Since 1.04 the codepoints U+10000...U+10013 in Liberation Sans, U+10000...U+1000B in Liberation Serif, and U+10000...U+1000E in Liberation Mono, are filled with characters that used to be out of bound. The glyph don't actually match the codepoints. For example U+10005 LINEAR B SYLLABLE B001 DA is a fraction four in LiberationSans-Regular. The error was probably introduced when the font encoding went from Unicode BMP to Unicode Full.
Would you have any suggestions on the conversion? I have done my best to preserve that but seems at your feedback it was failed.
Created attachment 363148 [details] path reajusting encodings The patch does several things: - clears duplicates of .notdef, .null and nonmarkingreturn - renames the out-of-bound glyphs to previous names - adjusts glyph id and references.
Thanks Denis. Is this patch a proper fix which maintain closer quality with the original TTF delivered from Ascender?
(In reply to comment #3) > Is this patch a proper fix which maintain closer quality with the original TTF > delivered from Ascender? No, the tables DSIG, hdmx, LTSH and VDMX from the original TTF are not there. Fontforge can generate a DSIG. hdmx, LTSH and VDMX are ignored by Fontforge ; they would have to be included in the generated fonts in post-processing.
As far as I know, no hinting has gone missing and the kern table is the same as in the original font.
Denis how did you generate the patch btw? :)
patch is applied on rawhide