address@hidden: Re: [Xpert]Using character sets without a locale]

[Richard Stallman]

This message has major implications for the issue of
internationalizing menus.  It seems we might be able to do this now
in many cases by encoding in utf-8.

When Emacs converts to a representation that is mostly Unicode but
with some extensions (for instance, to avoid Han unification), it
seems we will need to convert the other characters to Unicode somehow
in order to display them in menus.

------- Start of forwarded message -------
To: address@hidden
cc: address@hidden
cc: Keith Packard <address@hidden>
Subject: Re: [Xpert]Using character sets without a locale 
From: Keith Packard <address@hidden>
In-reply-to: Your message of "Tue, 25 Dec 2001 14:40:16 MST."
             <address@hidden> 
Date: Tue, 25 Dec 2001 15:27:59 -0800
Sender: address@hidden


Around 14 o'clock on Dec 25, Richard Stallman wrote:

> Is there any platform-independent way of using fontsets
> and multibyte characters with Xlib?

Others may know better, but as far as I know, the multi-byte support in 
Xlib is dependent on the ANSI/ISO C language locale model and hence relies 
on underlying C library support of each locale.

The lack of standardization in the underlying locales supported is part of 
the reason I believe many projects are moving away from a locale-based 
model for internal text support and towards a model where the 
locale-specific text is converted at the operating system boundaries of 
the application to a standard Unicode encoding.  

Once the text is internally represented in Unicode, there are convenient
Xlib functions that accept utf8 encoded strings (Xutf8DrawText et al.)
This will also ease the transition from core X text support to the newer 
client-side font support of Xft which provides only Unicode APIs.

Keith Packard        XFree86 Core Team        Compaq Cambridge Research Lab
------- End of forwarded message -------