[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Unable to set LC_COLLATE system-wide



On Thu, 2008-06-05 at 13:08 +0100, John Horne wrote:
> On Tue, 2008-06-03 at 22:09 -0430, Patrick O'Callaghan wrote:
> > On Tue, 2008-06-03 at 22:58 +0100, John Horne wrote:
> > 
> > > The trouble is I do not know what the login sequence is when logging
> > > into an X window system. As such, I cannot say what is being run between
> > > the first run of /etc/profile and the second.
> > 
> > It's not only X itself, it's the whole desktop environment. In principle
> > you can follow the breadcrumbs via the man pages, but this line
> > (in /etc/profile etc.) will add more info to your debug comments and
> > might be useful:
> > 
> > echo This shell called from `ps -p $PPID -o comm=`, pid=$PPID >> /tmp/BASH_DEBUG
> > 
> Runlevel 3 shows 'login' being used; runlevel 5 shows that 'init'
> calls /etc/profile first, and then 'gdm-session-worker' calls it again.
> 
> > > 
> > > Since I'm a bit stumped as to where to go from here, but it definitely
> > > seems that something 'odd' is going on, I think perhaps this should go
> > > up to bugzilla?
> > 
> > Could be. It certainly doesn't seem to be doing what it says on the tin.
> > 
> Bugzilla https://bugzilla.redhat.com/show_bug.cgi?id=450052

Well, if it isn't one of those, it must be something else :-) i.e. any
process in the hierarchy from init down to your shell can unset any
environment variable if it wants to. It would be useful if one could
somehow set a trigger on an envariable to log when it was changed, but
there is no such general facility that I know of (aside from running
everything through a debugger, which could get pretty hairy).

poc


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]