Bus error [was:Re: [XeTeX] Wanted: "exotic" OpenType examples]

Jonathan Kew jonathan_kew at sil.org
Fri Mar 4 14:28:57 CET 2005


On 4 Mar 2005, at 12:41 pm, Cyril wrote:

> Since Will asked for examples, I had a go at building some, using the 
> unicode consortium texts 
> <http://www.unicode.org/standard/translations/>, following the example 
> of Jonathan Kew.
> Well things were OK (as far as I can tell Amharic the various Greek, 
> Hindi and Tibetan (although not with Code2000) were fine) even though 
> I had a lot of overfull error messages, with results similar to Yves 
> Codet's Sankrit example where the lines bunch up at the right-hand 
> margin.
> Then I remembered the recent addition to XeTeX allowing for the 
> "hyphenation" of Chinese, and I was curious to see if it would work 
> with Japanese. I therefore pasted not the first little bit of the 
> text, but rather two paragraphs, saved the text and typeseted (sp?) 
> it.

To activate line-breaking without spaces or hyphenation positions, for 
scripts such as Chinese or Japanese, you need to enabled XeTeX's new 
"locale-dependent line breaks" feature. (Yes, I know this is 
undocumented as yet! I'm currently updating the XeTeX-notes file and 
samples.)

At the start of your file, add the commands:

	\XeTeXlinebreaklocale "ja"
	\XeTeXlinebreakskip 0pt plus 1pt

and all should be well.

Of course, even without that XeTeX has no right to crash in such a 
spectacular way! So that's something I should look into.

BTW, the "script=japanese" that I see on your font declarations won't 
achieve anything. It would be "script=hani" for CJK ideographs, 
"script=kana" for Kana (these are the registered OpenType tags); but in 
any case there's no script-specific shaping needed so you don't need a 
script tag at all.

JK



More information about the XeTeX mailing list