[XeTeX] Strange hyphenation

[Jonathan Kew]

On 10 Jul 2005, at 2:23 pm, Musa Furber wrote:

> Something I noticed after updating to the current version of XeTeX  
> is that the word Allāh (transliterated, just as it appears in this  
> email) is sometimes hyphenated after Al-. I can suppress this using  
> by putting the word within a \makebox{...}.
>
> Isn't the default that hyphenation is suppressed for words such as  
> this?

Think about what would cause it to be suppressed.... If you do  
something like "All\={a}h" in plain TeX or LaTeX, then the accent  
command means that the word is not just a plain sequence of  
characters, and so it becomes ineligible for hyphenation. And if you  
enter it as "Allāh", and use LaTeX macros such as \usepackage[utf8] 
{inputenc} to interpret this, then the input bytes will be getting  
translated into internal macros of some kind, probably rendering in  
the end via an \accent command or similar. So again, the word is not  
a simple character sequence. And TeX doesn't hyphenate such "complex"  
words.

But if you're using Xe(La)TeX with Unicode text and a Unicode- 
compliant font, then "Allāh" is a simple sequence of five characters,  
with no internal TeX commands or anything, and so is eligible for  
hyphenation according to the patterns currently in force.

If this is not appropriate, you could tag the word as being in a  
different \language, or you could add it to the \hyphenation  
exceptions, or (as you've observed) you could box it so that it isn't  
part of the top-level text in the paragraph.

JK