Re: Automatically Saved Files Employ Different Coding

[Alexander Winston]

On Fri, 2004-03-26 at 12:15 +0900, Miles Bader wrote:

> Alexander Winston <address@hidden> writes:
> > > > As you can see, the automatically saved file uses a different coding,
> > > > emacs-mule, instead of the default, utf-8.
> >
> > > That makes it (significantly) faster, and also makes it sure that the save
> > > will work and will not encounter a char that can't be encoded.
> > 
> > Which characters cannot be encoded using UTF-8?
> 
> Do `C-h h', and then save the resulting buffer as a file, forcing it to
> use utf-8.  Now re-read the file.
> 
> Note that it looks ... _mostly_ the same.  But not exactly.  [And this
> is assuming you have `utf-translate-cjk-mode' turned on; otherwise, it
> couldn't ready any of the CJK characters.]

Actually, the UTF-8 version appears better than the original. Compare
the first page at <http://home.comcast.net/~alexander.winston/p1.png>
and the second page at
<http://home.comcast.net/~alexander.winston/p2.png>.

I'm unsure as to why Emacs has problems converting Arabic to UTF-8 from
the original encoding, as I normally do not encounter any problems using
Arabic. See screenshots
<http://home.comcast.net/~alexander.winston/w3m.png> and
<http://home.comcast.net/~alexander.winston/ar.png>, for example. The
former was taken while using w3m and the latter while using Emacs; both
were taken while viewing <http://www.arabia.com/>, a site that seems to
use valid UTF-8.

signature.asc
Description: This is a digitally signed message part