| racarr | RAOF: How goes stub clientAPI stuff? | 00:36 |
|---|---|---|
| === no_m7u is now known as Nothing_Much | ||
| === chihchun is now known as chihchun_afk | ||
| robert_ancell | duflu, is mir_demo_server_shell and mir_demo_client_fingerpaint supposed to work in utopic? When I run them you don't seem to be able to draw with the mouse | 03:34 |
| duflu | robert_ancell: Yes, it does work. Did you remember sudo (required for input device access)? | 03:35 |
| robert_ancell | duflu, the server is run through sudo, yes | 03:35 |
| robert_ancell | I'm getting motion events. And if I click on the fingerpaint window it drags it rather than drawing | 03:35 |
| robert_ancell | duflu, are you running utopic? Can you confirm it working there? | 03:39 |
| duflu | otp | 03:39 |
| robert_ancell | ah, it seems that on first click it's in drag mode. If I hold down alt then drag the window it seems to work after releasing alt and then clicking. | 03:42 |
| duflu | robert_ancell: Ah, could be a minor demo-shell issue stealing the events then | 03:48 |
| duflu | Or not an issue at all | 03:49 |
| duflu | Will test | 03:49 |
| robert_ancell | It seems like the surface doesn't have focus. When I test GTK+ mir we're not getting any mouse up/down events even when clicking inside the surface until that first drag with alt. | 03:49 |
| duflu | robert_ancell: Works fine immediately on startup for me (now using vivid but always worked in utopic too) | 03:56 |
| robert_ancell | weird | 03:56 |
| duflu | robert_ancell: I fully expect we have input bugs though. I can't reproduce this one is all | 03:56 |
| duflu | robert_ancell: Oh, do you have multiple surfaces/clients when fingerpainting? | 04:00 |
| duflu | I think I have observed that fingerpainting on an unfocused window only works sometimes | 04:01 |
| duflu | Although I'm not sure it should work at all in our current design | 04:01 |
| robert_ancell | Who knows the correct way to interpret a current Mir motion event? GTK+ Mir was looking at the button_state field to determine which button event to generate. But touch events seem to have button_state empty. You can't tell it's a touch event without looking at the pointer_coordinates field. If we get a down event with button_state empty just assume it's the primary mouse button? | 04:02 |
| duflu | robert_ancell: There's a "tool" field to tell you if it's a finger or mouse (with buttons) | 04:03 |
| duflu | In theory | 04:04 |
| robert_ancell | duflu, only in the pointer_coordinates list - which could be a different tool for each event | 04:04 |
| robert_ancell | for each coordinate I mean | 04:04 |
| robert_ancell | We have the magic numbers for the device but no way to interpret them :) | 04:04 |
| duflu | robert_ancell: Yeah you could be using multiple tools at once | 04:04 |
| robert_ancell | I'm just going to generate a primary mouse button event if the field is empty | 04:05 |
| duflu | robert_ancell: Technically you should look at all the pointer_coordinates because any one of them could be a mouse | 04:05 |
| robert_ancell | But which one generated the up event? | 04:05 |
| duflu | robert_ancell: Hmm sounds like a good question. I'm reaching the limits of my mental memory of the event struct | 04:06 |
| robert_ancell | Based on previous discussions I think the answer would be "we're going to replace that struct" | 04:06 |
| RAOF | robert_ancell: This discussion is going to be obsolete reasonably soon. | 04:06 |
| RAOF | Hah, yes! | 04:06 |
| robert_ancell | there we go :) | 04:06 |
| duflu | robert_ancell: Kind of. It will hang around for a fair time because breaking the client API will be painful | 04:07 |
| duflu | ... for distro | 04:07 |
| RAOF | Do what you want for the moment, we'll provide you with an actual pointer in MirInputEvent 2.0 | 04:07 |
| RAOF | (As in - there'll be a core-pointer like device that is what you care about) | 04:07 |
| duflu | robert_ancell: A reasonable kludge is only test pointer[0] | 04:07 |
| robert_ancell | I'm just going to do if (button_state == 0) generate_event (PRIMARY) | 04:08 |
| duflu | robert_ancell: Actually it's probably a good idea to separate button state from the pointer list. Assuming you don't care about multi-mouse support, you might want to use buttons in concert with some other device. So the device that has the buttons is irrelevant | 04:11 |
| robert_ancell | What is with X generating out of order keyboard events when it gets laggy? Very annoying... | 04:12 |
| * RAOF is confused and alarmed as to what the hell is happening on https://jenkins.qa.ubuntu.com/job/mir-utopic-amd64-ci/342/consoleFull | 06:59 | |
| duflu | RAOF: The leak? | 07:09 |
| RAOF | That, and the 25 failing tests are failing with valgrind errors in totally unrelated code. | 07:10 |
| RAOF | If you look at it, ClientLibraryErrors.create_surface_returns_error_object_on_failure completes fine. | 07:10 |
| duflu | RAOF: Yeah I think that's "normal" with valgrind errors... some tests will fail delayed | 07:10 |
| RAOF | And then 25 *unrelated* tests fail with by 0x5EA268: ClientLibraryErrors_create_surface_returns_error_object_on_failure_Test::TestBody() (test_client_library_errors.cpp:206) | 07:10 |
| RAOF | Oh, it's probably something to do with forking. | 07:16 |
| duflu | That's fun. Compiz is mentioned in a kernel header... #define KEY_SCALE 120 /* AL Compiz Scale (Expose) */ | 07:26 |
| === Saviq is now known as Saviq-codedive | ||
| === chihchun_afk is now known as chihchun | ||
| === dandrader is now known as dandrader|afk | ||
| === dandrader|afk is now known as dandrader | ||
| === chihchun is now known as chihchun_afk | ||
| === chihchun_afk is now known as chihchun | ||
| alan_g | alf_: anpok_ any opinions pending on this? https://code.launchpad.net/~alan-griffiths/mir/mir_Server-apply_settings/+merge/240612 | 15:00 |
| alf_ | alan_g: commented | 15:03 |
| alf_ | alan_g: (approved) | 15:03 |
| alan_g | alf_: thanks. (I'll now base my next MP on it.) | 15:05 |
| === chihchun_afk is now known as chihchun | ||
| === dandrader is now known as dandrader|lunch | ||
| === chihchun is now known as chihchun_afk | ||
| === dandrader|lunch is now known as dandrader | ||
| === alan_g is now known as alan_g|EOD | ||
| racarr | AWS outage causedirssi to die | 18:55 |
| racarr | private message bankruptcy :) | 18:56 |
| === Saviq-codedive is now known as Saviq | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!