On Wed, 20 May 2026 at 21:10:18 +0100, Colin Watson wrote:
Do you think we should cherry-pick this patch or its equivalent into
pygame, then? I don't know this package especially well, but it looks
like a pretty simple patch.
Yes, I think so.
I would personally suggest completely deleting the last
`self.assertTrue(blit_surf.get_flags() & pygame.RLEACCEL)` assertion
from test_fill_rle(), rather than making it conditional as the pygame-ce
team did. My understanding is that it was true in the past as a result
of SDL implementation choices, but it was never an API guarantee that
performing edits to a RLE surface would leave it still in the RLE state.
smcv
