[00:08] yeah [00:08] I only branched because of the multiarch issue [00:08] I haven't branched the other projects [10:05] tvoss, morning [10:06] dandrader, morning :) [10:06] have you played with Unity's autopilot already? [10:06] dandrader, a little bit, but no serious result yet [10:07] dandrader, tried to prototype a class MultiTouchDevice in X11.py, like lamalex suggested yesterday [10:07] tvoss, I found it a bit fragile. When I run, a good number of tests failed [10:07] dandrader, did you look into autopilot yet? [10:08] dandrader, I think thomi sent around an email in the morning that mentioned failing autopilot tests [10:09] quite hard to make this kind of tests reliable, there are just so many variables. running from within a virtual machine with a pristine ubuntu installation might be "a must" for running it... [10:10] on the other hand it's quite neat to see the whole UI moving by itself, showing and hiding launcher, dash, etc :) [10:10] hmm... good to know [10:18] dandrader, yeah, it's a pretty cool technology [10:19] dandrader, are you planning on looking further into autopilot? I have pretty much stuff on my plate regarding chromium patch and jenkins maintenance [10:22] I have to provide some testing for the utouch stuff there one way or another, since I have two bug fixes that have been blocked because they are missing tests [10:23] but first I will invest some time in figuring out the feasibility of making unit tests [10:24] I fear the headaches that might come from autopilot (many dependencies) [10:38] dandrader, ack ... unity already uses google test [10:38] yes === MacSlow is now known as MacSlow|lunch === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [12:19] Someone want to test this: lp:~jpakkane/arsenal/utouch-bug-script [12:19] Run the file scripts/get_touch_bugs.py and redirect stdout to an HTML file. [12:34] Are three and four finger actions broken for anyone else? [12:44] Satoris, they were working last time I tried [12:44] that was last Friday. But I'm using lp:trunk versions of utouch-* and not stuff from repositories [12:46] I use the daily build PPA, which should be the same. [12:48] Three and four finger touches come to applications, but two finger touches disappear somewhere. [12:48] Even though I have disabled all scrolling thingies. [12:51] Satoris, one thing you could do is removing synaptics xorg driver and see if you get the two-touches events [12:51] that would confirm the culprit at least [12:52] But doesn't that driver read the evemu interface? Without it there should be no mouse/touch events at all. [13:01] I believe both synaptics and evemu use the /dev/input interface [13:02] I mean evdev, not evemu [13:02] There is a plain evdev driver for X now? [13:02] although evemu also uses the /dev/input interface [13:03] there has been for a few releases now [13:03] it's what gets used for touchscreens [13:03] it can also be used for touchpads, but we don;t normally recommend it [13:03] because it won;t work the way WIn 95 and Mac OS users expect [13:04] Ok, I see it. Is there a way to blacklist the Synaptics driver or does one have to uninstall the package? === MacSlow|lunch is now known as MacSlow [13:04] you can blacklist the synaptivs driver usiing fancy config scripts in, um, some /var/lib directory somewhere.... [13:05] /usr/share/X11/xorg.conf.d [13:05] see, /var/lib [13:08] It does not seem to have anything X or evdev related ... [13:27] Changing the input driver from synaptics to evdev causes all touches disappear. [13:28] Changing it back made touches work. [13:29] Gestures, too. [13:31] This does not make ANY SENSE. [13:34] dark magic? [13:34] blood magic? [13:35] I think this calls for magic written with a 'k'. [13:35] restarting X to pick up new shared libraries? [13:35] By which I mean magick. [13:36] In theory, yes. But the old X used the old libraries, which were working fine yesterday. === dandrader is now known as dandrader|lunch [13:54] * tvoss will miss the standup today, ps-qa meeting [14:06] tvoss, when you have a chance, just pop in to let us know how things are going and what you're up to [14:06] Satoris, I don't have any issues with trackpad 3 and 4 touch gestures [14:07] four touch gestures should always work, no matter what settings you have [14:11] The properties reported by xinput were strangely out of whack when the gestures were broken. Fiddling with them did not help. They are ok now, though. [14:11] interesting [14:11] Satoris, are you running the standard desktop with gnome-control-center, gnome-settings-daemon? [14:12] As it comes out of the box, yes. [14:13] Shared library wackiness is the most probable explanation, methinks. [14:13] huh [14:13] Satoris, I just tried your script [14:13] it looks pretty good! [14:14] I think all the data is there [14:14] it could be cleaned up visually, table borders and headers, etc. [14:14] but it'll be a great help [14:14] The API differences between launchpadlib and the helper lib thingie are minimal. The only difference seems to be how you get the launchpad object. [14:14] it's already found lots of bugs I didn't know existed :) [14:15] ok [14:15] Which made me suspect that one of them was deprecated. [14:15] * bregma stands up, scratches his arm, and yawns [14:15] heh [14:15] I'm going to be proposing a fix for bug 972012: Touchscreen touches are not rejected until they end [14:16] then probably weeding bugs from the new bug list that Satoris has created [14:17] Made the script. Then I fought the X, and the X won. [14:17] I'm seeing q few memory leaks when running the geis testsuite, I'm hoping they're poorly written test cases, but I want these to run clean, so..... [14:17] I shot the sherriff [14:17] dandrader|lunch, standups :) [14:18] I'm trying to do a dist-upgrade in the background and it's giving me a time-to-copletion estimate of between 4 hours and 9 days, depending on when I look [14:18] ugh [14:18] right now it's down to 1 hour 15 minutes [14:18] I strongly suggest having someone else do the CSSification of the bug list. I haven't done it in ages, so the end result probably won't be all that spectacular. [14:19] Satoris, hopefully we can reuse an existing template? [14:19] and change it slightly as needed? [14:20] The "templates" in Arsenal are mostly test thingies rather than production. Most scripts just do their own thing, some with Gnuplot, even. [14:20] hmm [14:20] Plus if there is not a template for the exact layout we have, it has to be recreated anyway. [14:26] And the test templates are very problem specific, such as bugs-by-team.mako. [14:27] They seem to be doing some sort of a JSON data transfer thingie of which I'm not entirely sure what it does. [14:28] ok [14:30] Satoris, I would love it if we could get a nice list like: http://reports.qa.ubuntu.com/reports/kernel-bugs/reports/_kernel_hot_.html [14:30] Since (not all of) their scripts are not in Python, this is probably just a way to get data from one place to another rather than just using Genshi directly. [14:32] Well, I _can_ do it, but someone with Web experience probably would get it done faster. Just sayin'. [14:32] dandrader|lunch, bregma: I just pushed a couple MPs for grail [14:33] biab [14:33] cnd: leaving now, Skype later? [14:33] Satoris, sure [14:33] Ok. Bye. [14:36] 22h 59min 27s [14:37] 5d 14h 6min 19s [14:38] I think I'm overloading a local network node or something [14:39] hey folks [14:40] 8d 9h 10min 7s [14:40] working on chromium patch set, helping out with various jenkins tasks [14:42] dived into autopilot today a little mor [14:42] e [14:42] but no definite result, yet === dandrader|lunch is now known as dandrader [14:46] damn, missed the stand up. [14:46] I'm working on a unit test for GestureEngine in unity [14:47] for bugs https://code.launchpad.net/~dandrader/unity/lp940612 and https://code.launchpad.net/~dandrader/unity/lp969554 [16:00] cnd, hangout? [16:00] sure [16:00] one sec [16:00] cnd, ack :) [16:02] tvoss, did you get an invite? [16:02] cnd, no [16:02] hmm [16:02] cnd, let me try [16:02] k [17:55] cnd, pint [17:55] s/t/g [17:55] pong [17:56] regarding the virtual d'tor: I think every test reyling on a a custom fixture inherits the fixture [17:56] iirc [17:57] oh right [17:57] didn't think of that [17:57] should we put it in the comment on the mp for reference? [17:57] I forgot about the distinction between the test class and the fixture class [17:57] I'll do it [17:57] thanks [17:58] np, happy to approve with that changes [18:05] tvoss, so I did some more reading [18:05] and the virtual destructor doesn't need to be specified if you are deriving from a base class that already has one [18:05] which testing::Test does [18:05] ah, like a transitive virtual d'tor? [18:06] I guess [18:06] http://www.parashift.com/c++-faq-lite/virtual-functions.html#faq-20.7 [18:06] it might also explain why it wasn't necessary in any of the other tests [18:07] yeah, wasn't aware that it is transitive, thanks for the hint :) [18:11] tvoss, I pushed the subscription deletion change to the test branch [18:12] and commented on the fix branch [18:12] makes sense. Do I just need to approve the test branch then? [18:13] ah, got it ... already approved [18:13] yay [18:13] cool [18:21] woo-hoo, I'm back [18:22] seems my ISP was having some network troubles, or something [18:34] bregma, are you all updated too? [18:35] as of a few hours ago, yes [18:35] I have about 16 kernels I need to purge [18:35] dandrader, bregma: still looking for a review of: https://code.launchpad.net/~chasedouglas/utouch-grail/fix-hold-reject/+merge/100621 [18:35] heh [18:35] I'm going to cut a new release of grail after it's merged