That sounds sufficiently "trivial" to me... :-) --Chris
Doug Hunt wrote: > Hi Chris: I understand your reluctance to make big build-related > changes now. > > On the other hand, the current 'make test' behavior is quite > obnoxious--it repeats the whole 'make' step, then runs the test and > *then* rebuilds some subdirectories in an effort to test the subdirs. > > I think the current 'make test' behavior could well put new users > off--you can't even tell if the tests passed because it scrolls many > pages beyond the success message. > > I do think a major rework of the build system is in order, perhaps > using Module::Build (but not now). > > My fix just re-wires the 'make test' step to: > > 1) Not do a 'make' first, and > 2) Not try to run subdirs tests, just the top-level test > > It is not perfect, but it makes things much nicer and quicker! > > Regards, > > Doug > > [EMAIL PROTECTED] > Software Engineer III > UCAR - COSMIC, Tel. (303) 497-2611 > > On Mon, 20 Oct 2008, Chris Marshall wrote: > >> Hi Doug, >> >> I'm not that familiar with subtleties of EU::MM so I can't help >> you there. I would vote against any non-trivial fixes right now: >> >> (1) PDL-2.4.4 is in "pre-release" this week, debugging this >> fully would slow down the formal release >> >> (2) it appears to work correctly---just v-e-r-y s-l-o-w-l-y >> >> (3) Module::Build work might be a better time investment >> >> Just my 3-cents, >> Chris >> >> Doug Hunt wrote: >>> Hi all: I'm looking into what is causing the inconvenient 'multiple >>> makes' that have happened since Craig DeForest's 'circular >>> dependency' fix of July 9. Craig's fix seems to be necessary--I >>> think the problem started with a MakeMaker change. >>> >>> Does anyone know what sub MY::test is for in top-level Makefile.PL? >>> >>> It re-writes the 'test' section generated by MakeMaker, but on my >>> linux box it does not change a thing, leaving the original 'test' >>> section as it would have been. I'm thinking I may have to fiddle >>> with this to get >>> the desired behavior: >>> >>> make -- builds PDL >>> make test -- tests PDL (and nothing more!) >>> >>> It seems to me that the problem may have to do with this Makefile >>> clause: >>> >>> test_dynamic :: pure_all >>> PERL_DL_NONLAZY=1 $(FULLPERLRUN) "-MExtUtils::Command::MM" >>> "-e" "test_harness($(TEST_VERBOSE), '$(INST_LIB)', '$(INST_ARC\ >>> HLIB)')" $(TEST_FILES) >>> >>> I'd like to see something more like: >>> >>> test_dynamic :: >>> PERL_DL_NONLAZY=1 $(FULLPERLRUN) "-MExtUtils::Command::MM" >>> "-e" "test_harness($(TEST_VERBOSE), '$(INST_LIB)', '$(INST_ARC\ >>> HLIB)')" $(TEST_FILES) >>> >>> Any thoughts? >>> >>> Would anyone mind if I got rid of the current sub MY::test in the >>> top level Makefile.PL? >>> >>> Regards, >>> >>> Doug >>> >>> [EMAIL PROTECTED] >>> Software Engineer III >>> UCAR - COSMIC, Tel. (303) 497-2611 >>> >>> _______________________________________________ >>> Perldl mailing list >>> [email protected] >>> http://mailman.jach.hawaii.edu/mailman/listinfo/perldl >>> >>> >>> > > _______________________________________________ Perldl mailing list [email protected] http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
