Are the changes in tottfgpos.c, lookup*.c automatic? makeutype.py doesn't seem to touch these files.

Only the first commit was automated. For the rest, I copied the data from unicode.org and learn.microsoft.com, applied regex substitutions to get it into the right format, and made various manual edits like adding Script| where necessary.

Only the first commit was automated. For the rest, I copied the data from unicode.org and learn.microsoft.com, applied regex substitutions to get it into the right format, and made various manual edits like adding Script| where necessary.

@dscorbett, could you please give more detail and links or perhaps outline the steps and relevant regexes? It would be good to have at least some documentation for this process.

For each script in scripts in tottfgpos.c, I copied its escaped, abbreviated code point set from this UnicodeSet utility and applied regex substitutions to get something like \u1B00-\u1B4C\u1B4E-\u1B7F into 0x1b00, 0x1b4c, 0x1b4e, 0x1b7f. For localscripts, I copied the table from OpenType’s script tag documentation and applied substitutions to get Beria Erfe 'berf' into { N_("Beria Erfe"), CHR('b','e','r','f') },. For scripts in lookupui.c, I copied that same table and applied substitutions to get { (unichar_t *) N_("Beria Erfe"), NULL, 0, 0, (void *) CHR('b','e','r','f'), NULL, 0, 0, 0, 0, 0, 0, 1, 0, 0, '\0'},. For languages, I copied the table from OpenType’s language system tag documentation and applied substitutions to get Eastern Abenaki 'AAQ ' aaq into { (unichar_t *) N_("Eastern Abenaki"), NULL, 0, 0, (void *) CHR('A','A','Q',' '), NULL, 0, 0, 0, 0, 0, 0, 1, 0, 0, '\0'},. I didn’t record the specific regexes I used. Besides that, there were many manual edits and miscellaneous special cases, especially to avoid modifying existing code unnecessarily.