> 1. i followed your advice and made t8aenc.def t8menc.def t8kenc.def
> files. This becouse there is Khutsuri window in UNICODE V.5 as
> GEORGIAN SUPPLEMENT http://unicode.org/charts/PDF/U2D00.pdf
Hmm. Are Khutsuri and Nuskhuri used together, similar to uppercase
and lowercase letters in Western scripts? Then you don't need three
encodings but two should be sufficient.
t8kenc.def: contains Khutsuri at uppercase and Nuskhuri at lowercase
positions.
t8menc.def: contains Mkhedruli at both uppercase and lowercase
positions.
Note that you have to extend [EMAIL PROTECTED] (within the *enc.def files) as
shown in t5enc.def (and fontenc.sty).
> like this
> \DeclareFontEncoding{T8A}{}{}
> \DeclareTextSymbol{\u10A0}{\LastDeclaredEncoding}{192}
> ......
Instead of `\u10A0' I would use more friendly names, like \grgan,
\GRGAN, and \grgAN for U+2D00, U+10A0, and U+10D0, respectively.
Note that the use of \LastDeclaredEncoding makes mostly sense if you
create the t8*enc.def files from a common DTX file, which I strongly
recommend especially for documentation purposes.
> 2. dfu files lok like this
> \ProvidesFile{t8enca.dfu}[2006/08/19 v1.1 UTF8 support for Georgian]
>
> \DeclareUnicodeCharacter{10A0}{\u10A0}
> ......
OK (except the typo in the first argument of \Provides File).
> 3. t8a.enc file looks like this
> /UnicodeGlyphNamesGeorgianEncoding [ % now 256 chars follow
>
> % 192
> /u10A0
> % 196
> /u10A1
> .....
>
> question is do i need to define Latin (English) letter codes in this
> (.enc) files
Well, the LaTeX encoding `T8K' *must* contain some ASCII characters as
documented in the `encguide' file (which is part of the latest LaTeX
release). `*.enc' files aren't read by LaTeX but by dvips and
friends. It's not necessary that they contain ASCII characters -- for
example, you could put all Georgian glyphs into a single font -- but
then you need to construct virtual fonts.
> 4. when i supply fonts with package (ttf,pfb) with their support
> files for tex/latex they need to be 8bite fonts with 256 glyps. am i
> right?
It depends. Maybe it's really better to have a single Georgian font
containing all three letter forms, then constructing virtual fonts to
satisfy the LaTeX constraints. In that case, you would do something
like that:
georgian.enc: Contains all three Georgian alphabets and nothing
else. Since this encoding isn't directly related to a
Georgian LaTeX encoding I wouldn't call the file
t8k.enc or t8m.enc.
T8K: A virtual font with 256 slots constructed from the Georgian
font and, say, the EC glyphs for everything else (you should
choose a font which fits best to Georgian).
T8M: Ditto.
On the other hand, I strongly believe that existing Georgian fonts
already contain ASCII character *and* Georgian letters... It's really
your choice.
All what you and I have written in this mail exchange should be
documented somewhere. I thus strongly suggest that you use the
`fontinst' package to create FD, PL, VPL, and ENC files, writing
proper ETX files (and DTX for the stuff not covered by fontinst). As
an additional guide you could use how I've set up the C90 encoding for
Thai: You can find the stuff in the CJK package, directory
`utils/thaifont/tools' (however, the C90 doesn't use virtual fonts).
Finally, please remember that you have to set up the Georgian
hyphenation patterns to use the lowercase positions of the output
encoding(s).
Werner
_______________________________________________
Cjk maillist - [email protected]
https://lists.ffii.org/mailman/listinfo/cjk
