00:06 cmarcelo: gitlab ci question: do i need to setup "runners" to use gitlab ci in mesa/vkrunner repo? Idea is to make it run the basic unit tests already in the repo... nothing like the mesa ci tests etc.
00:07 cmarcelo: also: do I need to be the owner to set this up? I can't find the Settings option described in https://docs.gitlab.com/ci/quick_start/#ensure-you-have-runners-available
03:33 anholt: cmarcelo: shouldn't need to do anything to use the shared runners, to my knowledge.
03:39 cmarcelo: anholt: tks. will try to follow the remaining instructions to create the ci files in the repo.
03:49 cmarcelo: I should be using https://freedesktop.pages.freedesktop.org/ci-templates/ right?
04:49 anholt: cmarcelo: yeah. might find a smaller project than mesa to crib from as far as usage. ;)
08:41 MTCoster: Hi, I understand I can request a merge from upstream into drm-misc-next. We have a couple fixes waiting to make their way back that are blocking landing our series to support BXS-4-64. Is this the right way to go about that?
11:14 dwt: melissawen, emersion: I was wondering the other day if we should revive the "info on why DRM commits fail" topic for this hackfest. But I still can't decide whether it's at all possible to express things like this sensibly.
11:25 emersion: yeah, I have the same feeling...
11:26 emersion: also it seems like we spend a lot of time talking and that rarely results in anything tangible
11:26 emersion: we lack volunteers to do the work :(
13:41 dj-death: alyssa: interesting you're passing maintenance8 caching tests
13:42 dj-death: alyssa: unless there is a hidden requirement that the serialized shaders are 8bytes aligned, it looks like the runtime is not capable of maintenance8
13:42 dj-death: alyssa: there is also some weird ordering issues in there
13:43 alyssa: dj-death: CTS fixed the ordering issue last week
13:43 dj-death: ah thanks :)
13:43 dj-death: so it's just 8byte alignments
13:43 dj-death: okay, enough for today
13:43 dj-death: there is still RT shaders to debug tomorrow...
13:45 alyssa: dj-death: see https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/5689 and https://gitlab.khronos.org/vulkan/vulkan/-/issues/4236 for backdrop
13:46 alyssa: https://github.com/KhronosGroup/VK-GL-CTS/commit/d21f6df68644f6fac8a97fab9f7e011d142d8af4 is the fix
14:04 kusma: eric_engestrom  daniels: seems CI is broken when building the docs: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34385
14:04 kusma: ...as far as I can tell, no docs-changes were merged into main since the infrastructure update, so... I guess this could be related?
14:06 kusma: Oh wait, this seems kinda relevant: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34165
14:22 jani: kusma: you have typedef for simple_mtx_t twice for !UTIL_FUTEX_SUPPORTED
14:22 jani: with different typedefs
14:23 jani: and presumably the docs build is when you have !UTIL_FUTEX_SUPPORTED. though you could hit that in normal builds too
14:23 kusma: It's not like this has changed lately, though...
14:24 jani: perhaps the docs build never included that header before, and now it does?
14:25 jani: ...and that could be through transitive includes, and hard to track
14:29 jani: __OPENCL_VERSION__ could also play into it
14:43 alyssa: jani: uh-oh
14:43 alyssa: wasn't me
14:43 alyssa: :P
14:46 alyssa: jani: kusma: yeah, that's a regression from cb31b5a9588 ("clc,libcl: Clean up CL includes")
14:46 alyssa: sorry about that :(
14:47 kusma: ok, thanks for checking... I hope we can get this fixed ASAP? :)
14:48 alyssa: kusma: sure. do you want me to write the patch?
14:48 sima: vsyrjala, maybe team up with mripard for your compile test series so it's also included in gitlab ci?
14:48 kusma: Well, I don't know what the patch would look like :P
14:49 kusma: alyssa: That would be awesome
14:50 alyssa: fixing
14:52 alyssa: kusma: MR up
14:53 alyssa: please cherry-pick and land with whatever you have, no need to bother Marge with tihs
14:53 alyssa: this
15:06 melissawen: dwt, emersion, yes, I think this is always a valid discussion, at least to keep it always in our minds.
15:06 melissawen: During the kms meeting in 2024 xdc we retook this discussion about "Better Commit Failure Feedback". A short summary here: https://melissawen.github.io/blog/2024/11/19/summary-display-kms-meeting-xdc2024
15:07 melissawen: but I think we can be more assertive if we select a set of cases to address in each meet-up
15:08 melissawen: because the issue seems to me very widespread
15:11 melissawen: it also reminds a bunch of issues in the AMD issue tracker after Kwin logging: kwin_wayland_drm: Pageflip timed out! This is a kernel bug
15:12 zamundaaa[m]: I put pageflip timeouts on the topic list already
15:13 melissawen: and it seems hard to identify what is the problem even for kernel devs due to missing logs
15:13 melissawen: zamundaaa[m], nice!
18:16 boywasIinforajob: i remember all those big guns sima was daniel vetter, as to what anholt became and how those names changed after your abuse, and HdkR is another one of such heros, i know all jacks friends names and i saw them, we fucking get to the ends of to destroy them completely i know all who those scrubs are, i have been peaceful long enough i liked jefferey sachs's speech he understands very
18:16 boywasIinforajob: well, where those proteins came from, how they extracted substances from my circulation, it's all known, i did not miss the protein folding part, HdkR was long since decided to be left intact cause i know who the main violator is behind the name of David Airlie Adam Jackson etc, and with that abuse slut and others going on everyday how big wanker i was has nothing to do with her
18:16 boywasIinforajob: lesbian or bisexual state, they demoed their interactions calling me names and I was facing their sexual harassment and terror every day, those people are done in the pipes of shitloads soon. all the nigerians all the south africans all the english everyone one after another who committed that abuse along with estonians and finnish and swedish, and HdkR your code in for a job boy is
18:16 boywasIinforajob: incredibly wrong and sick persons deluded shit. But ponytail you are a sick man with wrong visions already and not cause of your occulars, they are probably only things that work borrowed from science, convex or concave i do not know.
19:27 rburton: So... are there any plans for a mesa-demos release as the last one was March 2023, or does it Not Do Releases and we should just package git head?
20:05 mattst88: I think kusma did the last one^
20:06 kusma: It's probably time for another release, yeah.