[03:18] RAOF: mirout is only printing half my EDID (compared to xrandr --verbose). I wonder if it's useful to have a length value without having to interpret bytes? [03:19] duflu: I did think about that. The only time it's interesting is to a client that is unable to parse the EDID but, for some reason, wants to store it? [03:19] RAOF: Yes indeed. Or a user who wants to copy and paste into a decoder [03:19] If they want to do that they can just read /sys/class/drm/$CONNECTOR/edid [03:20] In a much more friendly fashion. [03:21] parse-edid /sys/class/drm/$CONNECTOR/edid is going to be much easier to do than taking the output of mirout, stripping off the whitespace, converting back to binary, and dumping it into parse-edid. [03:21] RAOF: Is there a whole length field or is it incremental? [03:21] The second-last byte is the number of extensions which follow. I'm not sure if all extensions are fixed-size, though. [03:22] So there is indeed no single length field. [03:22] And it's anywhere up to 32K [03:22] Strange also DRM can't get subpixel order from my Dell [04:08] duflu: If you care, lp:~raof/mir/dump-edid-size [04:09] Maybe. Working on that topic right now [04:11] Then maybe you'd like to review https://code.launchpad.net/~raof/mir/dump-edid-size/+merge/312223 :) [04:13] RAOF: I'll let yours land. I have a branch that adds text (but will improve it after lunch)... [04:13] Output 48: DisplayPort, connected, "DELL U2413", 1920x1200+0+0, enabled, on, 520mm x 320mm (24.0"), normal, 1.00x, unknown, monitor [04:13] Ah, parsing out the name? Sure. [04:31] This is the point at which I wish we had a comprehensive set of failure tests. === marcusto_ is now known as marcustomlinson === chihchun_afk is now known as chihchun [05:15] Also a proper RPC debugger. [05:22] Oh, right. [05:33] This fails because we don't have any eventloop in place. [05:52] And now, when you I both actually pump the eventloop *and* don't use freed data, things work better. [06:40] Oh, yay! I *wrote* a bunch of failure tests in the past. [06:40] Thanks, past, me! [11:19] * alan_g has miral-shell running on Unity8/X+O [11:35] oh cool, mir0.25 works fine [11:39] for what? [11:43] for recording the screen, with 0.24.1 on 17.04 (mirscreencas | ffmpeg) although it did record, the video was showing only the first frame [11:44] hm.. and unity8 works, does it already use miral? [11:46] oh and i'm using hexchat snap on unity8 /17.04 now [11:47] iirc there is a test sio right now that integrates the qtmir/miral integration for unity8 [11:48] that mp has it.. https://code.launchpad.net/~unity-team/qtmir/miral-qt-integration/+merge/310001 [11:49] http://i.imgur.com/mkEATx8.png [11:51] i have #2180 installed https://bileto.ubuntu.com/#/ticket/2180 [11:52] interesting, unity8 is becoming usable :D so i took a screen shot, launched firefox dev from the container, uploaded to imgur, copy the link, alt tab to hexchat snap, and paste the link [11:53] hm.. now to get the child windows working, this will make me cry [11:54] greyback: can you advise? ^^ [11:57] romex_: unity8 not quite yet using miral, we have a silo2160 with that implemented but child window support isn't in there yet [11:57] https://code.launchpad.net/~dandrader/unity8/miral [11:57] this, so.. i'll have to compile u8? [11:57] do i need anything else? [11:58] romex_: that is pre-built in silo 2160 [11:58] oh this https://code.launchpad.net/~dandrader/unity8/miral [11:58] https://bileto.ubuntu.com/#/ticket/2160 [11:58] sorry wrong c/p [11:58] Unity8 CI Bot: Needs Fixing (continuous-integration) 16 hours ago === alan_g is now known as alan_g|lunch [11:59] thanks greyback [11:59] to be expected, as branch depends on a unity-api change [11:59] romex_: don't expect child windows to work with that! [11:59] it's a very work-in-progress silo still [12:00] but but daniel had it working, bu then again he knows what he's doing [12:00] ok then, i'll just wait to land it in a silo :D [12:02] for me unity8 is almost ready :D [12:02] i have child windows i think i'll try to move away from unity7 [12:02] and try to report more u8 bugs [12:10] romex_: yep, daniel had it working, but it needs lots of polish yet before we it can be shared === alan_g|lunch is now known as alan_g [13:08] greyback: after QA tried testing miral under Unity8 I started wondering about how that might work. The main blocker I see is that U8 will interpret a lot of user interactions and not pass them to the "application". Is that likely to changes, or is this just an unsupported scenario? [13:09] alan_g: can you give me some examples of this blocker? [13:09] Dragging [13:10] and when did QA test miral&unity8? [13:10] yesterday [13:12] Seems that if the test plan doesn't say "try this on a non-virtual machine running Unity7" they get inventive [13:12] ah, now I follow you, running miral shell in a window inside unity8 [13:12] :D [13:13] dragging should just work, all the mouse events should be delivered to the client [13:13] but apparently not [13:13] I'll give it a look [13:13] You'll need a MP I just put up [13:14] https://code.launchpad.net/~alan-griffiths/miral/basic-Unity8-compatibility/+merge/312246 [13:15] ok [13:15] I guess I need a polite way to say "if Mir doesn't work on a configuration don't bother testing MirAL on it" [13:16] /sigh === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === chihchun is now known as chihchun_afk [15:02] alan_g: for some reason, the miral makes itself fullscreen when I try it on unity8 here. Any idea why? [15:03] greyback: that's Mir's "nested server" - it assumes that it takes over the outputs. Bug 1646449 [15:03] bug 1646449 in Mir "Nested Mir does not support a windowed mode" [Medium,Triaged] https://launchpad.net/bugs/1646449 [15:05] In the distant (and recent) past that made sense - the only nesting was inside USC === dandrader is now known as dandrader|afk [16:10] alan_g: sorry for delay, too many distractions. I've booted to unity8, run "miral-desktop -socket ${XDG_RUNTIME_DIR}/not_mir_socket -launcher qterminal" and I get fullscreen miral with qterminal working inside it. I can drag the window around, and drag the scroll bar inside it [16:11] greyback: OK, cool. Is that the release version of the stack or development? (I was using X+O) [16:11] alan_g: this is zenial [16:11] updated just before trying [16:11] zesty? [16:12] bleh, yes [16:13] darn linguistic redundancy [16:18] greyback: I've got a machine on zesty, but can't log into the U8 session on it. [16:19] Mostly because the greeter doesn't display and I'm typing blind [16:19] alan_g: maybe remove the unity8 greeter and install the old "unity-greeter" ? [16:22] it is the unity-greeter [16:22] Maybe I should try the unity8 one [16:34] greyback: while I'm distracting you. I can't reproduce locally, but any thoughts on this? https://bugs.launchpad.net/miral/+bug/1646532 [16:34] Ubuntu bug 1646532 in MirAL "Cannot run kate or qterminal in miral-desktop" [Undecided,New] [16:38] alan_g: worked for me on today's Zesty [16:38] trying my xenial box [16:39] yep, ok there too. [16:39] thanks, I'm wondering if there's a "missing" dependency I have installed [16:39] alan_g: qtubuntu-desktop installed? [16:40] I'd expect some error output from qt at least [16:40] according to the test case [16:41] am unsure, it all looks legit [16:41] he inside a VM? [16:42] greyback: I was checking you didn't recognise it. Don't waste time on it [16:42] Yes, in a VM [16:42] ok [16:42] if VM, qt using gl to draw stuff, might be gl issue [16:46] Any quick way to vaiidate that guess? (Like one of the Mir demos?) === dandrader|afk is now known as dandrader