orbea has quit [Quit: You defeated orbea! 2383232 XP gained!]
orbea has joined #zink
<ajax>
zmike: how ready do we think zink is? time to take Draft: off and see if anyone else wants to take a look?
<ajax>
s/zink/kopper/
<ajax>
(i think yes, if it's not clear)
<zmike>
I think it's super ready
<zmike>
rip the bandaid off
<zmike>
kusma: were you saying d3d12_lower_int_cubmap_to_array handles the nonseamless case?
<kusma>
Lowering cubemaps to arrays results in non-seamless, yes
<zmike>
I suppose that's obvious
<zmike>
hm
<kusma>
atemu12: depends on what you mean with works. There's no winsys integration, someone would have to write that. Other than that, none of the regular zink devs have macs, but patches are welcome!
<kusma>
atemu12: No, that would depend on the missing winsys bits...
<ajax>
enh. i have several macs, i'm typing from one in fact but it's running a better OS than it shipped with.
<kusma>
ajax: congratulations, you're now the maintainer of the mac port of zink!
<kusma>
atemu12: But if all you want in glx, then I would suspect that it's not too much work to wire up!
<zmike>
no, don't make him run away, we need ajax
<ajax>
i'm not super interested in working on macos support though (i assume that's what's being discussed anyway, i'm only seeing kusma's text)
<zmike>
(I'm also only seeing half of this conversation)
<zmike>
matrix is truly a cancer
<kusma>
ajax: Yeah, I should probably have worded that differently.
<ajax>
no offense taken
<kusma>
ugh, was I not authenticated again?
<zmike>
no, you're fine
<kusma>
Ah, I guess atemu12 isn't?
<ajax>
hm. making kopper work on macos/win32 would be... somewhat challenging.
<ajax>
kopper+glx anyway. under egl it's boring.
<ajax>
but i think i'd need to like port xquartz to glamor first
<ajax>
which... you know, also not the worst idea, if caring about macos is a thing one were into
<ajax>
personally my concern for macos is more along the lines of how do i get darling working
<kusma>
Actually, there's several macs in the household here. But I'm not really allowed to touch them, because they're corporate equipment owned by a bank.
<kusma>
ajax: I think GL -> Metal should be its own thing, TBH.
<kusma>
But I don't feel strongly enough about it to do anything about that.
<kusma>
(yet, I might have some bookmarked pages with details on the MSL binary format etc)
<ajax>
kusma: not a moltenvk fan?
<kusma>
ajax: MoltenVK is great. But layering on top of layering tends to get a bit... semantically lossy.
<kusma>
It's great for running Vulkan content.
<kusma>
That being said, Zink + MoltenVK is by far the shortest path to open source GL on mac.
<kusma>
atemu12: If you have a link, feel free to share! :)
<kusma>
Ah, yes. That one... I looked at it before, and it seems to kinda be taking an approach that's going to take a few decades to get right ;)
<zmike>
kusma: happy news, I've managed to lightly mangle a copy of the d3d12 cube pass into working on zink with (hopefully) minimal overhead
<zmike>
doing a cts run now but it fixes the test that needs to be fixed and also passes piglit
<kusma>
zmike: awesome :)
<zmike>
in retrospect, had I known such a pass existed I might have used that as a starting point
<zmike>
but mesa is too big
<kusma>
Yeah, totally. I've duplicated stuff out there without knowing it too
<kusma>
Especially nir-passes ;)
<zmike>
if you don't reinvent the wheel at least once a week are you really even writing code
<kusma>
I prefer deinventing the wheel by deleting the code for it.
<zmike>
uh oh
<zmike>
have you paid the appropriate licensing fees to jekstrand to be saying such a thing?
<kusma>
I have a long and proven track-record of deleting code
<zmike>
he has a patent for his Delete The Code methodology
<kusma>
I'll get his patent invalidated in court due to prior art
<kusma>
Or just destroy his stats by doing 'git commit --author="TOTALLY JASON" -- /dev/random'
<zmike>
I think as long as you target those into src/mesa for shader image handling he'll be okay with it
atemu12[m] has left #zink [#zink]
<ajax>
working on some of the review feedback, will update kopper branch in a few
LexSfX has quit []
LexSfX has joined #zink
LexSfX has quit []
LexSfX has joined #zink
<jekstrand>
zmike: I'm happy to grant licenses of the Delete the Code method to anyone as long as they're deleting the right code. :D
<zmike>
jekstrand: great news for all the enthusiasts out there
<zmike>
kusma: I'm guessing the tg4 part of the d3d12 cube pass has never been tested on KHR-GL40.texture_gather*