[XeTeX] XeTeX for Linux (first experiencies)

[Yves Codet]

Hello.

Le 6 juin 06, à 18:02, Jonathan Kew a écrit :

> On 6 Jun 2006, at 4:56 pm, Ralf Stubner wrote:
>
>> Yves Codet <ycodet at club-internet.fr> writes:
>>
>>> \documentclass[12pt,a4paper]{book}
>>> \usepackage{fontspec}
>>> \setromanfont[Script=Greek]{Gentium}
>>> \language=\greek % this is for hyphenation
>>
>> Strange, I get the follwoing message here:
>>
>> Package fontspec Warning: Font 'Gentium' does not contain script
>> 'Greek' on input line 4.
>
> Gentium has no OpenType Greek support, so the [Script=Greek] option
> here doesn't actually achieve anything.

My bad, I hadn't noticed. Gentium has neither OT nor AAT support for 
Greek apparently. But Greek is quite correctly displayed in XeTeX's 
output.

>
>> ! Undefined control sequence.
>> l.5 \language=\greek
>>                      % this is for hyphenation
>> ?
>>
>> I can understand the warning from fontspec, but it seems that the
>> '\language' command isn't known.
>
> No, that's \greek that is unknown. I suspect Yves is using a custom
> format file with extra hyphenation patterns loaded and \greek defined
> as a language number.

Indeed, and you helped me to fix the file. Here it is:

-------------- next part --------------
A non-text attachment was scrubbed...
Name: uni-grhyph.tex
Type: application/x-tex
Size: 5115 bytes
Desc: not available
Url : http://tug.org/pipermail/xetex/attachments/20060606/9a7092ee/attachment-0001.tex 
-------------- next part --------------


It's an adaptation of "grhyph.tex". There are a few patterns at the end 
which are commented out. I suppose they're useless with Unicode, but I 
left them there in case my supposition turns out to be wrong and they 
have to be used.

By the way I have an advice to ask. There are more elaborate 
hyphenation files including etymological patterns, and I planned to do 
the same adaptation with one of them, when I had a sudden doubt. There 
was a short discussion on this list about hyphenation in transliterated 
Sanskrit, which is a similar problem. In original scripts Indian 
scribes simply break after any vowel, regardless of etymology. If one 
does the same in transliteration there will be hyphenations like:
	(1) ?dr?-?ya
But some will complain that the root of that verb is "dr??" and that the 
hyphenation should be:
	(2) ?dr??-ya
In my opinion there's no reason to invent practices which never existed 
in India (as far as I know), and I would stick to (1). Being no 
specialist of Greek I've never read any manuscript. Did Greek scribes 
break words after etymological patterns or not? I presume they didn't 
but I may be wrong and some members of this list can enlighten us. If 
they didn't take etymology into consideration I think there's no reason 
to invent... (almost the same as above). In that case that 
"uni-grhyph.tex" is fine for ancient Greek. For modern Greek I don't 
know.

Kind regards,

Yves