[05:27] Hi, I want to know if it's possible to run custom shell commands from the dash, for example a function created in .bashrc === tvoss|sick is now known as tvoss [05:59] anyone with 3 or 4 monitors? https://bugs.launchpad.net/unity-2d/+bug/930147 would need testing on 12.04 [05:59] Launchpad bug 930147 in unity-2d "all panels have same window focus on multiple monitors unity2d" [Low,Fix committed] [06:03] wow, three monitors. [06:03] stupid intel GPUs only have 2 PLLs [06:07] hey Mirv [06:07] Mirv: any progress on the skype/qt4 issue? [06:29] hey didrocks. I didn't get new answers yesterday, but I haven't found any similar crash reports on other libqtwebkit4 rdepends (like amarok, arora, k3b, ktorrent, rekonq, quassel), so I believe it's skype only. [06:30] but I'm not certain if anyone is willing to upload skype with the LD_PRELOAD hack... [06:30] Mirv: do you mind adding a comment on the bug report + ping jdstrand who had concerns with it? [06:30] didrocks: ok [06:30] Mirv: if he agrees, I can sponsor it :) [06:30] (the hack) [06:30] Mirv: just ping me so that we can conclude that today if possible :) [06:32] ok, I'll ping + send e-mail [06:32] thanks ;) [07:13] hi [07:17] good morning [07:17] this is about ubuntu interface ? [07:19] jim00234: yeah [07:22] why they choose a tablet interface for a desktop os? [07:27] because it's not a tablet interface [07:47] hey sil2100, how is it going? [07:48] didrocks: hi! Had to do tax-related stuff in the morning so I'm starting just now ;) [07:48] sil2100: did you get to the new package yesterday evening, or not yet? [07:48] sil2100: what's on your plate, to sum up? ;) [07:49] (and what I've eventually to review since Tuesday) [07:50] didrocks: I have the packaging prepared according to the guidelines for android-audiosystem, will be doing testing on my system yet before proposing a merge [07:50] sil2100: excellent! do not hesitate to poke me for a review :) [07:51] sil2100: how are you going to test it manually btw? just curious :) [07:51] I had some build problems due to the changes yesterday ;p some strange dpkg- things [07:51] sil2100: needing help? [07:54] mzanetti: any reason we don't publish generic-mediumtests-builder jobs in the public jenkins? [07:56] tsdgeos: yeah... because it's not supposed to fail unless all the other builders fail too. [07:56] :D [07:56] tsdgeos: which has been working 100% reliable until yesterday [07:56] seen the answer i just gave to michi? [07:56] no [07:56] link? [07:57] https://code.launchpad.net/~michihenning/unity/phablet-coverage-tests/+merge/157781 [07:59] mzanetti: do you have a minute for hangout? I want to discuss some tests-autopilot-jenkins-job-thingy-yes-I-cant-make-a-more-fuzzy-title-but-will-try ;) [07:59] Saviq: pushed the nested proxy rolenames test [08:00] tsdgeos: really weird... the mediumtests-builder is exactly the same as the i386-ci expect it builds without coverage measurements and does not execute tests [08:00] tsdgeos, nice [08:00] didrocks: sure [08:00] mzanetti: shrug [08:01] tsdgeos: well... it does execute the core "make check" suite [08:01] tsdgeos: as pbuilder does that on its own [08:01] right [08:01] mzanetti: https://plus.google.com/hangouts/_/1ce5181c7e07bfd3df3ccb932563480d1966d625?authuser=0&hl=en [08:52] What's the best way to type check in QML? [08:53] nic-doffay, there's no way to do that unfortunately [08:53] nic-doffay, the only thing is «typeof something, "object"» [08:53] nic-doffay: https://bugreports.qt-project.org/browse/QTBUG-24799 [08:54] Saviq, hmm ok. [08:54] nic-doffay, we could try and use a hack [08:54] nic-doffay, by declaring properties of certain types [08:54] and trying to assign the values to them [08:54] * mzanetti wonders if the problem isn't somewhere else... [08:55] I haven't felt the need for it in more than 10000 lines of QML [08:55] mzanetti, it's for the interface tests [08:55] mzanetti, when we want to make sure that an exposed property is of a certain type [08:56] tsdgeos, so that really means the QTBUG re: nested proxy roleNames is fixed? [08:57] Saviq: not really, the bug is about nested proxies, not nested "our" proxies [08:57] Saviq: hmm, ok... how about something like this? http://stackoverflow.com/questions/332422/how-do-i-get-the-name-of-an-objects-type-in-javascript [08:57] didrocks: hmmm... [08:57] didrocks: can I poke you about something? [08:57] sil2100: isn't what you have just done? ;-) sure, what's up? [08:58] tsdgeos, hmm right [08:58] Saviq: next step in "improving" our proxy is checking if they correctly return sourceModel()->roleNames from the QSortFilterProxyModel and then remove our implementation, but i'd like to defer that a bit (and get this one merged), working on the hud prototype now [08:58] tsdgeos, sure, approved already [08:59] i did fix it for the QIdentityProxyModel almost for sure [08:59] don't remember if also proposed a fix for the SortFilter one [08:59] didrocks: so... I have a situation I don't really understand ;) The android-audiosystem package, I wanted to build it locally but it fails - the reason is, the debian/rules they use uses dh and override_dh_auto_build/clean/install [08:59] didrocks: the package itself uses .sh scripts to build, so there's no 'real' buildsystem that dh can use [08:59] didrocks: now, the thing I don't understand... [09:00] sil2100: hangout + screensharing maybe? [09:00] we can look at that together :) [09:00] Does anyone know if it's possible to get a view on all blueprint workitems assigned to a user? [09:00] didrocks: when building it locally with bzr bd, it fails building, as the override hooks for build, clean and install doesn't work [09:00] sil2100: maybe a good training for both of us ;) [09:00] Saviq, just a reminder about that email for the interface tests with a summary of what needs to be done. [09:00] didrocks: but when building on a PPA, I saw it working fine :o OK, one moment and I'll prepare for a hangout [09:01] Got to connect the power and things like that [09:01] sil2100: and dress up! :p [09:01] nic-doffay, yeah, I'm on it, but doing it properly from the get go, so might not make it today [09:03] No problem, still got things to implement. Taking slightly longer due to QML noobness too. [09:05] didrocks: invited you [09:09] mzanetti, I'll have a look, thanks [09:12] mzanetti, "[native code]" :/ [09:13] mzanetti, the constructor.toString() just tries to print the source of the constructor [09:13] but obviously that only works in JS [09:13] Saviq: right... [09:13] [native code] :D Hooray for propert code [09:13] proper [09:15] and even root.constructor.name is empty [09:15] where root is a Rectangle [09:15] it's actually an empty string [09:16] so yeah, the only solution I can think of is declaring a property of a type [09:16] and check if assignment works [09:16] but that's probably not really reliable... [09:19] hmm seems to work, actually [09:21] mzanetti, what do you think? [09:22] Saviq: I could immagine implicit casts happening at some cases [09:22] mzanetti, yeah, but if the property is of a type [09:23] Saviq: what kind of type? [09:23] mzanetti, well, you're right [09:23] I don't think you can distinguish a ByteArray from a String [09:23] while it might work for QObject* types [09:24] but especially basic types will be implicitly casted, and Polymorphic types might be problematic too [09:24] mzanetti, well, basic types we can check with typeof [09:24] mzanetti, the only problematic ones are objects [09:24] and enums, which are just numbers, but that should be somewhat fine [09:25] still, it would be something [09:25] Saviq: not really sure if the approach of checking types in a untyped language is the right one [09:25] maybe we should have some C++ tests for that? [09:26] mzanetti, we want to test the QML interface [09:26] and sure, we could test it in C++ [09:26] and it actually will be tested in C++, too [09:27] I guess the most sensible approach would be to split them... checking types of QML items in javascript feels like the sledgehammer method [09:27] mzanetti, and well, QML is typed [09:27] Saviq: yeah... but the engine it runs in is not [09:28] mzanetti, so deferring to it (by trying to assign a value to a typed property) to QML [09:28] Saviq: might as well be enough... actually if QML casts it implecitly and the test passes it will most certainly also work in real code - casted or not [09:28] mzanetti, I'd really like to avoid splitting the test, it's supposed to be a definition of the API in a sense [09:29] yeah exactly [09:29] mzanetti, and obviously when the aforementioned bug is fixed [09:30] right... that would help, yeah [09:30] we can move to a more proper way, but I'd rather not split it out [09:32] Saviq: the bug was last edited in Jul 2012 by a guy who has most likely been fired the same day as I was... wouldn't hope too much tbh [09:32] mzanetti, lol [09:33] mzanetti, yeah, I rather thought we'd have to take care of it if we really want it [09:33] but that's fine [09:34] Saviq: but doesn't sound too hard actually... How about creating a simple qml plugin that just the type comparison as a C++ method? I guess that could be possible [09:34] mzanetti, indeed [09:34] or would you need to include all the types there? [09:35] mzanetti, don't think so [09:35] mzanetti, if you define a component in pure QML [09:35] you can still pass it back to C++ [09:35] question is how do you get its type hierarchy, then [09:36] Saviq: for QObjects it would be easy [09:36] and I guess to pass them to QML they have to be QObjects anyways [09:36] mzanetti, exactly [09:36] obj->metaObject() [09:36] right [09:37] * Saviq hacks [09:40] did someone fix the startup crash of qml-phone-shell? [09:49] mzanetti, what startup crash? the one on the phone [09:49] ? [09:49] Saviq: no, on the desktop [09:50] Saviq: but its back now... just hat a lot of lucky runs [09:50] mzanetti, I barely ever see it crash [09:51] Saviq: happens in most of the cases for me [09:51] mzanetti, but I did manage to make Qt crash, by trying to pass a component (not an instance) to C++ ;) [09:51] Saviq: passing -geometry 1600x1400 or so decreases likelyhood [09:51] hehe [09:51] we'll have to compare by string, unfortunately [09:52] Saviq: can you remember what exactly was the reason for copying the ListItems? [09:52] mzanetti, there were a bunch, just look at the diffs [09:52] mzanetti, greyback was the one that did most of it, IIRC [09:53] mzanetti: which listitems? [09:53] seb128: hey, a friend of mine has just upgraded to 13.04, but his applications/defaults.list had not the proper handler for inode/directory... Is that only applied after fresh-install? [09:53] greyback: Components/ListItems/* [09:54] greyback: unity-phablet [09:54] seb128: I didn't do that manually as far i remember [09:54] Trevinho, maybe he has a local config for his user? [09:54] Trevinho, we don't overwrite user configs [09:54] seb128: no, changing it manually on /usr/share/applications, fixed it [09:54] weird [09:54] seb128: it was only a global thing [09:54] yeah [09:55] well it you changed it manually we can't get debug infos now [09:55] mzanetti: they're heavily customized versions of SDK ListItems. [09:55] seb128: damn :P, I can revert it back... :P [09:55] Trevinho, next time somebody has the issue tell them to keep the system in buggy state for debugging [09:55] Trevinho, that will not restore the timestamp from before you changed it... [09:55] greyback: yeah. I'm in the process of determining if we need to keep them or if we should go back to upstream ones at some point [09:55] Trevinho, so we don't know when the file was updated before your edit [09:56] seb128: oh... I'm sorry :/ [09:56] Trevinho, it's supposed to work on update, dunno what happened there [09:56] Trevinho, no worry [09:56] Trevinho, dpkg -l | grep desktop-file-utils [09:56] seb128: fine, I've other PC to upgrade anyway [09:57] mzanetti: mostly the customisations are for lists with section headers which have transparent backgrounds - and the section headers stick to the top of the view. Right now listview draws the section header on top of the delegate, but if there's transparency you see both. [09:57] Trevinho, can you get the dpkg info? [09:58] Trevinho, did you get a conffile diff prompt during the upgrade? [09:58] mzanetti: It's a workaround for what I suspect is a Qt bug [09:58] Trevinho, the issue is that the file is stored in /etc, so it preserve local changes [09:58] Trevinho, and some buggy software edit the files for you, which makes the system think you edited the file and want your old version preserve [09:59] Trevinho, acrobat reader does that iirc, and some others [09:59] greyback: I find it hard to believe that the ListView section header has such a bug [09:59] seb128: ah, he has 0.21-0ubuntu3 anyway [10:00] seb128: but if one changed that file should't it keep the .dkpg-old or original? [10:00] mzanetti: open the SDK component gallery. You'll see it there right away in the list on the left [10:00] Trevinho, well, when there is a change it asks you what to do on upgrade [10:00] seb128: yeah, sure [10:00] Trevinho, it shows you the diff and ask if you want to install the version of the package or keep your local one [10:01] Trevinho, ls /etc/gnome/defaults.list* [10:01] ? [10:01] greyback: where can I find that again? [10:01] the components gallery [10:02] mzanetti, ubuntu-ui-toolkit-examples [10:03] mzanetti: From the Dash, search for "Qt Components". You may have it installed. If not, lp:ubuntu-ui-toolkit, qmake, make, and run ./launch-demos. [10:03] no dash here [10:03] but found it [10:03] * Trevinho running between rooms like a crazy... [10:03] mzanetti, /usr/lib/ubuntu-ui-toolkit/demos/launch_componentshowcase [10:04] seb128: yeah, so it seems that the old one was kept [10:04] mzanetti, and I don't actually think is a bug in its usual definition, it's just an omission [10:04] seb128: apparently there were some google- earth and acroreader changes on it [10:04] Saviq, greyback: yeah... QML does not do any clipping on its own. [10:04] Trevinho, that's your issue, those are buggy and modify the file [10:05] seb128: yeah :/ [10:05] in our case that does suck indeed a little bit [10:05] mzanetti, yeah, there should really be an option to enable clipping under the section header [10:05] seb128: what would be the proper way for doing that for them? just a proper .desktop file or a local setting? [10:05] mzanetti, it's just assuming you will always have a background in your section header [10:05] Saviq: InverseClippingArea :D [10:05] mzanetti, ;) [10:06] mzanetti, if I go `some_function(Rectangle)`, do you know of a way to then create an object of that type? [10:06] Trevinho, what mimetype google-earth needs to claim? if nothing else is using the same mimetype they don't need to set a default [10:06] Trevinho, acroread is a bit trickier [10:07] * mzanetti thinks [10:07] mzanetti, or is that at all supported [10:07] Trevinho, the current xdg spec doesn't allow for a new installed package to take over a default, opensuse has hacks for that and they suggested them on the xdg list some years ago [10:07] Trevinho, but that never went to a conclusion [10:07] mzanetti, I seem to get an Object from that just fine, but it doesn't seem to have any members :/ [10:07] seb128: ah, I see [10:08] seb128: I guess that using a defaults.list.d/ would be a solution, isn't it? [10:08] seb128: with priority-names .... i.e. 00_adobe.....list [10:08] Saviq: I would need to try experimenting... could immagine to get a Loader somehow to create an object from its name [10:09] mzanetti, don't think so, it either takes a Component (which Rectangle is not) [10:09] mzanetti, or a source file [10:09] same as Qt.createComponent, really [10:09] Qt.createObject() ? [10:10] or was it Component.createObject()... that wouldn't help either in that case :/ [10:10] yeah, it was Component [10:10] there is Qt.createQmlObject [10:10] but that doesn't help either [10:12] Trevinho, right [10:13] Trevinho, some of the discussion/proposal for improvements were around "by desktop default" as well [10:13] Saviq: what you could do - not the most elegant thing though - is to have a file with all the Types wrapped in a Component {} and a getByName() that returns the Component... [10:13] let's make new standards! :) [10:13] Trevinho, e.g http://lists.freedesktop.org/archives/xdg/2011-February/011806.html [10:13] Saviq: actually pretty ugly... [10:13] mzanetti, yeah, what with custom types [10:13] mzanetti, I'll just go for string comparison [10:13] Saviq: yeah... would require constant extending [10:15] greyback: so in your opinion we can't get rid of them but rather should write test cases in our codebase? [10:16] * mzanetti wonders how to test clipping - which is really the only thing they do :D [10:16] mzanetti: for now we're stuck with them. I also think it's a bug, or at least an omission in Qt. [10:17] mzanetti: you could test to make sure that the math is correct, so it's clipping the correct height [10:21] greyback, but some of the changes were upstreamable to SDK, no? [10:21] mzanetti, please make sure of that ^ [10:22] Saviq: mzanetti: yes they could be upstreamed, that would be option 2. But I think option1 is seeing if we can fix ListView [10:22] greyback, right [10:22] greyback, but there were some other changes [10:23] greyback, like the height of the item to (not) include the separator [10:24] Saviq: yep, those bits can go upstream, but I know the SDK's ListItems are due to be reviewed heavily so I'd be tempted to hang on for that first [10:24] right [10:25] +1 [10:25] that doesn't seem to be a task to do in between our testing efforts [10:26] Saviq: I'll postpone them then... given that the basics (size and clicking) are already tested through the peoplepreview now [10:27] mzanetti, k [10:31] this is ridiculous [10:31] 14/16 Test #14: whitespace ...........................***Failed 0.55 sec [10:31] /home/phablet/shell/builddir/CMakeFiles/CompilerIdC/CMakeCCompilerId.c: trailing whitespace in lines 224, 279 [10:31] /home/phablet/shell/builddir/CMakeFiles/CompilerIdCXX/CMakeCXXCompilerId.cpp: trailing whitespace in lines 212, 267 [10:31] really? [10:35] tsdgeos: I don't think this is intentional [10:36] tsdgeos: but yeah... in my opinion we should do license and whitespace checks _before_ the build run, not afterwards, or this kind of things will keep on biting [10:43] tsdgeos, yes, really [10:44] tsdgeos, just make test before you push [10:44] tsdgeos, ah [10:44] paulliu: you around? [10:44] Saviq: and how do i fix whitespace there? [10:44] if i may ask? ;-) [10:44] tsdgeos, wait, that should be handled [10:44] tsdgeos, the script ignores stuff in ${CMAKE_BINARY_DIR} [10:44] it does [10:44] or, should [10:44] it is true [10:45] the problem is that i was dpkg-building [10:45] tsdgeos, should work just the same [10:45] on a "dirty" checkout [10:45] tsdgeos, ah [10:45] tsdgeos, :P [10:45] tsdgeos, bzr bd -S [10:45] so the builddir was not really the ${CMAKE_BINARY_DIR} [10:45] tsdgeos, or bzr bd [10:45] tsdgeos, that will only take files checked in (modified, too) [10:46] mzanetti, did you ever get expectFail() in QtQuickTest to work? [10:47] Saviq: never tried tbh [10:47] Saviq: didn't know existed [10:47] ;) [10:47] Saviq: I use verify(something != something) or the like [10:48] mzanetti, yeah, not working with exceptions ;) [10:48] mzanetti, http://pastebin.ubuntu.com/5701083/ that, more or less, works [10:48] Saviq: don't use exceptions with Qt :D [10:48] mzanetti, yeah, it was automagic [10:48] need to protect [10:49] Saviq: isInstanceOf.... is that the thing you hacked together? [10:49] mzanetti, yeah [10:49] Saviq: and you attached it to every object? [10:50] mzanetti, no, not yet [10:50] I guess I'll see a MP that explains it at some point... just curious how you did it [10:50] mzanetti, just a Q_INVOKABLE now [10:50] mzanetti, testObject is just an object from a plugin [10:50] Saviq: ah... ok... net yet the actual objects to be tsted [10:51] mzanetti, yes they are [10:51] * Saviq pastes the whole thing [10:53] mzanetti, http://pastebin.ubuntu.com/5701091/ [10:54] Saviq: ok... got it... this line comfused me: testObject.isInstanceOf(testObject, "QQuickItem") [10:55] Saviq: I got the impression that suddenly all objects know of isInstanceOf [10:55] mzanetti, yeah, WIP [10:56] mzanetti, I think that should do for now [10:56] Saviq: seems good imho [10:56] Saviq: autopilot does the same btw [10:57] mzanetti, I'll build it into UnityTestCase [10:57] Saviq: while you're at it... mind moving UnityTestCase to /tests/imports/ [10:57] Saviq: we're using it in unittests too now [10:57] mzanetti, yup [10:57] thanks [11:00] hmm === MacSlow is now known as MacSlow|lunch [11:01] hey guys [11:02] mzanetti, tsdgeos hmm, plugins vs. imports vs. modules? [11:02] Saviq: good question [11:02] Saviq: context? [11:02] hey kgunn [11:02] mzanetti, tsdgeos, SDK uses "modules", we use "plugins", mzanetti just mentioned "imports" [11:03] hey kgunn [11:03] Saviq: are 3 different things [11:03] Saviq: a Module would be something like Ubuntu.Components [11:03] Saviq: plugins is c++ stuff loaded as a plugin into the enigne [11:03] Saviq: imports is small qml or .js file imports [11:04] mzanetti, Ubuntu.Components has a plugin, too [11:04] Saviq: yeah... a module can consist of plugins and imports [11:04] mzanetti, now we're talking ;) [11:04] QQuick 2.0 is a module [11:05] mzanetti, so module[plugins (C++), imports (qml)] [11:05] Saviq: rule of thumb, yeah... [11:05] mzanetti, but then do we want them separately in our root? [11:05] that's 3 import paths where one should generally suffice [11:05] and then [11:05] Saviq: you thinking about moving the FakeHud plugin etc too? [11:06] mzanetti, considering [11:06] mzanetti, want to avoid duplicates [11:06] mzanetti, where we, for example, need mocked Unity or something [11:06] Saviq: hmm... could make sense... could cause things I don't see now though [11:06] mzanetti, so yeah that was my question, too [11:06] but yeah... I guess a mock-hud plugin could indeed be useful in some other places than the hud qmluitests [11:06] should we have /modules, /plugins /imports /tests/modules /tests/plugins /tests/imports [11:07] Saviq: doesn't sound like a friday afternoon task, but would sound like a clean solution, yeah [11:07] mzanetti, lol [11:07] mzanetti, you've learned something there :D [11:08] :D [11:08] So, Im on Kubuntu and I have the unity webapps thing coming up on my chromium. I have 3 questions: 1. what is this?/what does it do? 2. If its useful, how do I install it (it currently gives me a html file of a LP page when I try install it) 3. if its not useful on kubuntu, how do I get rid of it... ? [11:10] jussi:+1. I think its weird that all the browsers pop up those webapps thingies when not running unity... however, I don't think #unity is the right channel to ask this :D [11:10] jussi: try something like #webapps [11:10] jussi: i don't think it does anything in kubuntu, just click "don't ask me anymore". That's what i did [11:10] hrm, #webapps ? not seen that, but Ill give it a try... [11:11] mzanetti: that channel doesnt exist :/ [11:11] tsdgeos: I dont have that option... :/ [11:12] jussi: #ubuntu-webapps [11:12] it was there in firefox at least afair [11:12] tsdgeos: http://i.imgur.com/tIdd2yN.png [11:12] mzanetti: ahh, that makes sense [11:12] ok, that's definitely not what i got in firefox [11:13] it was more combo-like [11:13] and you could say "ignore me forever" [11:14] yeah, I asked in -webapps, lets see what they say [11:44] we have a typo in the license checker :D [11:44] Found 1 license prolems: [11:44] mzanetti: ↑↑ where is that code? [11:45] tsdgeos: I think somewhere in tests/ [11:45] tsdgeos: depends... we have 2 license checkers [11:45] really? grep sound nothing [11:45] may it be the jenkins one? [11:45] tsdgeos: one in jenkins that runs before the build [11:45] that one [11:45] tsdgeos: and one in the tests from Michi [11:46] tsdgeos: lp:ps-qa-tools [11:47] tsdgeos: where is the typo? [11:47] "Found 1 license prolems:" [11:47] hehe... right [11:47] https://jenkins.qa.ubuntu.com/job/unity-phablet-quantal-armhf-ci/498/console [11:47] maybe it was fixed [11:47] tsdgeos: http://10.97.9.46/hooks/A10checklicenseheaders [11:47] nope [11:47] grep doensn't find it there either [11:48] there = ps-qa-tools [11:48] tsdgeos: then lp:pbuilderjenkins [11:48] sorry [11:50] mzanetti: https://code.launchpad.net/~aacid/pbuilderjenkins/fix_typo/+merge/158572 [11:50] dednick: something weird happened in https://code.launchpad.net/~nick-dedekind/unity/phablet-test-panel/+merge/157927 [11:51] it says you removed and added tests/qmluitests/Panel/qml/CMakeLists.txt :D [11:51] tsdgeos: I guess you don't mind if it takes a week or two until it is actually installed on all jenkins nodes? [11:51] tsdgeos: otherwise you would need to make a release [11:51] mzanetti: lol no [11:51] can't live without that fixed [11:51] mzanetti: hi [11:51] tickling my inner mind [11:51] j/k [11:52] i was just doing something for 3 min before lunch is ready [11:52] paulliu: hey. I just wanted to ask on where you stand with the IndicatorMenuWindow tests to avoid conflicts [11:52] paulliu: but I've found your branch [11:53]