------------------------------------------------------------------------ r4656 | khaled | 2013-12-03 15:26:34 +0100 (Tue, 03 Dec 2013) | 11 lines Changed paths: M /trunk/source/texk/web2c/luatexdir/luafontloader/fontforge/fontforge/parsettfatt.c Do not duplicate glyph names when guessing them When reading fonts without glyph names (e.g. with version 3 ?\226?\128?\152post?\226?\128?\153 table), we try to guess glyph names for unencoded from GSUB substitution, but there were not any checks for duplicate name which could happen if font has two different substitutions in the same feature for the same base glyph (e.g. Palatino Lintype shipped by Microsoft has two ?\226?\128?\152i?\226?\128?\153 substitutions in the ?\226?\128?\152smcp?\226?\128?\153 feature, one to dotted i for Turkish under different language tag and a general one). We now check if the name is duplicated and just give up on that glyph. ------------------------------------------------------------------------ r4657 | khaled | 2013-12-03 16:13:20 +0100 (Tue, 03 Dec 2013) | 1 line Changed paths: M /trunk/source/texk/web2c/luatexdir/luafontloader/fontforge/fontforge/parsettfatt.c Get it to actually compile ------------------------------------------------------------------------