Re: CVS Head - patch to remove extraneous code in temp_rpath calcs.

[Ralf Wildenhues]

* Bob Friesenhahn wrote on Mon, Nov 29, 2004 at 03:08:45PM CET:
> On Mon, 29 Nov 2004, Gary V. Vaughan wrote:
> >We can simplify more than that now:
> >
> >>--- config/ltmain.m4sh    28 Nov 2004 19:55:40 -0000    1.27
> >>+++ config/ltmain.m4sh    29 Nov 2004 03:14:53 -0000
> >>@@ -3516,7 +3516,6 @@
> >>         if test -n "$shlibpath_var"; then
> >>           # Make sure the rpath contains only unique directories.
> >>           case "$temp_rpath " in
> >>-          *" $dir "*) ;;
> >>           *" $absdir "*) ;;
> >>           *) temp_rpath="$temp_rpath $absdir" ;;
> >>           esac
> >
> >This hunk can become:
> >
> >          if test -n "$shlibpath_var"; then
> >            # Make sure the rpath contains only unique directories.
> >            case ":$rpath:" in
> >              *":$absdir:"*) ;;
> >              *) rpath="$rpath$absdir:" ;;
> >            esac
> 
> The reason I didn't do this is that 'rpath' is used as a temporary 
> variable in a number of other places.  It seems to me that a variable 
> used to carry state throughout the script should have a consistent 
> meaning and should not be used as a temporary variable.  I do agree 
> that 'rpath' is a much better name than 'temp_rpath'.

Yes, but..

> We could change existing 'rpath' to 'temp_rpath' and change existing 
> 'temp_rpath' to 'rpath'. :-)

No.  Got burnt with the incomplete `dlsymsobj' renaming, don't want to
go through that pain again.

Simply use a new, good name, or maybe two new, good names.  And don't
produce any leftovers.  People will compare passages of branch-2-0 with
older and newer branches and kill you over this (if they find out).

I know you didn't mean this, I just needed to mention that it was
painful.  :-)

Regards,
Ralf