On Fri, Jan 23, 2009 at 20:31, Will Coleda <[email protected]> wrote: > On Fri, Jan 23, 2009 at 11:30 PM, jerry gay <[email protected]> wrote: >> On Fri, Jan 23, 2009 at 20:23, Will Coleda <[email protected]> wrote: >>> On Fri, Jan 23, 2009 at 11:07 AM, Jonathan Worthington >>> <[email protected]> wrote: >>>> Will Coleda wrote: >>>>> >>>>> Why we can't just run this test via t/harness? >>>>> >>>>> >>>> >>>> That's exactly what the patch does: >>>> >>>>>> + $(PERL) t/harness $(EXTRA_TEST_ARGS) -r t/op/annotate.t >>>>>> >>>> >>>> Or am I missing the point? But just feel free, to commit whatever better >>>> way >>>> you want - the aim is just to run this test also in >>>> compile-to-bytecode-first mode on make test, as well as in the normal "just >>>> run it" way. >>>> >>>> Thanks, >>>> >>>> Jonathan >>>> >>>> >>> >>> Ok, so the -r makes it hard to run as part of the single invocation of >>> t/harness that we do with 'make test' >>> >>> Why do we have to run this one test with -r when we run 'make test'? >>> Why not catch it with 'make testr' which is run as part of 'make >>> fulltest'? >>> >> because nobody runs make fulltest. the smokers don't. the users don't. >> the developers don't. only the release manager does, once or twice >> before a release. i'm not saying that's good, that's just the way it >> is. >> >> ~jerry >> > > What makes /this test/ so important we have to run it here? > this test is for bytecode annotations. we need to make sure that they're frozen to bytecode properly, and when they're loaded, that they're correct. the -r runcore does just that--it freezes the code to bytecode on disk, then loads it, and runs it. so this test is run once without -r, which tests that bytecode annotations work in the interpreter, and once with -r, to make sure they work in frozen bytecode.
it's possible we should be testing other parrot features with the -r core more regularly, but i haven't thought about that enough to suggest we run always run the testr make target when we run test. ~jerry _______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
