluar at some point. | 15:08 |
Saviq | dednick, sure, that's why unity7 has its own loading implementation, too ;) | 15:09 |
dednick | Saviq: although that may have been due to some aspect | 15:09 |
Saviq | dednick, battery is rectangular | 15:09 |
Cimi | mterry, but not bug on our side? | 15:09 |
Cimi | mterry, do you know what can we do to workaround this property? | 15:09 |
mterry | Cimi, I don't know if it's a bug on our side or not | 15:10 |
Cimi | mterry, I don't think it is | 15:10 |
Cimi | mterry, commenting out that code makes the wizard stop crashing | 15:10 |
mterry | Cimi, does other code that uses UnityMenuAction have this problem, is what I'm wondering? | 15:11 |
mterry | Cimi, i.e. is it a problem with UnityMenuAction, or how we use it | 15:11 |
Cimi | dednick, about the crash with UnityMenuAction | 15:12 |
Cimi | http://paste.ubuntu.com/7710506/ | 15:12 |
Cimi | dednick, this makes the wizard crash on some occasions | 15:12 |
Cimi | dednick, do we have similar crashes somewhere else on the platform? | 15:12 |
Cimi | dednick, are we using it wrong in the wizard? | 15:13 |
dednick | Cimi: can you send a stacktrace? | 15:13 |
Cimi | not sure hot to get it | 15:13 |
Cimi | mterry, to get stacktrace we attach gdb right? | 15:13 |
mterry | Cimi, yeah | 15:13 |
Cimi | which dbg symbols we need? | 15:13 |
dednick | Cimi: apport-cli /var/crash/XXX.crash | 15:14 |
Cimi | mterry, in order to reproduce bug I need to have wifi off | 15:14 |
Cimi | so might try to manually get dbg | 15:14 |
mterry | Cimi, you probably want qtbase5-dbg at least | 15:14 |
mterry | Cimi, not sure which package has UnityMenuAction | 15:15 |
Cimi | mterry, qmenumodel | 15:15 |
Cimi | but I don't see dbg for it | 15:16 |
dednick | libqmenumodel | 15:16 |
Cimi | dednick, and debug symbols? | 15:16 |
=== Malsasa_ is now known as Malsasa | ||
dednick | Cimi: libqmenumodel0-dbgsym | 15:19 |
Cimi | it's not on the phone | 15:19 |
Cimi | what shall I enable? | 15:19 |
Cimi | who knows guys? | 15:20 |
dednick | Cimi: https://wiki.ubuntu.com/DebuggingProgramCrash | 15:20 |
dednick | on phone presumably | 15:20 |
dednick | Saviq: you need me to check that branch out now, or you doing something in usc for it? | 15:21 |
Saviq | dednick, yeah, I'll have another usc branch for this in a few mins hopefully | 15:22 |
Cimi | dednick, mterry http://paste.ubuntu.com/7726901/ | 15:23 |
Cimi | so yeah I need those symbols | 15:23 |
dednick | Cimi: that trace was fine. | 15:27 |
=== pete-woods is now known as pete-woods-eodlu | ||
dednick | Cimi: where is the code? | 15:31 |
Cimi | dednick, http://bazaar.launchpad.net/~system-settings-touch/ubuntu-system-settings/trunk/view/head:/wizard/qml/Pages/20-wifi.qml | 15:34 |
Cimi | line 121 I think | 15:34 |
tsdgeos | kgunn: Saviq: some managementarial decision needed at https://bugs.launchpad.net/unity8/+bug/1332598 | 15:35 |
ubot5 | Ubuntu bug 1332598 in Unity 8 "make testShell crashes" [Critical,In progress] | 15:35 |
tsdgeos | upstream guys that can decide on that bug seem to be away for a while | 15:36 |
Saviq | dednick, https://code.launchpad.net/~unity-team/ubuntu-settings-components/status-icon/+merge/225034 | 15:37 |
tsdgeos | dandrader: can you confirm that your crash is also gone if you use the patch from https://codereview.qt-project.org/#/c/88717/ ? | 15:37 |
dednick | hm. network indicator doesnt seem to work on desktop anymore... | 15:38 |
dednick | tedg: ^ ? | 15:38 |
dandrader | tsdgeos, will try it now | 15:38 |
dandrader | tsdgeos, yes, it does fix it! | 15:43 |
tsdgeos | Saviq: so how do we proceed in here? ↑ | 15:44 |
dednick | Cimi: hm. doesnt seem to crash on desktop. | 15:44 |
dednick | Cimi: does it always happen under certain cirumstances? how to reproduce? | 15:46 |
dandrader | mzanetti, hey, what's the situation with "dash as an app"? | 15:49 |
dandrader | ie, why did we stop working on it? | 15:50 |
mzanetti | dandrader: we decided to land that separately, no? | 15:50 |
mzanetti | dandrader: we didn't really stop working on it | 15:50 |
dandrader | mzanetti, I don't remember. that's why I'm asking :) | 15:50 |
kgunn | tsdgeos: so iiuc, upstream says "your patch is our best guess" | 15:50 |
kgunn | "use it, until we fix it" | 15:50 |
mzanetti | dandrader: that's one step towards that: https://code.launchpad.net/~mzanetti/unity8/new-header/+merge/224585 | 15:50 |
tsdgeos | kgunn: upstream says "the guys that knows how this works are on holiday, your patch doesn't look terribly wrong to the guys left" | 15:50 |
mzanetti | dandrader: once this has landed there shouldn't be any links between shell and dash any more in the code. | 15:51 |
mzanetti | dandrader: I already have a branch here that creates an application binary for it | 15:51 |
dednick_ | Saviq: has that usc branch gone into the silo? | 15:51 |
kgunn | tsdgeos: based on your experience & dandrader, we should go for a distro patch & stick it in ci-train...we can always ask for more extensive testing | 15:51 |
tsdgeos | kgunn: +1 | 15:51 |
mzanetti | dandrader: one thing missing is we need ApplicationManager pick up applications started by upstart (as opposed to ubuntu-app-launch) so the dash would appear in the running apps list | 15:52 |
kgunn | e.g. get a qa resource at least, and maybe pull in some app folk | 15:52 |
dednick_ | Saviq: ubuntu-settings-components branch that is | 15:52 |
tsdgeos | does anyone know if Mirv is on holiday? | 15:52 |
mzanetti | dandrader: I decided not to make Gerry do that before we have QtComp in, or at least in the silo | 15:52 |
kgunn | tsdgeos: he was... | 15:52 |
tsdgeos | he always does the ci-train-ing of Qt stuff for me | 15:52 |
kgunn | mmm he's Finnish....might be long vacation | 15:52 |
Saviq | dednick_, not yet, lemme add | 15:53 |
mzanetti | kgunn: tsdgeos: yeah, I think its another week | 15:53 |
dednick_ | Cimi: not sure if you got my message. How do you reproduce the crash? doesnt seem to occur on desktop | 15:53 |
kgunn | mzanetti: did you get a Qt-patch in ci-train ? | 15:53 |
kgunn | ...i know the process is a little diff | 15:53 |
mzanetti | into the train yes, not sure if it made it into the image yet | 15:53 |
dandrader | mzanetti, because I think "dash as an app" is an important piece in the "shell rotation" story | 15:54 |
mzanetti | dandrader: I agree that might make things a lot easier | 15:54 |
kgunn | mzanetti: did you use a canonical branch ? or just local mod and give src ball to core dev | 15:54 |
kgunn | assuming you did this w/o Mirv | 15:54 |
mzanetti | dandrader: well. afaik we were also not enabling rotation for the first batch of QtComp goodness | 15:54 |
dandrader | mzanetti, on the phone. Dash stays in portrait but the indicartors panel, launcher etc do rotate according to app preferences | 15:54 |
mzanetti | kgunn: I used the kubuntu-packagers branch and proposed my change towards that | 15:55 |
mzanetti | kgunn: after I figured we got the packages from that one lately | 15:55 |
dandrader | mzanetti, and we hopefully also mitigate (or even eliminate) that horrible ~2 secs gui choke when resizing unity8 due to a rotation | 15:55 |
mzanetti | dandrader: huh? we're going to rotate launcher and panel but not the rest? | 15:55 |
* mzanetti wonders how we even do that | 15:56 | |
dandrader | mzanetti, with dash as an app? | 15:56 |
Saviq | mzanetti, well, dash might choke, shell won't ;) | 15:56 |
mzanetti | ah... yeah, once its an app, yes | 15:56 |
mzanetti | I thought dandrader was doing that with current codebase | 15:57 |
Saviq | dednick_, is building | 15:57 |
mzanetti | dandrader: so yeah, after QtComp is in the silo and requires less attention we're going to do the appman changes and finish my branch with the separate app for the dash | 15:58 |
mzanetti | dandrader: then you can finalize the rotation along with that | 15:58 |
Cimi | dednick_, on the phone | 15:59 |
Cimi | dednick_, https://bugs.launchpad.net/qmenumodel/+bug/1334203 | 16:00 |
ubot5 | Ubuntu bug 1334203 in ubuntu-system-settings (Ubuntu) "ubuntu-system-settings-wizard crashes/freezes when using the back button" [High,New] | 16:00 |
Cimi | dednick_, basically on the phone, you have to do | 16:00 |
Cimi | rm /etc/NetworkManager/system-connections/% | 16:01 |
Cimi | rm /etc/NetworkManager/system-connections/* | 16:01 |
Cimi | then | 16:01 |
Cimi | rm /home/phablet/.config/ubuntu-system-settings/wizard-has-run | 16:01 |
Cimi | reboot | 16:01 |
Cimi | at the wizard, wait till it loads wifi then you tap back | 16:02 |
Cimi | happens 100% here | 16:02 |
kgunn | tsdgeos: are you familiar with that part of the process ? (see zanetti's post wrt kubuntu-packagers) | 16:03 |
mzanetti | tsdgeos, kgunn: fyi, this is the branch: https://code.launchpad.net/~mzanetti/kubuntu-packaging/qtdeclarative-opensource-src/+merge/224517 | 16:04 |
kgunn | yep...i see it now, tsdgeos, if you can upload in the same fashion & get me the mp, i can get you on the ci-train | 16:06 |
dednick_ | Cimi: need to connect to a wifi hotspot? | 16:06 |
tsdgeos | kgunn: tomorrow :) | 16:06 |
kgunn | tsdgeos: no worries...you can hit up Saviq you're morning, he can get you on the train | 16:07 |
tsdgeos | cool :) | 16:07 |
* tsdgeos leaves for the day | 16:07 | |
* kgunn feels qt code control less out of control than he thot... | 16:08 | |
kgunn | still seems funky | 16:08 |
Cimi | dednick_, no | 16:09 |
dednick_ | Cimi: didnt work for me. | 16:10 |
Cimi | dednick_, clear as I said | 16:10 |
dednick_ | Cimi: i'm flashing device. | 16:10 |
Cimi | dednick_, do not connect to any wifi | 16:10 |
dednick_ | Cimi: i didnt | 16:10 |
Cimi | dednick_, otherwise bootstrap | 16:10 |
dednick_ | Cimi: heh. good luck getting me to do that ;) | 16:10 |
Cimi | dednick_, https://bugs.launchpad.net/qmenumodel/+bug/1334203 | 16:10 |
ubot5 | Ubuntu bug 1334203 in ubuntu-system-settings (Ubuntu) "ubuntu-system-settings-wizard crashes/freezes when using the back button" [High,New] | 16:10 |
=== om26er is now known as om26er|away | ||
=== gatox is now known as gatox_lunch | ||
Cimi | dednick_, luck? | 16:28 |
dednick_ | Cimi: :( nope | 16:29 |
Cimi | dednick_, do you want me to test code? | 16:29 |
Cimi | dednick, do you want me to test code? | 16:35 |
dednick | Cimi: give me a minute. | 16:35 |
Cimi | ok | 16:35 |
=== dandrader is now known as dandrader|lunch | ||
=== om26er|away is now known as om26er | ||
=== alan_g is now known as alan_g|EOD | ||
dednick | Cimi: reproduced. | 17:12 |
=== gatox_lunch is now known as gatox | ||
mhr3 | Saviq, here? | 17:32 |
mhr3 | Saviq, when you will be - https://code.launchpad.net/~unity-team/unity8/cache-network-data/+merge/224995 seems to work fine and ready for landing in 005 | 17:34 |
dednick | Cimi: https://code.launchpad.net/~nick-dedekind/qmenumodel/unitymenuaction.lp1334203 | 17:45 |
=== dandrader|lunch is now known as dandrader | ||
Saviq | mhr3, kk | 19:42 |
=== dandrader is now known as dandrader|afk | ||
Saviq | seb128, could you please resubmit https://code.launchpad.net/~seb128/ubuntu-system-settings/use-theme-icons/+merge/214950 using lp:~unity-team/ubuntu-system-settings/use-theme-icons instead? | 20:31 |
=== Daviey_ is now known as Daviey | ||
=== dandrader|afk is now known as dandrader | ||
=== salem_ is now known as _salem |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!