
On 24 Jan 2018, at 18:06, Attila Lendvai wrote:
My iterate directory was missing some patches. With the latest version from darcs 5 tests fail:
5 out of 268 total tests failed: ITERATE.TEST::ALWAYS.FINALLY, ITERATE.TEST::NEVER.FINALLY, ITERATE.TEST::THEREIS.FINALLY, ITERATE.TEST::BUG/WALK.2, ITERATE.TEST::BUG/COLLECT-AT-BEGINNING.
yep, this is what i see here, too:
5 out of 268 total tests failed: ITERATE.TEST::ALWAYS.FINALLY, ITERATE.TEST::NEVER.FINALLY, ITERATE.TEST::THEREIS.FINALLY, ITERATE.TEST::BUG/WALK.2, ITERATE.TEST::BUG/COLLECT-AT-BEGINNING.
unfortunately iterate's test suite is... well, less than ideal. (sb-)rt is annoyingly cumbersome and just simply archaic...
hu.dwim.reiterate has an iterate-compat package that is meant to be a drop-in replacement for iterate. it uses a proper code walker and some nice features (like retaining slime's source-location feature within its body).
unfortunately it's only around 80% completed and i didn't work on it for quite a while now.
and i also don't feel entitled to do a major refactor of iterate's test suite to e.g. use hu.dwim.stefil.
sorry for the mess,
Didn't mean to be complaining. One possibility would be to use the set of expected failures in the definition of the ASDF test-operation, and raise an exception only for unexpected failures. I think we have to hack exceptions into the TEST-OP if you want to be able to easily check tests, since ASDF:OPERATION doesn't return anything (and really, it can't). I'll have a look, but unfortunately, I'm drowning in Real Work(TM) right now. I should look at what's going awry on ACL, first, too, since that seems more critical. Anyone testing iterate on CCL or any of the other implementations? Cheers, r