[mew-int 2459] Re: A patch: make-string in emacs 23

[Kazu Yamamoto ( 山本和彦 )]

Hello,

Sorry for keep waiting for a long time. After long discussion, Emacs
23 has come back to the original aset.

I think Mew can now work on Emacs 23 without any modification.

--Kazu

From: Harald Hanche-Olsen <hanche at example.com>
Subject: [mew-int 2420] Re: A patch: make-string in emacs 23

> > > when I was retrieve the mails from gmail, it works because 3 of 10 messages
> > > had been retrieved, how can I resolve this?
> > 
> > I guess that this is a bug of Unicode Emacs.
> > 
> > As I mentioned many times, Mew does not support Unicode Emacs at this
> > moment. I have compiled it in the past but it was really buggy. 
> > 
> > If you want to use Unicode Emacs because of the anti-alias feature,
> > you should use it at your own risk (including the bug above).
> 
> The problem that was discussed then still persists.  What is different
> is that now unicode emacs IS the main development branch of emacs, so
> more people will be using it.  For better or worse, I think mew will
> have to adapt.
> 
> I may add that I have been using mew with unicode emacs for a couple
> years, with hardly a problem at all.
> 
> The problem referred to in the above discussion came about because
> (make-string len ?x) will create a unibyte string in unicode emacs.
> It did then, and it still does.  This bombs out, of course, when you
> try to stuff multibyte characters into the string.
> 
> A simple cure is to replace the ?x in the code
> 
>     (let* (...
> 	   (tmp (make-string len ?x))))
> 
> at the top of mew-addrstr-parse-syntax-list by
> 
>   (if (char-valid-p 256) 256 ?x).
> 
> This fix costs next to nothing, it does not rely on any emacs version
> number magic, and it solves the problem once and for all.
> 
> (I got bitten by this all over again after upgrading my mew, since I
> had forgotten that I had patched my previous copy.)
> 
> - Harald
> 
>