Re: Unicode/Mule (Re: null-device)

[Eli Zaretskii]

On 19 Jul 2001, Karl Eichwalder wrote:

> Mule has issues (and you all know about them); maybe, the \201 issue is
> fixed by -- unfortunately, I've still to handle broken files coming in
> (on account of the Translation Project).  It was (and is?) Emacs
> creating those files.  And Emacs 21.x is still clueless to detect this
> mess and to guide the user to fix this files.

IIRC, this is the first time someone suggests a feature that would try to 
fix those messed up files.  I think it's a good idea, if those files can 
be repaired in many cases.

What happens if you visit those files with "C-x RET c emacs-mule RET" 
before "C-x C-f"?

> And then this rule "é" != "é" (in case you're using different charsets).
> Why doesn't this rule apply for "a", too?

Because the current representation of characters has only one ASCII 
charset, but many disjoint Latin-x charsets.  So the above is a 
consequence of the design of the character representation used by Emacs.

Since the decision that Unicode-based representation is better was made 
several years ago, it is futile to reopen these issues: we all basically
agree the "é" != "é" is not good.  The issue at hand now is to work on 
carrying out those decisions.  I find it sad that a decision that took so 
long to arrive at is still on paper, after all those years.

> The consequence is, Gnus often thinks it has to create a
> multipart message...  Yes, it will only do so if you'll enter three 'y'
> (yes) in a row -- this isn't "user-friendly" (Eli).

I don't understand that last remark: I don't have anything to do with how 
Gnus handles this.

> I again admit my techinical knowledges are rather limited; for details
> please re-read Eric Naggum's statements (I'm still convinced he's
> right).  It isn't Eric alone who argues against Mule.

That argument is long over; see above.  It's time to put one's money 
where one's mouth is, and start coding.  Otherwise, it is my belief that 
this will remain a pipe dream.

> But it's Eli who permanently complains "we don't have enough developers".

That's what I think.  I'm entitled to tell that whenever I think it's 
appropriate: a development team that is not large enough cannot be 
expected to make radical changes too frequently, especially when the 
subject matter is outside of the immediate area of expertise of many of 
us.

> It isn't nice to blame us who we try to show ways out.

The ways out are known and agreed upon.  I wasn't assigning any blame.  
Either you want Emacs to switch to Unicode-based internal representation, 
or you don't.  If you do, and if you need it badly, please consider 
helping us make it happen.