problem with 8-bit characters

problem with 8-bit characters

[Paul Falstad]

[-- Attachment #1: Text --]
[-- Type: text/plain , Size: 1565 bytes --]

Hi,

I would like to mention a bug that's been there since the beginning
(and has come back to haunt me).  If you have a filename which
includes one of the character tokens (like 0x84) then zsh will
of course have lots of problems dealing with that file.

example:

zsh% ls Albumbl?tter' Op. 124.wav'     
ls: Albumbl
           tter Op. 124.wavv: No such file or directory
                        ^^^^

bash$ ls -l Albumbl?tter' Op. 124.wav'
-rwxrwxr-x   1 pf       users    11277884 Apr 19 00:30 Albumbl?tter Op. 124.wav
bash$ echo Albumbl?tter' Op. 124.wav' | od -c
0000000   A   l   b   u   m   b   l 204   t   t   e   r       O   p   .
0000020       1   2   4   .   w   a   v  \n
0000031

This particular file is on an smbfs filesystem on linux.  The file is
called "Albumblätter Op. 124.wav" on the PC side but is showing up
with a \204 (0x84) instead of an "ä" on the linux side because of a bug in
smbfs.  Fortunately there appears to be a fix for the smbfs problem
so I'm going to try that out.  However zsh isn't handling this very well.

I expected that characters like 0x84 would never show up anywhere but
apparently that isn't the case.  This is probably a real low priority
bug and it seems very hard to fix, but I just wanted to point out that
it is a real problem in some cases in case anyone comes up with any
clever ideas for how to fix this.

Thanks much.


-- 
Paul Falstad, paul@falstad.com, 805-966-4935, http://www.falstad.com/
work: paul.falstad@openwave.com, 805-957-1790, http://www.openwave.com/

Re: problem with 8-bit characters

[Peter Stephenson]

Paul Falstad wrote:
> I would like to mention a bug that's been there since the beginning
> (and has come back to haunt me).  If you have a filename which
> includes one of the character tokens (like 0x84) then zsh will
> of course have lots of problems dealing with that file.
> 
> example:
> 
> zsh% ls Albumbl?tter' Op. 124.wav'     =
> 
> ls: Albumbl
>            tter Op. 124.wavv: No such file or directory
>                         ^^^^

Are you sure this is still there?  This was supposed to be fixed some time
ago in both 3.0 and 3.1.  I can't get this to happen now.

-- 
Peter Stephenson <pws@pwstephenson.fsnet.co.uk>
Work: pws@csr.com
Web: http://www.pwstephenson.fsnet.co.uk

Re: problem with 8-bit characters

[paul]

> > example:
> > 
> > zsh% ls Albumbl?tter' Op. 124.wav'     =
> > 
> > ls: Albumbl
> >            tter Op. 124.wavv: No such file or directory
> >                         ^^^^
> 
> Are you sure this is still there?  This was supposed to be fixed some time
> ago in both 3.0 and 3.1.  I can't get this to happen now.

yeah it's fixed in 3.1.9.  I'm way back on 3.0.1.  I looked at the
3.1.9 code and saw that #define Pound 0x84 was still there but didn't
notice the trick that was used to fix this bug (until just now)..  Sorry.