New to Typophile? Accounts are free, and easy to set up.
I'm prototyping a cursive font where, at least on paper, any word with a "t" in it has the crossbar on the t extended to the left and right over letters as long as this fits; verticals in h, l, d, etc. will cut off the bar, so something like:
becomes something like:
I'm not entirely sure how to implement this, though. I was thinking of using many-to-many GSUB rules based on glyph classes, but no matter how I look at it, it turns into needing recursive GSUB behaviour...
1) t [low] ->[t with bar that continues right] [low with left-connecting overbar]
If you load http://pomax.github.io/CFF-glyphlet-fonts in Firefox, you'll note that the "small" font, without GSUB table, renders the character "~" as a rectangle with a cutout. The second font has a GSUB table for the 'liga' feature with a lookup type 4 to turn the sequence "c,u,s,t,o,m" into "~", http://github.com/Pomax/CFF-glyphlet-fonts/blob/gh-pages/binaries/with%20GSUB/customfont.gsub.ttx#L271, which works in Firefox, but not in Chrome or IE (of course IE doesn't like plain .otf, so the fonts are repackaged as uncompressed WOFF as well, as fallback option).
For which reason and when are 'subtables' created? Adobe's OpenType Feature File Specification does not help me any further understanding the structure of GPOS/GSUB.