Re: XeTeX encoding problem

[Masamichi HOSODA]

>> I've created a patch that uses native unicode support of both XeTeX and 
>> LuaTex.
>> It works fine in my XeTeX, LuaTeX and pdfTeX environment.
>> Except, LuaTeX create broken PDF bookmark.
>>
>> How about this?
> 
> It looks mostly all right. We'd need to wait until we have your
> copyright assignment on file before merging a patch of this size.

I sent it, today.

> The main change I'd make is to turn it off by default to avoid the
> risk of breaking something that worked before (for example: if someone
> didn't have the right fonts installed), but it would be easy to modify
> your patch to do this.

If XeTeX/LuaTeX is used and @documentencoding is UTF-8,
using native UTF-8 support is very natural for me.

Of course, even if using XeTeX/LuaTeX, if @documentencoding is US-ASCII,
I think that the native UTF-8 support may be turned off.

Most users use pdfTeX instead of XeTeX/LuaTeX.
The risk of using native UTF-8 support is few
because pdfTeX does not have the function.

If you consciously use XeTeX/LuaTeX instead of pdfTeX
and consciously set @documentencoding to UTF-8 instead of US-ASCII,
it means that you want to use native UTF-8 support in my humble opinion.

Otherwise, if you use UTF-8 characters on XeTeX/LuaTeX
without native UTF-8 support, some problems can be happen.
(something like ``Table of Contents'' broken etc.)
I think this is obviously a risk.

Additionally, CJK characters can not be used without native UTF-8 support.