debwarrior not triggering blitstacking

[dankamongmen]

I'm convinced that our fix to #1068 is good, as are our unit tests. debwarrior, however, is still showing the old behavior. It showed the desired behavior before introducing the crender->s.blittedquads condition to paint(), but we need that, and it's correct. No, the real problem (as verified by debugging) is that somehow the overworld map is being blitted (using NCBLIT_2x2) without quadblit bits. Celes does get the expected bits, and propagates them down, but then we hit the overworld, have a valid non-0 crender->s.blittedquads, but no cell_blittedquads(vis).

This isn't Notcurses proper, but get it resolved as part of the blitter stacking saga. There's definitely a bug here, though it might just be out in debwarrior.

[dankamongmen]

Nope, it was a bug of ours, one I figured out immediately after cutting 2.1.8 (argh). Fix is incoming.

[dankamongmen]

Yep, debwarrior now appears to work perfectly. w00t! =]

[dankamongmen]

Indeed, she finally looks right =]