| robert_ancell | duflu, hey, is there anything in particular that needs working on in xmir while you are away? | 02:41 |
|---|---|---|
| duflu | robert_ancell: Yeah I've tried to keep the bug lists accurate. I'll email the URLs to you | 02:42 |
| robert_ancell | duflu, ta | 02:42 |
| bschaefer | duflu, when does your vacation start? | 02:51 |
| duflu | bschaefer: One week | 02:51 |
| bschaefer | duflu, cool, enjoy! | 02:51 |
| tjaalton | duflu: hey, so will mir revert back to the old pkg-config filename for mir-client-platform-mesa-dev, or provide both? | 06:00 |
| duflu | tjaalton: The plan is 0.17.1 provides both. But 0.18.0 only the new one | 06:01 |
| tjaalton | ah, good | 06:01 |
| duflu | If and when 0.17.1 goes out | 06:01 |
| tjaalton | so I can upload this new mesa to xenial then | 06:01 |
| duflu | tjaalton: So long as it doesn't contain -dev and assuming we don't release 0.17.1 in xenial :P | 06:01 |
| tjaalton | right | 06:02 |
| duflu | Simple :) | 06:02 |
| RAOF | Even *if* we release 0.17.1 in xenial; 0.17.1 provides both names. | 06:03 |
| duflu | RAOF: Good point but we don't want xenial packages getting into the habit of the old name | 06:03 |
| RAOF | Indeed. | 06:03 |
| duflu | Which reminds me, Xmir will probably need to support both | 06:04 |
| duflu | That's a bit ugly | 06:04 |
| duflu | Oh, no it won't | 06:04 |
| RAOF | It can always just use the new name. | 06:04 |
| duflu | yes | 06:04 |
| RAOF | Right. | 06:04 |
| duflu | It does already | 06:04 |
| === ara is now known as Guest43853 | ||
| === greyback__ is now known as greyback | ||
| alan_g | greyback: @https://code.launchpad.net/~alan-griffiths/qtmir/test-harness-for-MirWindowManager/+merge/274221 - are there known problems on wily? (The latest failure doesn't look related to the MP) | 10:52 |
| greyback | alan_g: I think CI on wily is busted. As long as vivid+o is ok, I'd be happy | 10:52 |
| alan_g | greyback: I won't look any deeper then. Thanks | 10:52 |
| zzarr | alan_g, I think I'll skip porting mir to freon, it's too easy to install ubuntu on a sd-card | 11:18 |
| zzarr | if I download an image for a device with the same SoC as I have in my device, can I use the binary blob with libhybris? | 11:22 |
| alan_g | zzarr: ok | 11:29 |
| zzarr | am I correct in my assumption that I can use GPU drivers from any Android device with the same GPU type (same SoC, RK3288)? | 11:34 |
| alan_g | zzarr: I'm not familiar with the issues involved, but I suspect you'd at least need compatible firmware drivers. kdub may have a clearer idea (but it is still early on his part of the planet.) | 11:41 |
| kdub | zzarr, in theory it should work | 11:42 |
| zzarr | kdub, okey, what firmware will I need? | 11:43 |
| kdub | whatever firmware works with the android drivers | 11:44 |
| zzarr | kdub, do I need to modify the kernel? (I have an kernel for Arch Linux right now) | 11:44 |
| kdub | android will have its own kernel drivers that have to be around in whichever kernel you're using | 11:45 |
| zzarr | kdub, is there a step by step explanation how to do it? | 11:46 |
| kdub | not really, we use the android kernels in ubuntu touch, i would just go with those | 11:48 |
| zzarr | but it's not a Android based device I have, it's a ChromeBook (ASUS ChromeBook Flip) | 11:49 |
| zzarr | so there's no Android kernel for it (if I can't use a Android kernel for another device) | 11:50 |
| kdub | well, to use the android drivers and hybris, you'll need the same environment that those drivers run in under android | 11:51 |
| zzarr | yes, but I thought that meant I needed just an lxc | 11:53 |
| zzarr | this it what I thought, I could install Ubuntu, have and install the necessary Android drivers in a lxc and install the firmware for the drivers | 11:56 |
| kdub | the userspace drivers have to talk to the kernel drivers properly | 11:58 |
| kdub | and if there are no kernel drivers for the device, they wont work | 11:58 |
| zzarr | ohh, right, I forgot about that | 11:58 |
| zzarr | alan_g and kdub, thanks for your help | 12:07 |
| kdub | yep, no problem | 12:07 |
| === alan_g is now known as alan_g|lunch | ||
| === chihchun_afk is now known as chihchun | ||
| === dandrader is now known as dandrader|afk | ||
| === alan_g|lunch is now known as alan_g | ||
| === dandrader|afk is now known as dandrader | ||
| anpok_ | dandrader: which version of mir did you try to run in kvm? | 15:36 |
| alan_g | \\o/ Got a failure on krillin | 15:52 |
| alan_g | /o\ not the one I was trying to reproduce | 15:52 |
| anpok_ | i think I have to revert the change to the mesa hack | 15:53 |
| alan_g | anpok_: just thinking... could we change mesa to do something like lp:~vanvugt/mir/fix-1510218 and lose the hack completely? | 15:55 |
| bschaefer | alan_g, nice | 15:56 |
| bschaefer | o | 15:56 |
| bschaefer | alan_g, sooo another error that could come up later? | 15:56 |
| anpok_ | the odd thing is .. we do access the native egl display (which does the dlopen for mir symboles) within the mgn::Display .. which already has the guestplatform | 15:56 |
| anpok_ | which means we already reloaded that library | 15:56 |
| anpok_ | because that happens just before the platform creation.. | 15:57 |
| anpok_ | alan_g: hm I dont understand | 15:57 |
| anpok_ | ah change mesa! | 15:58 |
| anpok_ | yeah we could do sane things in mesa instead of what we currently do | 15:58 |
| alan_g | bschaefer: I don't believe in coincidences. This one proves there's racy code about and we all know races. | 16:00 |
| dandrader | anpok_, the one in wily. 0.17 I think | 16:00 |
| anpok_ | dandrader: with the one in wily I do get something on screen .. but for some reason use input does not work | 16:01 |
| bschaefer | alan_g, very true, hopefully caused by the same issue | 16:01 |
| dandrader | anpok_, I downloaded the wily iso from ubuntu.com. installed it in a vm. installed the unity8 session package in it and tryied logging into it | 16:01 |
| anpok_ | which one? | 16:01 |
| anpok_ | i mean which session package | 16:01 |
| dandrader | anpok_, unity8-desktop-session-mir I think | 16:02 |
| dandrader | anpok_, so you can select a unity8 mir session from the lightdm greeter | 16:02 |
| anpok_ | dandrader: hm ok .. this works... then screen turns black for a few seconds.. then I see the u8 greeter.. and run into the input problem | 16:02 |
| dandrader | anpok_, oh, so you get way further than I. for me the greeter UI gets stuck | 16:03 |
| dandrader | anpok_, and running "ps -rf" I see that usc is not even running yet | 16:04 |
| anpok_ | greeter ui gets stuck .. that means usc is not starting up | 16:04 |
| anpok_ | could you paste the unity-system-compositor.log from var/log/lightdm? | 16:04 |
| dandrader | -ef I mean | 16:04 |
| anpok_ | bbl | 16:04 |
| bschaefer | dang Failed to enter Recovery | 16:05 |
| bschaefer | so the version on CI is 311? | 16:06 |
| * bschaefer wonders why 312 is the current image im pulling | 16:06 | |
| === chihchun is now known as chihchun_afk | ||
| alan_g | bschaefer: fginther said CI is 312, where do you see 311? | 16:11 |
| bschaefer | alan_g, o i though someone in google hangout mentioned 311... | 16:11 |
| bschaefer | that could have been their run or a prev run | 16:11 |
| alan_g | bschaefer: that might have been in the context "it started working with 311 and stopped with 312" | 16:12 |
| alan_g | We should record all hangouts for reference | 16:13 |
| bschaefer | ooo i see | 16:13 |
| bschaefer | yeah | 16:13 |
| bschaefer | alan_g, i wish there were logs... | 16:13 |
| bschaefer | just simple text logs | 16:13 |
| bschaefer | but a recording would be nice | 16:13 |
| bschaefer | huh, my krillin does not want to be flashed | 16:14 |
| * bschaefer tries again | 16:14 | |
| bschaefer | alan_g, you just used: ubuntu-device-flash touch --channel=ubuntu-touch/rc-proposed/ubuntu --bootstrap | 16:18 |
| * bschaefer hasnt flashed this krillin in some time | 16:18 | |
| greyback_ | quick C++ question, how can one forward declare the MirFormFactor enum? | 16:19 |
| alan_g | I didn't "--bootstrap", but otherwise yes | 16:20 |
| greyback_ | I'm unable to specify a working storage type | 16:20 |
| alan_g | enums are a PITA because of C legacy. The underlying type needs to be known | 16:21 |
| greyback_ | that's what I'm suspecting. oh well | 16:21 |
| bschaefer | enum class? | 16:23 |
| bschaefer | i suppose that wont work for that though :( | 16:23 |
| camako | bschaefer, http://askubuntu.com/questions/602035/how-do-i-use-ubuntu-device-flash-with-the-bq-aquaris-e4-5-and-aquaris-e5/602037#602037 | 16:30 |
| camako | bschaefer, you need to use a recovery image | 16:30 |
| camako | wget http://people.canonical.com/~jhm/barajas/recovery-krillin.img | 16:30 |
| camako | ubuntu-device-flash touch --channel ubuntu-touch/rc-proposed/bq-aquaris.en --recovery-image /home/camako/Downloads/recovery-krillin.img --developer-mode --password=0000 --bootstrap | 16:31 |
| bschaefer | camako, oo ok, thanks! | 16:31 |
| camako | (channel will be different for you) | 16:31 |
| bschaefer | i thought i'v done this before with that but things could have changed | 16:31 |
| camako | bschaefer, yeah they have | 16:32 |
| bschaefer | camako, cool thanks | 16:32 |
| camako | bschaefer, also you need to put the device in "bootloader" mode as explained in that link by pressing certain buttons | 16:32 |
| === dandrader is now known as dandrader|bbl | ||
| * camako wonders if alan_g 's way of flashing didn't update everything, which might explain the difference | 16:45 | |
| * alan_g wonders too. Will try again tomorrow if there's been no progress. | 16:46 | |
| bschaefer | camako, oo ok, just got down downloading the recovery | 16:48 |
| bschaefer | huh... set developer mode it lights up green, go back it goes back to off | 17:06 |
| * bschaefer reflashes with --developer-mode | 17:14 | |
| bschaefer | kind of strange you cant do that anymore... | 17:14 |
| bschaefer | from the UI | 17:14 |
| alan_g | bschaefer: works for me | 17:23 |
| bschaefer | strange, when ever i press the check, it doesnt ask for a password | 17:23 |
| bschaefer | and when i leave the menu it goes back to being off | 17:24 |
| alan_g | bschaefer: I have a passcode set | 17:24 |
| bschaefer | same | 17:24 |
| bschaefer | i even went and set a new passcode to check | 17:24 |
| alan_g | Odd | 17:25 |
| alan_g | tried rebooting | 17:25 |
| alan_g | ? | 17:25 |
| bschaefer | yeah i just re-flash in dev mode... | 17:25 |
| bschaefer | alan_g, o i should have | 17:25 |
| bschaefer | i had once rebooted, but it was still doing it maybe i needed reset password and reboot | 17:26 |
| bschaefer | well its in dev mode now | 17:26 |
| anpok_ | 311 was yesteday i believe.. | 17:26 |
| bschaefer | soo,ill have to check that again | 17:26 |
| anpok_ | dunno | 17:26 |
| alan_g | what's the command to get the image data? we could add it to the test script | 17:27 |
| bschaefer | what like adb pull? | 17:27 |
| * bschaefer didnt know you could grab that | 17:27 | |
| alan_g | I mean the phablet version of uname -a | 17:28 |
| alan_g | returns the channel, version etc | 17:28 |
| bschaefer | o... /me looks for that | 17:28 |
| bschaefer | alan_g, adb shell system-image-cli -i | 17:30 |
| alan_g | that's the one. At least we can get that info into the test logs. | 17:30 |
| bschaefer | yeah that would be nice | 17:30 |
| * alan_g goes hacking | 17:31 | |
| alan_g | bschaefer: done | 17:36 |
| bschaefer | alan_g, haha that was quick :) | 17:36 |
| alan_g | write access to the repository the script is pulled from is useful | 17:37 |
| bschaefer | yeah that would be nice for debugging | 17:37 |
| === alan_g is now known as alan_g|eod | ||
| === dandrader|bbl is now known as dandrader | ||
| anpok_ | dandrader: maybe the mesa-kms driver package is missing? | 18:42 |
| dandrader | anpok_, hmm... let me check | 18:44 |
| dandrader | anpok_, you mean mir-platform-graphics-mesa-kms6? | 18:48 |
| dandrader | anpok, you there? | 18:50 |
| anpok | yes switched network | 18:53 |
| dandrader | anpok, http://pastebin.ubuntu.com/12992020/ | 18:53 |
| dandrader | anpok, this is what I get when I try to log in | 18:53 |
| dandrader | anpok, just sent you an e-mail with the same content as I though you were gone for good from IRC :) | 18:53 |
| dandrader | thought | 18:53 |
| anpok | sooo you need to turn the machine off and configure spice + qxl | 18:54 |
| anpok | and it should work | 18:54 |
| anpok | should be somewhere in display settings in | 18:55 |
| anpok | virt-manager | 18:55 |
| dandrader | anpok, never heard of this stuff... ok, let me search for it... do you use VirtualBox as well? | 18:56 |
| anpok | oh you are using virtual box | 18:57 |
| anpok | > | 18:57 |
| dandrader | anpok, is VirtualBox bad? should I use something else like VMWare? | 18:59 |
| anpok | the drm driver of virtual box is still a separate kernel module.. I havent come around to update it | 18:59 |
| anpok | hm vmware and kvm would both work | 19:00 |
| dandrader | anpok, is it worth filing a bug about virtual box? | 19:00 |
| anpok | hm sure | 19:00 |
| robert_ancell | kgunn, ack on the email about xmir | 20:39 |
| kgunn | hey thanks robert_ancell | 20:40 |
| kgunn | robert_ancell: i wonder, does it just need a wm ? | 20:40 |
| kgunn | but i figure ChrisTownsend would have suggested that....if thats what it was | 20:41 |
| robert_ancell | kgunn, duflu did create a built in wm, but I suspect it doesn't set a lot of things. | 20:41 |
| robert_ancell | Most modern X apps do behave a bit weird without one. It might also be settings that are set by u-s-d / g-s-d (since u-g runs without a WM but needs u-s-d for things to work). | 20:41 |
| robert_ancell | i.e. no-one really tests toolkits / apps without running inside a full shell so the defaults are often wrong. | 20:42 |
| ChrisTownsend | kgunn: robert_ancell: Compiz is running, so there is a WM under Xmir. | 20:43 |
| ChrisTownsend | And it works fine on a desktop, so I'm thinking it may be specific to arm support. | 20:44 |
| robert_ancell | ok | 20:45 |
| ChrisTownsend | robert_ancell: The awful artifacting has been there for some time now. The huge scaling inside the Xmir window is pretty new- it doesn't occur with the version of Xmir in the libertine PPA. | 20:46 |
| robert_ancell | ChrisTownsend, where is the libertine PPA? | 20:46 |
| robert_ancell | Is that an old version of Xmir, i.e. a regression? | 20:46 |
| ChrisTownsend | robert_ancell: https://launchpad.net/~libertine-team/+archive/ubuntu/devel/+packages | 20:47 |
| ChrisTownsend | robert_ancell: Yes, I believe it's a regression. | 20:47 |
| ChrisTownsend | robert_ancell: You uploaded that Xmir on Oct. 13, so something between then and now. | 20:48 |
| kgunn | robert_ancell: fwiw, i also was able to just now connect external monitor and it survived/displayed as expected (granted the FF is still huge) | 20:51 |
| kgunn | ChrisTownsend: ^ | 20:51 |
| kgunn | have you been able to do multimonitor on arm ? | 20:52 |
| kgunn | thot you had trouble...so i've at least seen it once :) | 20:52 |
| kgunn | w/ xapps specifically | 20:52 |
| ChrisTownsend | kgunn: I was kind of able to get it to work one time, but it went to the, uh, screen with the dots first for a bit before actually displaying. I haven't been able to get it to work since. | 20:54 |
| ChrisTownsend | kgunn: But I think any Xapps stuff is a separate issue from external monitor stuff. | 20:55 |
| kgunn | me too actually | 20:55 |
| kgunn | wasn't sure if you were claiming xmir was directly impacting it, so good to confirm | 20:56 |
| ChrisTownsend | kgunn: Nope, I just think external monitor support is fragile period. | 20:56 |
| === greyback__ is now known as greyback | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!