Pixelarray SDL3 #3309
Pixelarray SDL3 #3309
Conversation
| @@ -828,13 +859,12 @@ _compare(pgPixelArrayObject *array, PyObject *args, PyObject *kwds) | |||
| return 0; | |||
| } | |||
|
|
|||
| new_format = surf->format; | |||
There was a problem hiding this comment.
This function goes from 3 pixel formats to 2. new_format == format as the code was written previously.
Should it actually be a format derived from new_surf instead of surf? Was the previous implementation a typo?
I'm not fully sure, but this translates it as was written.
Starbuck5
force-pushed
the
pixelarray-sdl3
branch
2 times, most recently
from
644a2ec to
4696146
Compare
Starbuck5
force-pushed
the
pixelarray-sdl3
branch
from
92f0a40 to
ec2bdd0
Compare
| @@ -1021,20 +1029,23 @@ _array_assign_array(pgPixelArrayObject *array, Py_ssize_t low, Py_ssize_t high, | |||
| } | |||
| break; | |||
| case 3: { | |||
| // Note: | |||
| // Why is the 24 bit case pixelformat aware but none of the rest are? | |||
| // - Starbuck, jan. 2025 | |||
There was a problem hiding this comment.
How are the rest of the paths not pixelformat aware? I think they are assuming that source and destination pixelformats are same and therefore directly copying over the data. But I didn't find this check enforced in the code anywhere on a quick search?
There was a problem hiding this comment.
I just mean they don't take into account the pixelformat. As you say, it looks like they just assume the source and destination are the same. Going through the code we have a lot of pixel stuff in pygame-ce and lots of it has little suspicious items like that.
There was a problem hiding this comment.
oh well, guess that can be addressed in future PRs, it shouldn't block this one
Starbuck5
force-pushed
the
pixelarray-sdl3
branch
from
ec2bdd0 to
66a7879
Compare
|
Rebased so it uses latest SDL2/SDL3 updates. (has it on the branch, so you don't need to swap dep versions between main and this). |
Starbuck5
force-pushed
the
pixelarray-sdl3
branch
from
66a7879 to
11961d7
Compare
Today I ported a few modules and worked on a few others.
I first did everything with explicit SDL2 vs SDL3 code to observe patterns, then I made a few new compatibility macros to match what I saw.
This PR is those macros + the pixelarray module.