06:06 <chewitt> alyssa reading the backlog .. few people actually ship LE on hardware (some x86/64 and Raspberry Pi vendors mostly) but some of the Amlogic SBC vendors have lots of users

06:06 <chewitt> the main one for S912 users is Khadas, with their VIM2 board

06:08 <chewitt> and you wouldn't be the first to confuse LibreELEC with LibreComputer :)

08:50 <jernej> alyssa: just checking, have midgard_alu_op_ibitcount8 and midgard_alu_op_imov same op code? This patch here has two 0x7b opcodes: https://lists.freedesktop.org/archives/mesa-dev/2019-April/217593.html

10:17 <urjaman> https://www.youtube.com/watch?v=NAEQhoIh8x0 neverball actually runs quite nicely (under sway, the only thing i've done myself is disabling buffer_age ...)

16:44 <alyssa> chewitt: Good to know

16:44 <alyssa> jernej: Uh, no, I probably made a typo, thank you for pointing out!

16:45 <alyssa> urjaman: Indeed it does, it's one of my favourite time wasters on a panfrost laptop ;)

16:45 <alyssa> Personally I have to prefer STK to Neverball as far as gameplay goes, but performance is much better on the latter so

16:48 <alyssa> jernej: Rechecking notes, it should've been 0x7A

16:49 <alyssa> Thank you for noticing :)

16:49 <jernej> no problem :)

16:50 <alyssa> urjaman: It's good to know I haven't broken T760 support yet ;)

17:00 <urjaman> i'll try and tell you once you do :P

17:05 <urjaman> https://urja.dev/whodunthat.txt anyone have an idea on what bit of code would say '*** Caught unhandled unknown exception; terminating'? (kicad is based on wxGTK...)

17:18 <urjaman> i guess i'll just need to get compiling kicad and springle it with a bunch of printfs :P

17:19 <urjaman> *sprinkle

17:24 <alyssa> urjaman: :)

18:04 <ente> https://paste.xinu.at/VVZdC2/ rare footage of urjaman sprinkling printfs over a project

18:15 <hanetzer> ente: https://www.reddit.com/r/TenseiSlime/comments/as8lf5/did_yall_see_that_salt_bae_meme_in_the_latest_ep/

21:01 <urjaman> GL_MAX_COLOR_ATTACHMENTS

21:02 <urjaman> is 1 whilst kicad needs 3

21:03 <alyssa> urjaman: Don't even think about it :P

21:03 <HdkR> MRT woo

21:05 <urjaman> i think this is actually (one) of the same things that made the nouveau driver for the FX5200 (ancient geforce FX) to be unable to run the kicad gal ... but that had bigger problems too (like npot textures...)

21:05 <alyssa> urjaman: MRT is introduced in ES 3.0

21:06 <HdkR> ^

21:06 <alyssa> Panfrost is still hacking at ES 2.0

21:06 <alyssa> ES3 is a moonshot at this point

21:06 <alyssa> If I had an extra 40h/week, it'd be a different story :P

21:50 <alyssa> urjaman: That statement is so ominous :P

21:56 <urjaman> oh well, the list is just FX5200, the pandora (well, on that i had software trouble enough that i couldnt get it built non-accelerated for the ancient linux it has), the C201 with gl4es (couldnt deal with the shaders kicad has, among probly more things) ...

21:56 <urjaman> and panfrost as it is now :P

21:57 <alyssa> Pah

21:58 <HdkR> Eventually ;P

21:59 <urjaman> yeah atleast this time it should eventually be possible (just feels kinda wrong... always like "yeah a bunch of games work" "okay how about this thing that just draws lines onto 2D planes" "uhh nope")

22:00 <alyssa> Hehe

22:01 <alyssa> urjaman: If it requires MRT, it's definitely not "just" drawing lines

22:01 <urjaman> yeah i was checking what it uses it for (and since i'm not very into GL things, also what does that even do...)

22:03 <urjaman> the 2 target buffers it makes on initialization are called mainBuffer and overlayBuffer

22:04 <HdkR> I guess it is using a combined shader that does the main rendering and overlay at the same time?

22:05 <urjaman> yeah afaik it only has one shader

22:05 <HdkR> Bit silly since in the end it is going to have to blend together anyway

22:05 <HdkR> but w/e

22:05 <urjaman> one vertex shader and one fragment shader that is

22:10 <urjaman> and yeah the final operation is splatting one buffer to the output, then the other one on top of it and done... atleast in the case with no anti-aliasing

22:10 <HdkR> yep

22:11 <urjaman> there seems to be some code for a SMAA mode that i didnt read much

22:13 <HdkR> Probably just applies SMAA to the main output and then puts the overlay on top of the SMAA output

22:13 <HdkR> Since you don't want a post processed AA on your HUD elements