| === jono is now known as Guest43238 | ||
| tvoss | hmmm, taking eglSwapBuffers and eglSwapInterval strict, a turned off screen does not have vsync. With that, waiting for an eglSwapBuffers with an eglSwapInterval > 0 should never return | 06:39 |
|---|---|---|
| duflu_ | tvoss: Yep. And so all compositors and apps sleep nicely when the screen is off. | 06:43 |
| duflu_ | Unless you have the fglrx driver, which was buggy there | 06:43 |
| === duflu_ is now known as duflu | ||
| tvoss | duflu, I was thinking: If an application would decouple from vsync via a call to eglSwapInterval(0) in reaction to the screen being turned off, immediately returning false from eglSwapBuffers would be valid | 07:00 |
| duflu | tvoss: Returning false sounds correct. However I have seen drivers differ and buggy with respect to return values for an "off" screen. | 07:01 |
| === tvoss is now known as tvoss|afk | ||
| === tvoss|afk is now known as tvoss | ||
| === alan_g is now known as alan_g|lunch | ||
| === alan_g|lunch is now known as alan_g | ||
| === dandrader is now known as dandrader|afk | ||
| mterry | kgunn, so I think the biggest risk factor with split greeter silo is the Mir bugs. I'm going to look into the screen on/off one today a bit. I think racarr was looking at my bugs too? Do you know if he got anywhere? | 14:09 |
| kgunn | mterry: not sure...but good to know...do we have logged bugs for those ? | 14:12 |
| kgunn | to enable me to be really annoying :) | 14:12 |
| mterry | kgunn, um. No, probably not. /me is bad at paperwork | 14:12 |
| mterry | kgunn, will file | 14:12 |
| mterry | kgunn, will work on | 14:13 |
| kgunn | mterry: no worries....still steamy its all so new :) | 14:15 |
| kgunn | mterry: i do know anpok_ was poking around on the flicker | 14:16 |
| kgunn | like first few frames being transparent from greeter | 14:16 |
| mterry | kgunn, yeah... That is something I was willing to live with and fix later | 14:16 |
| mterry | Not something users will always see | 14:16 |
| anpok_ | kgunn: current status is in the review | 14:18 |
| anpok_ | approved it as it is not cause by the change.. | 14:18 |
| anpok_ | i tried with a non-qt greeter (basic_server + eglplasma :)) but that is not a proper greeter.. | 14:19 |
| anpok_ | so I couldnt get to that sequence of turning off and on.. | 14:20 |
| anpok_ | I wanted to look at the qt side of things next.. | 14:20 |
| mterry | kgunn, bug 1297876 and bug 1297878 | 14:20 |
| ubot5 | bug 1297876 in mir (Ubuntu) "Screen turns on when a new session/surface appears" [Undecided,New] https://launchpad.net/bugs/1297876 | 14:20 |
| ubot5 | bug 1297878 in mir (Ubuntu) "Keyboard presses "fall through" and hit surface behind them" [Undecided,New] https://launchpad.net/bugs/1297878 | 14:20 |
| anpok_ | oh 1297876 is a mir issue | 14:21 |
| mterry | kgunn, I'll look at the first one because that seems like something I could track down (see where screen turns on and work back). But the second one looks like it requires more knowledge of Mir than I have | 14:21 |
| mterry | kgunn, I think racarr was looking at one or both of them | 14:21 |
| mterry | anpok_, I *think* second one is too | 14:22 |
| kgunn | anpok_: want to dive in to 1297876...and maybe work with racarr when he comes on.... | 14:23 |
| kgunn | let's see if we can kill these quickly | 14:23 |
| * kgunn wants to land split greeter soon! | 14:24 | |
| anpok_ | kgunn: sure - this one is very annoying, thought it was in a different part of the system. | 14:24 |
| anpok_ | brb | 14:24 |
| === dandrader|afk is now known as dandrader | ||
| AlbertA | so now that facebook will buy oculus do we still have to "pass the oculus rift test"? :) | 14:40 |
| anpok_ | there is another company with augmented reality glasses | 14:43 |
| AlbertA | sony? | 14:43 |
| AlbertA | :) | 14:43 |
| anpok_ | or in austria there is a vr startup | 14:44 |
| anpok_ | i think we just rename it .. or have an indirection | 14:44 |
| ogra_ | AlbertA, sure you do, we need to be prepared for when we acquire facebook as sub-devision after making billions and billions in the phone market | 14:44 |
| anpok_ | "coolest displaying tec test" | 14:44 |
| AlbertA | I kinda like passing the "morpheus test" | 14:45 |
| AlbertA | though probably not applicable :) | 14:45 |
| anpok_ | hmm i am sure the sony equipment will work in linux likewise.. | 14:46 |
| AlbertA | unless we get ubuntu running on the ps4 :) | 14:46 |
| ogra_ | lets start small, Ubuntu Phone on the PSP first ;) | 14:56 |
| anpok_ | vita might be simpler .. no mips ... | 15:00 |
| alan_g | AlbertA: look at DemoPrivateProtobuf | 15:11 |
| AlbertA | alan_g: thanks | 15:17 |
| alan_g | yw | 15:17 |
| === alan_g is now known as alan_g|tea | ||
| === alan_g|tea is now known as alan_g | ||
| === chihchun is now known as chihchun_afk | ||
| === chihchun_afk is now known as chihchun | ||
| === dandrader is now known as dandrader|lunch | ||
| === tvoss is now known as tvoss|afk | ||
| alan_g | Does anyone know why ConsumingPlacementStrategy exists in the codebase? | 16:32 |
| kdub | alan_g, I don't know | 16:44 |
| pietro10 | Quick dev-related question: does Mir use libxkbcommon? Thanks. | 16:52 |
| kdub | its one of our dependencies | 16:54 |
| alan_g | $ grep libxkbcommon debian/control | 16:54 |
| alan_g | libxkbcommon-dev, | 16:54 |
| pietro10 | thanks | 17:02 |
| === dandrader|lunch is now known as dandrader | ||
| === alan_g is now known as alan_g|EOD | ||
| kgunn | AlbertA: trying to be better about public chat...so powerd will still contain brightness value ? | 19:02 |
| kgunn | and u-s-c will simply be a proxy for the ui in terms of requests to change that value...? | 19:02 |
| kgunn | i assume wrt actual screen state (on/off) that's still going to end up in u-s-c ? | 19:03 |
| AlbertA | kgunn: yes powerd will still handle the brightness logic | 19:03 |
| AlbertA | kgunn: I was pulling that into usc but from the conversation in the morning | 19:03 |
| AlbertA | kgunn: that was better left in powerd | 19:03 |
| kgunn | ok | 19:03 |
| AlbertA | kgunn: but usc will inform powerd when to act on brightness etc | 19:04 |
| kgunn | AlbertA: makes complete sense, any hw ctl & related hw logic should be in powerd...and usc is more about policy & req/resp on powerd | 19:05 |
| AlbertA | kgunn: yes basically usc will do the input event timeouts and display on/off, and it will inform powerd when to to dim the screen | 19:06 |
| AlbertA | kgunn: the settings will be obtained from the sessions possibly from a sidechannel using the same server socket mir uses | 19:07 |
| AlbertA | kgunn: or maybe we can incorporate it into display configuration - for now I'm leaning towards side channel as far as settings | 19:07 |
| kgunn | ack | 19:19 |
| kgunn | mterry: out of bug 1297876 and bug 1297878 | 19:58 |
| ubot5 | bug 1297876 in mir (Ubuntu) "Screen turns on when a new session/surface appears" [Undecided,New] https://launchpad.net/bugs/1297876 | 19:58 |
| ubot5 | bug 1297878 in mir (Ubuntu) "Keyboard presses "fall through" and hit surface behind them" [Undecided,New] https://launchpad.net/bugs/1297878 | 19:58 |
| kgunn | you really wanted help with 78 right ? | 19:58 |
| kgunn | just wanted to make sure i was asking racarr about the right thing | 19:59 |
| kgunn | or...asking him to look at the right thing | 19:59 |
| mterry | kgunn, both are blockers, but I think I can at least investigate 76 myself. Plus 78 affects more than just split mode, so it's more critical. I think racarr was looking more into 76 so far, but I'm not sure of his progress in either | 19:59 |
| kgunn | agreed...78 looks like the nastier of the 2 | 20:00 |
| kgunn | in terms of ux effect | 20:00 |
| mterry | kgunn, 76 is mad annoying bro | 20:01 |
| mterry | but yes | 20:02 |
| kgunn | mterry: sorry...i meant 76 | 20:02 |
| kgunn | no nvmd... | 20:02 |
| mterry | kgunn, they are both super annoying | 20:02 |
| kgunn | ah..my eyes o.O | 20:02 |
| mterry | kgunn, one drains battery and is startling. One makes keyboard useless | 20:02 |
| kgunn | keyboard useless is worse imho | 20:03 |
| mterry | sure | 20:03 |
| === dandrader_ is now known as dandrader|bbl | ||
| mterry | I'm on the same page :) I just also hate 76 | 20:03 |
| racarr | should I just | 20:21 |
| racarr | flip a coin/ | 20:21 |
| racarr | ok no 78 | 20:21 |
| racarr | mterry: kgunn: Anything known so far on 78 besides that is in the bug? | 20:22 |
| mterry | racarr, no | 20:23 |
| racarr | cool | 20:23 |
| racarr | ill be on it in an hour or so (have to do the usual flash phone clean, build clean stack, etc...) | 20:24 |
| kgunn | bregma: bschaefer ...meant to ask | 20:47 |
| kgunn | oops, wait...moving | 20:47 |
| mterry | racarr, regarding the 76 issue, your original instincts were right. If I uncomment the "power_mode = default_power_state" bits in DefaultDisplayConfigurationPolicy::apply_to, problem goes away. In your unity-mir branch, you say nested mode should do nothing for apply_to. Should NestedDisplay() just not call apply_to? | 22:06 |
| racarr | mterry: Which branch of unity-mir are you using against mir devel? trunk isnt building | 22:10 |
| racarr | mterry: Err...let me think | 22:10 |
| mterry | racarr, whichever ended up in silo 004 | 22:11 |
| mterry | kgunn_, which branches did you use for unity-mir in silo 004? | 22:11 |
| racarr | mterry: Err...so I cant remember exactly what I said in my unity-mir branch | 22:12 |
| racarr | I think it would make sense to say though, for example that | 22:12 |
| racarr | the nested server should inherit the display configuration from the system compositor by default | 22:13 |
| racarr | so the nested server could use a | 22:13 |
| racarr | DisplayConfigurationPolicy (overriding the_display_configuration_policy) | 22:13 |
| racarr | that doesnt initialize things to their default state | 22:13 |
| racarr | re-initialie/whatever | 22:13 |
| kdub | if they're different defaults, why not just provide an appropriate default implementation instead of overriding things | 22:21 |
| racarr | mterry: Ok where does silo 4 come from! | 22:23 |
| racarr | I guess maybe I can use binary packages of unity-mir | 22:23 |
| racarr | but...I think server abi broke | 22:23 |
| racarr | as early as like | 22:23 |
| racarr | every day for the past week | 22:23 |
| racarr | lol | 22:23 |
| racarr | so im not sure how this is possible | 22:23 |
| kgunn_ | mterry: one moment | 22:23 |
| mterry | racarr, https://launchpad.net/~ci-train-ppa-service/+archive/landing-004/+packages | 22:23 |
| kgunn_ | mterry: https://pastebin.canonical.com/107219/ | 22:24 |
| kgunn_ | racarr: ^ if needed | 22:24 |
| racarr | hmm I only see one unity-mir branch | 22:26 |
| racarr | which doesnt have the update | 22:26 |
| racarr | that the_session_authoried now has to implement screencast_is_allowed | 22:26 |
| racarr | maybe I got my local branches mixed up just a sec | 22:26 |
| racarr | mterry: Err, so I am not sure how you built mir/devel | 22:29 |
| racarr | against these packages | 22:29 |
| mterry | racarr, I apt-get sourced the package in the silo | 22:30 |
| racarr | / are there missing um branches... | 22:30 |
| racarr | the mir package? | 22:30 |
| mterry | racarr, and worked off that. That way, no rebuilding | 22:30 |
| mterry | racarr, yeah | 22:30 |
| racarr | oh | 22:30 |
| racarr | thats at least a week behind mir devel it seems | 22:30 |
| racarr | maybe more? screencast has been in for forever | 22:31 |
| mterry | racarr, well, good for fixing bugs anyway | 22:31 |
| racarr | mm | 22:31 |
| racarr | lemme get the PPA then | 22:31 |
| === dandrader|bbl is now known as dandrader | ||
| racarr | except apt-get update wont finish lol thats good | 22:39 |
| racarr | mterry: Just to make sure I understand how I am supposed to reproduce it | 22:42 |
| racarr | it should just show up in this silo | 22:42 |
| racarr | so there is nothing to isolate it to mir/mir devel in particular | 22:42 |
| mterry | racarr, yeah | 22:42 |
| mterry | racarr, uh. 76 needs silo. 78 should be reproducable with stock unity8 and silo | 22:43 |
| mterry | racarr, haven't tested 78 with stock unity8 and mir/devel | 22:43 |
| racarr | right but at this point it could be anything in the silo except unity8 right | 22:43 |
| racarr | i.e. papi/unity-mir | 22:44 |
| mterry | racarr, well papi/unity-mir are just rebuilds agains tmir | 22:44 |
| racarr | er...do you have to do something besides writable-image to make | 22:53 |
| racarr | apt work on the phones now | 22:53 |
| racarr | I am having all sorts of issues | 22:53 |
| mterry | racarr, no... | 22:55 |
| racarr | :/ apt-get update will never finish even though I can access the servers it claims to stall on | 22:55 |
| mterry | racarr, what errors? | 22:55 |
| racarr | and dist-upgrade is complaining about no space left ond evice | 22:56 |
| racarr | even though its a clean flash | 22:56 |
| mterry | racarr, I've gotten the stall before, a reboot can fix that sometimes | 22:56 |
| racarr | I dont even know space on what partition...and I check mount -v and see there are now liek | 22:56 |
| racarr | over 20 partitions lol | 22:56 |
| mterry | racarr, no space is an odd one after clean refresh. two things can help: | 22:56 |
| mterry | apt-get purge 'language-pack.*' mir-test-tools | 22:56 |
| mterry | rm -r /var/cache/apt/archives/ | 22:56 |
| racarr | I cant even run apt-get purge | 22:57 |
| racarr | because of no space | 22:57 |
| racarr | but I literally | 22:57 |
| racarr | emptied /var/cache | 22:57 |
| racarr | hmm | 22:57 |
| racarr | I guess I will reflash | 22:59 |
| racarr | with bootstrap | 22:59 |
| racarr | hopefully my flash isnt dieing | 22:59 |
| mterry | racarr, bootstrap shouldn't be needed, but couldn't hurt | 22:59 |
| racarr | shouldnt be right but | 23:00 |
| racarr | dont know what to do at this point | 23:00 |
| racarr | I mean I literally just flashed a new image without bootstrap immediately prior to this. | 23:00 |
| mterry | racarr, :( | 23:02 |
| mterry | racarr, did you accidentally do dd < /dev/random > /datafile? common mistake | 23:03 |
| racarr | mterry: Maybe, thats my password | 23:05 |
| racarr | so maybe I entered it in the wrong prompt | 23:05 |
| mterry | :) | 23:05 |
| racarr | or maybe my flash is dead | 23:05 |
| racarr | lol ok everything seems happy now. | 23:13 |
| racarr | just going to pretend that never happened | 23:13 |
| racarr | my usb cable is pretty flaky maybe the flashing process went wrong or maybe my flash really is failing | 23:13 |
| racarr | maybe it was just overheating *shrug* | 23:14 |
| mterry | :-/ | 23:16 |
| mterry | racarr, still before you re-install all the mir deps, you should probably run " apt-get purge 'language-pack.*' mir-test-tools" to help avoid in future | 23:17 |
| racarr | mterry: good call | 23:18 |
| racarr | woah | 23:19 |
| racarr | that was quite | 23:19 |
| racarr | the boot up experience | 23:19 |
| mterry | racarr, with the PPA? | 23:20 |
| racarr | yes | 23:20 |
| racarr | I can see also the osk | 23:20 |
| racarr | literally cant be dismissed lol | 23:20 |
| mterry | :) | 23:20 |
| racarr | I also got it stuck rotated somehow | 23:20 |
| mterry | If you re-summon it, it can come back unrotated | 23:21 |
| racarr | hmm ok well have to get unity-mir debug symbols now... | 23:29 |
| racarr | MirSurfaceManager::surfaceAttributeChanged is at least being called when the OSK comes up so cant understnad why | 23:29 |
| racarr | the surface wouldnt be getting the state | 23:29 |
| racarr | but we shall see | 23:29 |
| racarr | interesting | 23:55 |
| racarr | surface manager cant find the unity-mir side | 23:55 |
| racarr | osk surface at the time of | 23:55 |
| racarr | attribute | 23:55 |
| racarr | setting | 23:55 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!