On Wednesday 09 February 2011, Ralf Wildenhues wrote: > Hi Stefano, > > * Stefano Lattarini wrote on Sun, Feb 06, 2011 at 06:32:56PM CET: > > Hello Ralf, and sorry for the delay. > > No worries. I'm waay more behind. > > > On Thursday 03 February 2011, Ralf Wildenhues wrote: > > > * Stefano Lattarini wrote on Wed, Feb 02, 2011 at 12:10:15AM CET: > > > > A testsuite-enhancement patch stemmed from my brief foray into > > > > Automake's python support. This patch is in small part cosmetic, > > > > but IMHO offers real improvements and valuable additions, and > > > > also fixes a couple of glitches in python.m4. > > > > > > OK for master (branched off of maint, if you prefer) with nits > > > addressed. > > > I've now merged to patch (branched off of maint) into master, and pushed.
> > > Before pushing, please test on a system without a python interpreter > > > installed (you can rename you pythons temporarily). > > > > > Nah, IMHO is better to do something like: > > > > $ cd ~/src/automake/tests > > $ mkdir xbin && cd xbin > > $ for f in /bin/* /usr/bin/* /usr/local/bin/*; do > > > case $f in *python*);; *) ln -s $f .;; esac > > > done > > $ cd .. > > $ PATH=`pwd`/xbin make check TESTS='...' > > Neat; but also expensive. > Ah, but botching up system-wide python is much more expensive. And Murphy is always lurking ... ;-) > [CUT] > > > > --- a/tests/python5.test > > > > +++ b/tests/python5.test > > > > > > > @@ -24,16 +24,32 @@ set -e > > > > > > > > cat >>configure.in <<EOF > > > > # Hopefully the Python team will never release such a version. > > > > -AM_PATH_PYTHON(9999.9) > > > > +AM_PATH_PYTHON([9999.9]) > > > > > > Nice that you do it here, but up in python.m4 you should then, too. > > > > > Definitely. But in a follow-up patch IMHO (and since we are at it, we > > should fix underquoting in all the other *.m4 automake files). I will > > submit this patch in the coming week. > > OK. > Patch still on the todo list ... > > Attached is what I've squashed into the previous version of the patch. > > > > I will push in 72 hours if there are no further objections. > > OK but I have a question below. > > > --- a/tests/instdir-ltlib.test > > +++ b/tests/instdir-ltlib.test > > > @@ -14,7 +14,8 @@ > > # You should have received a copy of the GNU General Public License > > # along with this program. If not, see <http://www.gnu.org/licenses/>. > > > > -# If $(libdir) is the empty string, then nothing should be installed there. > > +# If $(libdir) or $(pyexecdir) is the empty string, then nothing should > > +# be installed there. > > # This test exercises the libtool code paths. > > > > required=libtoolize > > @@ -26,6 +27,7 @@ cat >>configure.in <<'END' > > AC_PROG_CC > > AM_PROG_CC_C_O > > AC_PROG_LIBTOOL > > +AM_PATH_PYTHON > > AC_OUTPUT > > END > > > > > --- a/tests/instdir-prog.test > > +++ b/tests/instdir-prog.test > > > @@ -14,7 +14,8 @@ > > # You should have received a copy of the GNU General Public License > > # along with this program. If not, see <http://www.gnu.org/licenses/>. > > > > -# If $(bindir) is the empty string, then nothing should be installed there. > > +# If $(bindir), $(libdir) or $(pyexecdir) is the empty string, then > > +# nothing should be installed there. > > # This test exercises the prog and libs code paths. > > > > . ./defs || Exit 1 > > @@ -25,6 +26,7 @@ cat >>configure.in <<'END' > > AC_PROG_CC > > AM_PROG_CC_C_O > > AC_PROG_RANLIB > > +AM_PATH_PYTHON > > AC_OUTPUT > > END > > > > Do these two tests require python now? > No, since I define the relevant configure cache values explicitly, and also force PYTHON to `echo'. Also, making such important and generic tests depend on python would have been a royally bad idea. > In that case maybe the merging > wasn't such a good idea after all; several of the machines I > test on do not have python in the default PATH. > In light of the above, this shouldn't be a problem. Regards, Stefano
