Red Hat

Hi,

We are trying a few things on our new RHEL machine and I noticed some strange behaviours happening in awk.
We are using the swedish locale sv_SE.iso88591, where you have three additional letters in the alphabet: ... X Z Å Ä Ö (in that order).

I can sort text files following the above natural order no problem.

Code:

[cscyabl@basstest03 ~]$ cat test
Å
Ö
Z
Ä
X
[cscyabl@basstest03 ~]$ sort test
X
Z
Å
Ä
Ö

But awk is giving me some serious trouble:

Code:

[cscyabl@basstest03 ip]$ echo | awk '{if ("Ä" < "Å") print "error"}'
error
[cscyabl@basstest03 ip]$

It seems to react like LC_ALL was set to C which it isn't...
Any suggestion for making the comparison follow the natural order required by the contents of my locales? I have also tried using gawk but it didn't make any difference...

Cheers