Re: File name completion with non-ascii filenames

[Kenichi Handa]

Eli Zaretskii <address@hidden> writes:
> On Tue, 25 Dec 2001, Kenichi Handa wrote:

>>  So, it seems that we must use a slow method at least in the
>>  case that completion-ignore-case is non-nil (it is nil by
>>  default).

> Yes, that's what I had in mind.  AFAIK, only the Windows port routinely 
> sets completion-ignore-case to non-nil.

> However, since working with multibyte strings means we need to decode 
> each file name returned by readdir, to address Richard's concerns, I'd 
> like to try to avoid decoding even if completion-ignore-case is non-nil.  
> So, could you please tell what is the best method of deciding whether 
> file-name-coding-system is the one for which the current unibyte 
> case-conversion tables are set up?

It's impossible to detect something about a coding system
from the setting of case table.

But, the current unibyte case table is set up by
set-language-environment.  So, the simplest way is to
register such a coding system that yields the same result on
encoding as string-to-unibyte in each language environment,
and set some variable to this coding system in
set-language-environment.

---
Ken'ichi HANDA
address@hidden