incomplete printout ending on german umlauts with SLES9 due to

first of all I apologize if this is not the best group to post my
question, I spent quite some time searching for some related to "SLES9
LC_ linux csh ..." but not really successful.

our application generates listfiles with mixed ksh/csh-scripts
consisting of data from oracle databases where of course our german
customers have german umlauts, like "oe"=3D"ö"=3Dx'F6' (and PLEASE DON't
ask me to change any NLS_LANG_* whatsoever variables ... ;-)

with SLES8 we had no problems setting LANG=3Dde_DE and LC_ALL=3DC in the
login-ksh
(also relying on the format/order of "ls" for example, related to
those variables)

now with SLES9 we suddenly realized crashing printoutput done by
(t)csh scripts (using lp ...) which only prints up to the character
before the first german umlaut.
This seems to be due to the /etc/csh.cshrc on that machine, which sets
LANG and LC_* to de_DE.UTF-8 and "deletes" LC_ALL

I know I could create ~/.cshrc with setenv LC_ALL C and LANG de_DE.

I know I might even export CSHRCREAD=3Dtrue in my login-ksh-environment
to prevent csh from resetting those (but maybe others as well?)

Are there any additional suggestions to get the old behaviour?

Furthermore of course I want to understand WHY these settings are made
now with SLES9 to determine if we will have to think about those
things sooner or later again and have will have change the application
anyway (and in which way?) which will be a great effort (additionally
testing a lot of incoming/outgoing data-interfaces for example which
all depend on *LANG* and other settings as well ...).

Thanks for any answers related to the topic,
bine

Re: incomplete printout ending on german umlauts with SLES9 due to LC_-settings from /etc/csh.cshrc

On 2008-04-03, bine wrote:
>
>
> first of all I apologize if this is not the best group to post my
> question, I spent quite some time searching for some related to "SLES9
> LC_ linux csh ..." but not really successful.
>
> our application generates listfiles with mixed ksh/csh-scripts
> consisting of data from oracle databases where of course our german
> customers have german umlauts, like "oe"="ö"=x'F6' (and PLEASE DON't
> ask me to change any NLS_LANG_* whatsoever variables ... ;-)
>
> with SLES8 we had no problems setting LANG=de_DE and LC_ALL=C in the
> login-ksh
> (also relying on the format/order of "ls" for example, related to
> those variables)
>
> now with SLES9 we suddenly realized crashing printoutput done by
> (t)csh scripts (using lp ...) which only prints up to the character
> before the first german umlaut.
> This seems to be due to the /etc/csh.cshrc on that machine, which sets
> LANG and LC_* to de_DE.UTF-8 and "deletes" LC_ALL
>
> I know I could create ~/.cshrc with setenv LC_ALL C and LANG de_DE.
>
> I know I might even export CSHRCREAD=true in my login-ksh-environment
> to prevent csh from resetting those (but maybe others as well?)
>
> Are there any additional suggestions to get the old behaviour?
>
> Furthermore of course I want to understand WHY these settings are made
> now with SLES9 to determine if we will have to think about those
> things sooner or later again and have will have change the application
> anyway (and in which way?) which will be a great effort (additionally
> testing a lot of incoming/outgoing data-interfaces for example which
> all depend on *LANG* and other settings as well ...).
>
> Thanks for any answers related to the topic,
> bine

In my opinion, the distributors of SLES want everyone to use UTF-8, but
they can't force you. If you aren't ready to convert all your files, you
can change /etc/csh.cshrc back the way it was (and do a web search for
"Csh Programming Considered Harmful", but that's another story).

Re: incomplete printout ending on german umlauts with SLES9 due to

I'm still not sure what to do best.
Any ideas how I could bring the printer to just accept these
characters?
That is: could I use some options to combine with "lp" just to print
those lists which the application has generated while using UTF-8
(when called from csh, which is most of the time).
If SLES* comes that way anyway in the future it might be better to
live with this than always resetting LC_ALL=C ?
thanks if there's still anybody listening ;-)