Re: [Groff] mapping of glyphs to Unicode

2006-02-15 Thread Werner LEMBERG
> > Personally, I prefer U+27E8 and U+27E9 which are never double-width > > glyphs. > > I agree; see the xterm screenshot in the other mail. > > So we _do_ need the possibility to map the same glyph to different > Unicode code points depending on the output device. So for Unicode > font files, t

Re: [Groff] mapping of glyphs to Unicode

2006-02-15 Thread Bruno Haible
Michail Vidiassov wrote: > Oh woe is me! I live under Mac OS X or Windows XP, > bdf fonts are off-limits to me :( > Maybe I have to use FontForge to convert bdf to bitmap truetype. There are also scalable fonts which have this glyph. The attached HTML file shows that on my Linux system, scalable

Re: [Groff] mapping of glyphs to Unicode

2006-02-15 Thread Michail Vidiassov
Dear All, On Wed, 15 Feb 2006, Bruno Haible wrote: Michail Vidiassov wrote: PS. BTW, can you point me to a font having that MATHEMATICAL * ANGLE BRACKETs ? Sure, the Misc-Fixed fonts from Markus Kuhn, such as -misc-fixed-medium-r-semicondensed--13-120-75-75-c-60-iso10646-1, have it. Oh w

[Groff] unicode support, part 7: cleanup glyph to name/number API

2006-02-15 Thread Bruno Haible
Hi, Now the glyph API is in good shape. This patch - fixes the naming, mentioning "glyph" instead of "index", - moves the 4 static functions that create and analyze glyphs out of the class 'font'. They semantically belong to the class glyph; the class 'font' is built on top of it. 200

[Groff] unicode support, part 6: public glyph to name/number API

2006-02-15 Thread Bruno Haible
Hi, Here is a patch that makes the glyph to name or number API available (for the Unicode font implementation) without the ugly hacks of the previous attempt. It makes the variants in nametoindex.cpp and troff/input.cpp more similar, and actually simplifies things: - Where character_indexer::nu

Re: [Groff] mapping of glyphs to Unicode

2006-02-15 Thread Bruno Haible
Werner Lemberg wrote: > > - devhtml maps "la" to U+2329, "ra" to U+232A. > > devutf8 and glyphuni.cpp map "la" to U+27E8, "ra" to U+27E9. > > I have a W3C document `entities.html' which describes HTML 4. In this > file, U+2329 and U+232A are used for ⟨ and ⟩ -- has this > changed meanwhile? Sti

[Groff] unicode support, part 6 X: public glyph to name/number API

2006-02-15 Thread Bruno Haible
Hi, The logical consequence of the last patch that added glyph::glyph_name(), needed for unicode fonts, is the reverse mapping from glyph to index or number. Please don't apply this patch! It is too ugly to have to convert a string from decimal to binary in order to find the number to which a gly

Re: [Groff] mapping of glyphs to Unicode

2006-02-15 Thread Bruno Haible
Michail Vidiassov wrote: > PS. BTW, can you point me to a font having that > MATHEMATICAL * ANGLE BRACKETs ? Sure, the Misc-Fixed fonts from Markus Kuhn, such as -misc-fixed-medium-r-semicondensed--13-120-75-75-c-60-iso10646-1, have it. It's used in xterm:

[Groff] unicode support, part 8: comments and tweaks for glyph API

2006-02-15 Thread Bruno Haible
Hi, Before freezing the glyph API, let me add more naming fixes and comments. Also, move some code from the name_to_glyph function to inside the character_indexer class. (This code is needed because character_indexer::ascii_char_index doesn't store the "charNNN" glyphs in the by-name table: a det