Support blitter with 8 horizontal levels (similar to NCBLIT_8x1)

[eschnett]

julia> [Char(ch) for ch in 0x2588:0x258f]
8-element Vector{Char}:
'█': Unicode U+2588 (category So: Symbol, other)
'▉': Unicode U+2589 (category So: Symbol, other)
'▊': Unicode U+258A (category So: Symbol, other)
'▋': Unicode U+258B (category So: Symbol, other)
'▌': Unicode U+258C (category So: Symbol, other)
'▍': Unicode U+258D (category So: Symbol, other)
'▎': Unicode U+258E (category So: Symbol, other)
'▏': Unicode U+258F (category So: Symbol, other)

[dankamongmen]

so, you can't really make a "blitter" out of this. i mean, you can, but it would be a very weird one. the whole reason why you can blit using halves, quadrants, sextants, and octants is that they regularly divide a set of pixels (2x1, 2x2, 3x2, and 4x2 respectively). you just have to solve for colors. these characters are only useful for dividing a cell into 8 columns, and you can only fill up successive columns. do you see? this is the same reason why we don't have a blitter built out of the existing vertical levels.

now, we do use the vertical levels in plots, which is why e.g. NCBLIT_8x1 exists. it probably never should have been called that, except that the plotting engine makes use of e.g. NCBLIT_2x1 (it makes use of any NCBLIT_Xx1). but we don't do horizontal plots. if i ever add them, this would definitely fit right in.

the one place where i can imagine it being useful is in progress bars. but it looks like we're already doing so.

so i'm going to close this, but thanks!

[dankamongmen]

so, you can't really make a "blitter" out of this. i mean, you can, but it would be a very weird one. the whole reason why you can blit using halves, quadrants, sextants, and octants is that they regularly divide a set of pixels (2x1, 2x2, 3x2, and 4x2 respectively). you just have to solve for colors. these characters are only useful for dividing a cell into 8 columns, and you can only fill up successive columns. do you see? this is the same reason why we don't have a blitter built out of the existing vertical levels.

there is a case to be made for what i've called "ultrablitter", which would be capable of using all glyphs existing outside of the (unpredictable) font (basically all block and line-drawing characters). basically it would build a histogram of colors, map each pixel to a color, and see which glyph out of all of them was best. but it would be very complex, probably slow, and i doubt you'd get much better imaging that what's already available outside of images constructed to use them. also, the space of available glyphs wold likely be irregular, so you'd have areas of different effective resolutions.

but that's for another day.

[dankamongmen]

does this make sense? do you disagree?