extend ncvisual_from_plane() to NCBLIT_PIXEL

[dankamongmen]

when ncvisual_from_plane() is run on a plane that holds a bitmap, we ought extract that as an NCBLIT_PIXEL ncvisual.

[dankamongmen]

is this not basically ncplane_rgba()?

[dankamongmen]

yeah, ncplane_rgba() is called by ncvisual_from_plane(). this would be about making ncplane_rgba() work. but what does that mean?

• we can't turn arbitrary glyphs into a bitmap...actually, shit, i guess maybe we can? nooooo...yes? no. sit on that.
• we could expand ncplane_rgba() to store back pixels, but that's already what it does
• we could actually extend ncvisual_from_plane() to work on sprixel planes, bypassing ncplane_rgba() and instead expanding the sprixel into an ncvisual...yes, i think that's what this bug is about

[dankamongmen]

when this is applied to a sixel, i think it ought return the quantized colors, so they can be inspected. to fill this from the actual sprixel, is what i'm saying. oooh it ought also reflect cutouts.

[dankamongmen]

we also want NCBLIT_BRAILLE.

[dankamongmen]

I've started the Braille work (defining the inverse table) in dankamongmen/braille-reverse.

[dankamongmen]

ack! still need NCBLIT_PIXEL!

[dankamongmen]

depending on what we do with #1539, this might suddenly become very easy indeed.

[dankamongmen]

hrmmm, note that this is going to be very difficult in kitty now that we're using reflected composition -- we have no copy of the visual. hrmm.

[dankamongmen]

yeah, i don't think we can really do this anymore with our current kitty scheme. closing as nonfix.