FreeBSD 10.0-CURRENT/amd64: Weirdness with LOCALE settings: ghostswitching in csh?

[Jeremy Chadwick]

On Fri, Nov 04, 2011 at 07:49:52AM +0100, O. Hartmann wrote:
> Am 11/03/11 23:48, schrieb Jeremy Chadwick:
> > On Thu, Nov 03, 2011 at 11:17:08PM +0100, O. Hartmann wrote:
> >> Hello.
> >> I realised something weird in FreeBSD 10.-CURRENT/amd64 (CLANG
> >> compiled), build as from today (buildworld).
> >>
> >> Working the whole day coding some pyhton scripts and committing the code
> >> to my subversion server (most recent subversion from the ports
> >> collection, the server is a FreeBSD 9.0-RC1/amd64 box, also system
> >> compiled with CLANG, most recent as compiled world of today), suddenly,
> >> oy of the blue, trying again to commit I get this error:
> >>
> >> svn: warning: cannot set LC_CTYPE locale
> >> svn: warning: environment variable LC_CTYPE is de_DE.ISO-8859-1
> >> svn: warning: please check that your locale name is correct
> >>
> >>
> >> Checking  csh shell setting with 'locale":
> >> LANG=
> >> LC_CTYPE="C"
> >> LC_COLLATE="C"
> >> LC_TIME="C"
> >> LC_NUMERIC="C"
> >> LC_MONETARY="C"
> >> LC_MESSAGES="C"
> >> LC_ALL=
> >>
> >>
> >> Checking my settings from /etc/csh.cshrc and ./.cshrc or .login reveals
> >> localised settings for some of the locales as I need those:
> >>
> >> (set in $HOME/.cshrc)
> >> setenv  LC_CTYPE        "de_DE.ISO-8859-1"
> >> setenv  LC_TIME         "de_DE.ISO-8859-1"
> >> setenv  LC_MONETARY     "de_DE.ISO-8859-1"
> >>
> >> What is going on?
> >>
> >> I realised this behaviour now several times, first time I thought I did
> >> something and I couldn't remember, but this time, only two terminal
> >> windows were opened and the whole day committing data to the repository
> >> wasn't an issue.
> >>
> >> Is there an explanation for this?
> > 
> > It sounds like a problem specific to the "client end", meaning your
> > -CURRENT box.  If that's the case: shouldn't this mail have gone to
> > freebsd-current@ instead of freebsd-stable@ ?  What am I missing?
> 
> Mea culpa, mea culpa, mea maxima culpa!
> 
> It was intented to send the mail to CURRENT. Sorry, missed the listentry
> by one row ... Can you please so kind and show mercy?

No worries.  I wasn't sure if there was a reason -stable was
involved; I saw it and thought "Hmm, he mentions a 9.0-RC1/amd64 box,
maybe that's where the problem is?  I must be missing something", so I
thought I'd ask.  Mistakes happen, especially ones from me!  :-)

> > As for your problem: your locale looks incorrect.  It's
> > "de_DE.ISO8859-1".  Note that yours has an extra hyphen, which probably
> > explains the error (sort of).
> > 
> > $ ls -ld /usr/share/locale/de_DE*
> > drwxr-xr-x    2 root      wheel     512 Sep 28 14:36 /usr/share/locale/de_DE.ISO8859-1/
> > drwxr-xr-x    2 root      wheel     512 Sep 28 14:36 /usr/share/locale/de_DE.ISO8859-15/
> > drwxr-xr-x    2 root      wheel     512 Sep 28 14:36 /usr/share/locale/de_DE.UTF-8/
> > 
> > As for the fact that it's "random": I cannot explain why a sub-shell
> > might get spawned in some cases but not others.
> 
> I corrected this. Sorry. I ffel a bit confused, since sometimes it is
> ISO-8859-1 and sometimes ISO8859-1. I got confused again.
> 
> After correcting that, the locale variables has been set correctly.
> 
> I will check now wether this also influences this weird random behaviour.

I find the "randomness" of the situation more perplexing than fixing
your locale (I have a feeling you do too).  I imagine this will probably
fix the errors you were seeing, but I'm still surprised that the errors
would happen seemingly intermittently.

I wonder how one could go about debugging such a thing.  Hmm.  Are there
any "environment complexities" you might have, such as using GNU screen or
tmux?  I'm cycling through ideas as they come to me.

-- 
| Jeremy Chadwick                                jdc at parodius.com |
| Parodius Networking                       http://www.parodius.com/ |
| UNIX Systems Administrator                   Mountain View, CA, US |
| Making life hard for others since 1977.               PGP 4BD6C0CB |