On Thu, Jan 25, 2001 at 02:29:21AM -0800, Greg Stein wrote:
> On Thu, Jan 25, 2001 at 03:41:12AM -0000, firstname.lastname@example.org wrote:
> > --- Makefile.am 2000/12/22 21:25:38 1.14
> > +++ Makefile.am 2001/01/25 03:41:07 1.15
> > @@ -12,6 +12,12 @@
> > ACLOCAL = @ACLOCAL@ -I ac-helpers
> > +## Note, we have to add these in so that they get included in the distribution
> > +EXTRA_DIST = ac-helpers/berkeley-db.m4 ac-helpers/config.guess \
> > + ac-helpers/config.sub ac-helpers/install-sh ac-helpers/ltconfig \
> > + ac-helpers/ltmain.sh ac-helpers/missing ac-helpers/mkinstalldirs \
> > + ac-helpers/svn-apache.m4 ac-helpers/svn-macros.m4
> We distribute an already-built configure. We don't need the .m4 files.
Good point, I just noticed that some of the files were missing, so I
added all of the ones from that directory.
> > +SVN_CLIENT_INCLUDES='-I$(top_srcdir)/subversion/client'
> > +AC_SUBST(SVN_CLIENT_INCLUDES)
> > +SVN_LIBSVN_CLIENT_INCLUDES='-I$(top_srcdir)/subversion/libsvn_client'
> > +AC_SUBST(SVN_LIBSVN_CLIENT_INCLUDES)
> > +SVN_LIBSVN_DELTA_INCLUDES='-I$(top_srcdir)/subversion/libsvn_delta'
> > +AC_SUBST(SVN_LIBSVN_DELTA_INCLUDES)
> > +SVN_LIBSVN_FS_INCLUDES='-I$(top_srcdir)/subversion/libsvn_fs'
> > +AC_SUBST(SVN_LIBSVN_FS_INCLUDES)
> > +SVN_LIBSVN_RA_DAV_INCLUDES='-I$(top_srcdir)/subversion/libsvn_ra_dav'
> > +AC_SUBST(SVN_LIBSVN_RA_DAV_INCLUDES)
> > +SVN_LIBSVN_SUBR_INCLUDES='-I$(top_srcdir)/subversion/libsvn_subr'
> > +AC_SUBST(SVN_LIBSVN_SUBR_INCLUDES)
> > +SVN_LIBSVN_WC_INCLUDES='-I$(top_srcdir)/subversion/libsvn_wc'
> > +AC_SUBST(SVN_LIBSVN_WC_INCLUDES)
> > +SVN_MOD_DAV_SVN_INCLUDES='-I$(top_srcdir)/subversion/mod_dav_svn'
> > +AC_SUBST(SVN_MOD_DAV_SVN_INCLUDES)
> > +SVN_TESTS_COMMON_INCLUDES='-I$(top_srcdir)/subversion/tests-common'
> > +AC_SUBST(SVN_TESTS_COMMON_INCLUDES)
> None of those directories are supposed to publish include files. I think
> that we should avoid the extra work (and the resulting sed substitution time
> caused by defining them). Defining these might also give the impression
> they're supposed to be published.
I was thinking about this yesterday as I was working on the patch. The
reason I included them is that several of the test suites rely on
non-published includes from the library they are testing. I guess we
could either remove those dependencies, or use -I@srcdir@/.. in the
Makefiles instead. I just thought I would put them in since they seem
to be used, but I agree that they really don't belong.
Let me know what you all think about it.
> Greg Stein, http://www.lyra.org/
Received on Sat Oct 21 14:36:19 2006
- application/pgp-signature attachment: stored