At Sunday 06 June 2010, Ralf Wildenhues <[email protected]> wrote: > * Stefano Lattarini wrote on Sun, Jun 06, 2010 at 07:50:57PM CEST: > > At Sunday 06 June 2010, Ralf Wildenhues wrote: > > > Thanks, and sorry for not having rerun the complete testsuite > > > before pushing the original patch. > > > > Well, I should say "sorry" for not having tested the patch > > thoroughly. > > The patch is still not right, so I'm not pushing it. A generated > test foo-p.test now needs to depend upon foo.test, > this is not reflected in the makefile. Right :-(
Not however that the older makefile didn't have a much better handling of tests' dependencies. Maybe the pending patch of mine "Improve declaration of dependencies in the testsuite" might help here, if properly extended. For reference: <http://lists.gnu.org/archive/html/automake-patches/2010-05/msg00012.html> > Either we rewrite defs.in so that it is idempotent and sourced > twice, Mhh... that would entail a rethinking of the whole 'required=...' stuff. Which wouldn't be bad in the long run, but we should think thoroughly before venturing down that road, to get the interface right this time. > or sets only the needed variables the first time it is sourced. I see no easy and natural way to do this. > Alternatively, it seems viable to just revert this patch > mini series un-aply titled "bikeshedding" in the first place. ;-) Maybe we should do so.... But I'd like to keep the series as-is anyway, and then integrate it with an extended-and-amended patch "Improve declaration of dependencies in the testsuite". Since I think I've made my point clear, I left the final decision to you (hoping you'll agree with me ;-). Regards, Stefano
