On Sun, 2 Nov 2025, Chad Jablonski wrote:
Reviewed-by: Chad Jablonski <[email protected]>
Tested-by: Chad Jablonski <[email protected]>
Hey Balaton!
This fix looks correct to me. Source and destination should use
their respective bits. I applied my latest HOST_DATA changes on
top of this and saw no problems when rendering text in xterm
with rage 128.
There is one thing I _did_ notice. I implemented similar behavior with the
GMC_SRC_CLIPPING and GMC_DST_CLIPPING fields (default and leave alone).
But I implemented them as a copy of the default values during the write
when the bits are set to default. Both approaches seem to work. If you want
to keep them consistent I can test the actual hardware to see which approach
it uses or I can just update my patch to match this style.
Matching hardware is more important than matching style, after all this is
supposed to emulate hardware and we already found places where the docs
were wrong. So if you can test and match what hardware does that would be
best.
Regards,
BALATON Zoltan
