Not supporting <use> increases the font file size and also increases the development cost of open source tools. googlefonts/nanoemoji#266
(for context, this is me with working code for ttLib wanting to avoid the need to cycle type-1 fonts through makeotf solely to make them readable. On the plus side, all I do with them is read them; I don't need to build any type-1 binaries...)
Cool; thanks!
Font stuff would sure be a lot easier without all the corner cases.
(I dunno if anyone else has noticed that.)
🚀🚀🚀
yes, it's impressive! haven't had a chance to dig into it yet though
specifically, it's in unicodes_missing, where it tries to log a missing unicode val, it's tripping up with
TypeError: %X format: an integer is required, not str
I'm populating the Subsetter with a unicodes=[] list. That seems like pretty basic stuff, so I'm obviously not seeing what's wrong for some simple. Any clearer eyes out there?
(Anyway, for anybody who stumbles across this late at night, the origin of the approach I used was this SO question, specifically mark2mark's answer: https://stackoverflow.com/questions/55009981/how-to-use-pyftsubset-of-fonttools-inside-of-the-python-environment-not-from-th )
passing a list of glyphs works as expected, but it seems like something different is necessary for the populate to get unicodes
instead of passing
subset.populate(unicodes=["U+0041", "U+0042", ...]), you should pass instead the decimal integer value for those hexadecimal unicode codepoints
just strip the leading "U+" and pass the rest to
int(..., 16)
hexadecimal is base 16 ;)
should
parse_unicodes() produce the right format?
because I also tried passing the ranges of interest through that & got the same error.
although those could be two separate issues
How should I set
sTypoAscender and sTypoDescender in a 1000-unit-per-em font such that ascenders of around 1000 units and descenders of around 1250 units do not overlap adjacent lines in HTML with the CSS declaration line-height: 1? The seemingly obvious solution of setting them to 1000 and -1250 does not work: the lines still overlap and the first line is too high within the containing box.
also remember the usetypometrics flag somewhere (head?)
I'm looking at booleanOperations and skia-pathops, but most of the example code I find is from people working on defcon or glyphs-source objects, not operating directly on TTGlyphs...
no idea but i think you can run bOps ops on ttglphs and if not, draw them into ufolib2 glyphs and bOps on that
you can even call it as a script
From the console
fonttools ttLib.removeOverlaps --help
requires skia-pathops
without roundtripping through glyphs, certainly
you create a pathops.Path from a TTGlyph, then call intersection then convert back to TTGlyph, replacing original
You want to call the pathops.op() method with two Paths and PathOp.INTERSECTION operator
thanks @anthrotype:matrix.org !