This is much simpler, for embedding of fonts into PDFs, than embedding the fonts by the method used in out-of-the-box ANT, which is to make a big font appear to the program as if it were chopped into many smaller fonts.
Subsetting – embedding only glyphs that appear in the document – should be pretty simple, given that I am already using FontForge to edit the font, and have had that sort of thing working before with FontForge and OpenType fonts (but not with Type 1 fonts).
(Adobe font nomenclature is really screwed up; the CID-keyed Type 1 fonts actually aren’t Type 1 fonts but rather are Type 2. Or they are Type 1C. Actually I don’t know what the hell they are called. And actually they are embedded as a Type 0 font, or as part of a Type 0 font, or children of a Type 0 font, or whatever.
CID-keying, on the other hand, seems to mean little more than assigning numbers to the glyphs in lieu of addressing them by a spelled-out name, as is done in regular PostScript fonts.)