[Ecls-list] Random tester
Juan Jose Garcia Ripoll
jlr at mpq.mpg.de
Fri Jan 7 06:15:19 UTC 2005
could you briefly explain me how to run the random tester. I am
currently using "make random-test" and "make rt_1000_8", but I am not
sure whether this is the right way. Is there also a recommended size for
the tests, number of variables, etc. Finally, is there a way to print
not only the offending form, but also the type of error it produces:
whether it fails to compile, to run, produces a wrong result, etc... I
would like to see this during the realization of the test, not after
waiting for the 1000 or 10000 tests to finish.
Besides that, I have verified that the errors that one sees with ECL and
the random tester are not, as you suggested, due to memory leaks, but
due to things like wrong inline forms and type propagation failures. It
seems just that your random tester tends to cluster the tests: for
instance now you see a lot of failures due to (AREF ...) forms where the
array is inferred to have the wrong type, and short after comes a large
period without errors.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 361 bytes
Desc: not available
More information about the ecl-devel