[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Groff] unicode support - questions
From: |
Werner LEMBERG |
Subject: |
Re: [Groff] unicode support - questions |
Date: |
Thu, 26 Jan 2006 20:51:39 +0100 (CET) |
> > classes
> > <Alike> = A :A 'A `A ... ;
> > <CJKpunct> = U+3000 - U+303F;
> > <Hiragana> = U+3040 - U+309F;
> > ...
> >
> > <CJK> = <CJKpunct> <Hiragana> ... ;
> >
> > properties
> > <CJK> width 24
> > ...
> > <Alike> kern V -3
>
> I see. I had imagined the same thing with just the properties and
> no classes, like the POSIX locale files.
Interesting. Maybe we can reuse some code for quicker coding. I'm
open to suggestions how the syntax should look like. Note that a
glyph name within groff is almost arbitrary; to mark classes as such
we need an escape mechanism in case the class marker(s) is used
elsewhere in the `classes' and `properties' sections.
> The classes make it easier to write small font files. On the other
> hand, the classes are arbitrary: two font files can describe exactly
> the same properties with very different class definitions. Not sure
> whether that's good or bad.
For the moment, I don't care. Later on, we could provide an `include'
section which makes it possible to load default classes from another
file, to be overwritten on demand.
> > I've no idea how to store such information efficiently within
> > memory.
>
> The three-level tables that are used in glibc's localedef program
> appear suitable here. They can store for example an complete width
> table for Unicode in ca. 12 KB, with an O(1) access (ca. 3 memory
> references).
Sounds great!
Werner