| duflu | RAOF: Do you recall why outputs have pixel formats? | 02:15 |
|---|---|---|
| duflu | Is that for nesting? | 02:15 |
| RAOF | Because CRTCs have pixel formats? | 02:16 |
| RAOF | How else would you scan out of rgb565, or rgbx1010102? | 02:16 |
| duflu | RAOF: Hmm, OK. Just noticed we've hardcoded xrgb in a couple of places for KMS (which might aggravate the vmware bug) | 02:16 |
| RAOF | Really? Odd. | 02:17 |
| RAOF | If you want to update the kms platform to use the available output formats that'd be fine :) | 02:17 |
| duflu | In other news, I found a second use for blank DVDs | 02:19 |
| duflu | To burn ThinkPad BIOS updates (which otherwise require Windows) | 02:19 |
| duflu | The first use is as a coaster | 02:19 |
| RAOF | Huh. I always used USB sticks for that. | 02:20 |
| duflu | RAOF: Not an option with ThinkPad ISOs | 02:20 |
| duflu | I have tried many times | 02:20 |
| duflu | The first 32K of the ISO is zero... so there's no boot sector there | 02:22 |
| RAOF | HUh. | 02:22 |
| === chihchun_afk is now known as chihchun | ||
| duflu | Is CI down? I've been waiting 5 hours on my current branch... | 07:21 |
| duflu | Oh I see others have been waiting 7 hours | 07:23 |
| anpok | hm device-runtests-mir seem to wait for krllin executors | 07:45 |
| anpok | hm but it is still running | 07:46 |
| anpok | at least one job.. | 07:47 |
| duflu | anpok: Finished after 6 hours | 08:20 |
| duflu | Just one revision on one branch | 08:20 |
| === chihchun is now known as chihchun_afk | ||
| alan_g | kdub: have I understood your review comment? https://code.launchpad.net/~alan-griffiths/mir/rework-throwback-and-staging/+merge/314351 | 12:09 |
| kdub | alan_g, didn't see the other file, will remove needs fixing | 12:15 |
| alan_g | thanks | 12:15 |
| * alan_g only needs the prerequisite to pass review now | 12:17 | |
| === hikiko is now known as hikiko|ln | ||
| === hikiko|ln is now known as hikiko | ||
| === andyrock_ is now known as andyrock | ||
| kdub | camako, been thinking... if we're deprecating mir_buffer_stream_get_egl_native_window in favor of Mir(Render)Surface, there's really no reason to have a 'hardware' buffer stream | 15:34 |
| kdub | so maybe we just drop the MirBufferUsage parameter from mir_render_surface_get_buffer_stream(), and that makes software buffers | 15:35 |
| kdub | android drivers don't need streams, mesa doesnt need streams, and the users will port to using MirRenderSurface as the native window (instead of the hop through mir_bs_get_egl_native_window()) | 15:35 |
| camako | kdub, umm yeah sounds reasonable... Future (non-egl) middleware can use PCs, I guess | 15:36 |
| kdub | right, and if we encounter any usage of it (which I think most the usage is the egl case) | 15:37 |
| kdub | we should add extensions for android and mesa to do the right thing | 15:37 |
| kdub | that is, 'add extensions for android/mesa to alloc a hardware bufferstream' | 15:37 |
| kdub | I'll pull that param out of there, and RAOF is probably interested in reviewing as well | 15:38 |
| kdub | 0.26 is going to be a big silo | 15:42 |
| bschaefer | kdub, it is | 15:43 |
| bschaefer | going to be | 15:44 |
| alan_g | kdub: your commit comment seems to be a C&P error - https://code.launchpad.net/~kdub/mir/usage-deprecation-nested/+merge/314518 | 17:01 |
| kdub | alan_g, thanks, correcetd | 17:04 |
| kdub | camako, darn, we have name collisions when trying to change MirRenderSurface to MirSurface :/ | 17:11 |
| camako | kdub, yeah we are not ready... We need to make a release | 17:12 |
| camako | then delete MirSurface | 17:12 |
| camako | then rename | 17:12 |
| kdub | yeah, so it looks like releasing the new platform won't make it into 0.26 | 17:13 |
| alan_g | kdub: unless you call it Mir5urface for the time being and then deprecate | 17:14 |
| alan_g | ;) | 17:14 |
| kdub | yeah :D | 17:14 |
| kdub | mir_5urface_release | 17:15 |
| kdub | not too late to change all the functions to l33t5p34k before the 1.0! | 17:15 |
| alan_g | LOL | 17:15 |
| camako | kdub, @new platform... correct | 17:16 |
| alan_g | If we don't have enough to release now, we /could/ release MirRenderSurface in 0.26, then deprecate it for MirSurface in 1.0. | 17:18 |
| alan_g | ...and delete it in 2.0 | 17:19 |
| bschaefer | kdub, i like that naming scheme, it expresses a lot about the functions | 17:20 |
| bschaefer | %s/*/l33t5p34k/g | 17:20 |
| alan_g | keywords too? | 17:21 |
| bschaefer | %s/./l33t5p34k/g | 17:21 |
| bschaefer | every char | 17:21 |
| * bschaefer assumes this would no longer compile though | 17:22 | |
| kdub | well, thats where macros come in | 17:22 |
| bschaefer | haha | 17:22 |
| bschaefer | good luck figuring out how many k's you need to disambiguate that grammar | 17:23 |
| kdub | alan_g, yeah, I think the plan was to not release the MirRenderSurface, because if its named 'MirSurface', then it falls into the vulkan standard | 17:23 |
| alan_g | kdub: ack, I thought that too | 17:24 |
| kdub | yay, we're in a cheatsheet for khronos https://www.khronos.org/registry/vulkan/specs/1.0/refguide/Vulkan-1.0-web.pdf | 17:25 |
| alan_g | But at that time I also thought 1.0 would make it into 17.04 | 17:25 |
| * bschaefer has hope but it fades by the hour | 17:31 | |
| kdub | r:74 | 18:02 |
| tjaalton | bregma: ping | 22:39 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!