view: Cycle through multiple fg colours if supplied

[wezm]

Closes #41

@n8willis had a go at implementing #41 here's some sample output:

allsorts view -f ../prince/tests/data/fonts/morx/Zapfino.ttf -t 'Zapfino'\!'' -s LATN --features kern,mark,mkmk,liga --fg-colour ff0000aa --fg-colour 00ff00aa --fg-colour 0000ffaa

allsorts view -f /home/wmoore/Downloads/lanna-export-pdf-issue/LN_TILOK_V6_05.woff -s thai -t 'เวรญชฯกณฑ' --features kern,mark,mkmk,liga --fg-colour ff0000aa --fg-colour 00ff00aa --fg-colour 0000ffaa

Note: there's an invisible/zero-width glyph in this one, which is why it skips the second use of blue.

[n8willis]

Looks great! Thanks for taking a crack at this.

On the syntax, would there be something that prevents using a comma-separated list, à la --fg-colour ff0000aa,00ff00aa,0000ffaa ?

Thinking about it overnight, I believe that counting the invisibles/spaces as part of the target sequence (as you've done) is the only approach that makes sense. It doesn't do any harm for those who don't care, and if someone wants to see that an invisible or space has changed or moved between two calls with different options or different fonts, it's valuable.

[n8willis]

Actually, another question. Is the most desirable model that the color list applies to the input sequence or to the output sequence — i.e., after reordering might have occurred?

I was thinking output originally; perhaps that's the only one that makes sense here. The wrinkle is that you provide the input sequence, but arguably you don't know if the same number of glyphs will be in the output, if you're doing it to debug the font, in order to specify the same number/order of colors. That might be overthinking it though....

[wezm]

Actually, another question. Is the most desirable model that the color list applies to the input sequence or to the output sequence — i.e., after reordering might have occurred?

I think colouring based on input sequence would be quite difficult with the current code due to the need for yeslogic/allsorts#31

The wrinkle is that you provide the input sequence, but arguably you don't know if the same number of glyphs will be in the output, if you're doing it to debug the font, in order to specify the same number/order of colors.

The way I implemented it, it will cycle through the colours supplied, looping around if necessary. So you can supply as many colours as you want and it will always work.