We should get rid of most of these fine-grained targets. They add silly
complexity. extraclean-swig-py vs extraclean-swig-pl. Really?!
How about just: extraclean.
Seriously. It is generally a bad idea to "extraclean" just a portion of the
tree. You could throw away outputs from gen-make, autoconf, or whatever.
Bad idea. The whole tree should go back to pristine all at once.
I would argue much the same for distclean targets.
The plain clean targets are arguable, as you may want to blast and rebuild
a subsystem. But who really uses them? They made a lot of sense a decade
ago when a complete build could take 30 minutes. But on modern hardware,
the granularity just doesn't make sense.
On Mar 16, 2012 9:45 AM, <julianfoad_at_apache.org> wrote:
> Author: julianfoad
> Date: Fri Mar 16 13:44:36 2012
> New Revision: 1301511
> URL: http://svn.apache.org/viewvc?rev=1301511&view=rev
> * Makefile.in
> (extraclean-swig): Depend on clean-swig so that it does everything that
> it was omitting to remove '.swig_checked'.
> Modified: subversion/trunk/Makefile.in
> --- subversion/trunk/Makefile.in (original)
> +++ subversion/trunk/Makefile.in Fri Mar 16 13:44:36 2012
> @@ -736,7 +736,8 @@ extraclean-bindings: extraclean-swig ext
> clean-swig: clean-swig-headers clean-swig-py clean-swig-rb clean-swig-pl
> @rm -f .swig_checked
> -extraclean-swig: extraclean-swig-headers extraclean-swig-py \
> +extraclean-swig: clean-swig extraclean-swig-headers extraclean-swig-py \
> extraclean-swig-rb extraclean-swig-pl
> EXTRACLEAN_SWIG_HEADERS=rm -f $(SWIG_SRC_DIR)/proxy/*.swg
Received on 2012-03-16 16:20:53 CET