Re: VTE on Solaris

[HideToshi Tajima <hidetoshi tajima sun com>]

Owen Taylor wrote:

> a) I think vte actually uses C library functionality on the resulting
>    wide characters - in particular wcwidth().
>
>    One the other hand, I think there is an assumption elsewhere in
>    VTE that wchar_t is Unicode, so maybe what is needed is to 
>    include a wcwidth() implementation on platforms where
>    __STDC_ISO_10646__ isn't defined.
I think using wchar_t type is incorrect if vte is assuming it as 
Unicode. It should define own char type.

Also, instead of using C librarry wide char functions, zve should have 
own implementations for
wcwidth, wcslen, wcscpy,  mbsrtowcs, and any other, without including 
<wchar_t>.

> b) UCS-4 may not be the right endianness.
How is "wchar_t" expected to be encoded in the current vte code? Using
either "UCS-4LE" or "UCS-4BE" will become right for it (at least for 
performing quick testing)?

Regards,
Toshi