On Thu, 10 May 2007, Peter Stephenson wrote:

> Timo Aaltonen <tjaalton@cc.hut.fi> wrote:
>> On Thu, 10 May 2007, Timo Aaltonen wrote:
>>
>>>>  Could you send a transcript of your session (as simple as
>>>> possible) made with "script" so we can see what is being sent to
>>>> the shell input?
>>>
>>> Attached.
>>
>> Forgot to mention that I pressed each of these keys twice;
>> ö - odiaeresis
>> ä - adiaeresis
>> å - aring
>
> Hmmm... nothing appears to be coming out; I'm looking for byte sequences
> like 0xc3 0xa4 (ä).
>
> Could we try a different tack?  Could you check the "multibyte" option has
> come on (use "setopt")?  This should happen automatically.  If it hasn't,
> the shell may not have built properly and we need to go back and look at
> the configuration.

Apparently multibyte is _not_ on, and running 'setopt multibyte' doesn't 
do anything. Configure does say that multibyte support is enabled. Setopt 
output:

replicant work 1 % setopt
interactive
monitor
norcs
shinstdin
zle


Attached is a script output of the configure/compile phase. The resulting 
binary is linked to libiconv.so:

#narya zsh 5080 # ldd 4.3.4/bin/zsh

         Main  =>   4.3.4/bin/zsh
         libiconv.so  =>   /usr/shlib/libiconv.so
         libm.so  =>   /usr/shlib/libm.so
         libc.so  =>   /usr/shlib/libc.so

> If multibyte mode is apparently enabled, try "cat >test.out", typing one of
> the characters, then ^D.  "xxd" or similar should show you the file
> contains a two-octet character like 0xc3 0xa4 (ä), 0xc3 0xb6 (ö), or 0xc3
> 0xa5 (å).  This may not tell us much, but it's a start...
>
> If that's all OK, try seeing if the shell is recognising multibyte
> characters internally.  Take the contents of a file with a multibyte
> character in it (say test.out) and do
>
> print ${#$(<test.out)}
>
> That should say 1 (i.e. one character found).  If it says 2 the shell is
> interpreting the multibyte stream wrongly.

Yep, printing works, and I could edit the file just fine.

thanks for helping out,

t