cleanup should be optional, and it's only really necessary if you're
worried about disk space. So in this case the lack of cleanup has caught a
bug in the testsuite, which is good.
actually, the bug in the testsuite was caused by faulty cleanup
in the first place, and the lack of cleanup on that machine only
hid the real bug!-)
my ghci025.script is referring to ../prog001/D.hs, but that is
a generated file (copied from either D1/D2), so it shouldn't
exist after cleanup. but it is left there, and i didn't notice it
is a ghost because validate on my or your machine showed
no problem. it would be nice if D and other generated
sources had a comment stating their generated status.
it was rather confusing to see all that variation
- validate runs clean
- some buildbots (eg, windows head) report:
module 'D' is not interpreted
because the compiled version is left
- some buildbots report:
can't find file: ../prog001/D.hs
because the source does not exist
since the 'import D' is hard-coded into C, i'll have to check
whether i can still use that prog001 directory, or whether i'll
have to find another collection of examples.
(also, I suspect the cleanup support is far from complete,
and it needs a lot of work to be multi-thread-safe)
if cleanup was consistent, that would help!-)
claus
_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
