monstr has quit [Remote host closed the connection]
i tried to dig into dEQP-GLES2.functional.fragment_ops.depth_stencil.random.* tests ...
they fail with lima and mali and i think this is a hardware bug. they fail, when we have a stencil test with GL_INCR/GL_DECR
my guess is, that hw has a problem, when it should do a GL_DECR with a stencil value of "0" or doing a GL_INCR with "255"
i have to do a little test program to verify that.
anyway, i wonder, if we can workaround this one in software.
so, i think of creating a shadow stencil buffer and updating the values during the draws/clears,
then we can change a GL_INCR on 255 to a GL_KEEP for example and handle all the possibilities. would that make sense or would this be an overkill :) any opinion if this is worth further investigation?
Barada has quit [Quit: Barada]
hm, not sure if this is the issue, i need to do some more tests :p
ente has quit [Read error: Connection reset by peer]
zombah_ has joined #lima
minicom0 has joined #lima
zombah has quit [*.net *.split]
minicom has quit [*.net *.split]
maciejjo has quit [*.net *.split]
egbert has quit [*.net *.split]
egbert has joined #lima
maciejjo has joined #lima
Ntemis has joined #lima
tlwoerner has quit [Quit: Leaving]
minicom0 is now known as minicom
Ntemis has quit [Read error: Connection reset by peer]