James Holderness

> Setting the colors with VT will set a runtime "colors have been overridden" flag, and previewing a scheme will clear that flag out. Can't we just make it so...

> Unfortunately we don't really have the architecture in place at the moment to say "this part of the settings changed" I get that, but I thought what you were...

There was a point that Carlos raised a while ago which I think may have got lost in the collapsed comments: could we please use the actual color names for...

> the thing that made me go "hm?" was the fact that can't people assign whatever colors they want to the indexes? Yeah, it's possible some users will assign weird...

I suspect this is a duplicate of #11794, in which case it should be improved by PR #11890 (once that's released).

I thought this had come up before, and I also thought there was an API somewhere that would force the frame to render before pass-through so we could trigger that...

OK, that's probably what I was thinking of, but that doesn't solve this problem. The content being flushed there is just the VT engine's output buffer. That includes the pass-through...

That seems reasonable to me. As I understand it, that's the equivalent of what we're doing manually in a couple of places in `AdaptDispatch` when we call `TriggerFlush` before returning...

> I love how removing UnicodeStorage obviates the need for rows to have IDs, which removes the entire machine built around making sure row IDs remain sane I think it's...

> Why not store or reference that data in `ROW` itself? In the case of sixel, 99% of the time you're not likely to have any data attached to the...