this isn't directly relevant to the discussion at hand, but more about timeouts in general.
one problem that timeouts can have is with a broken clock, the timeout will never fire. i don't think that atf can handle this itself, it needs to be in the testcase. there was one src/regress test case that i had patched to have a hard limit on iterations to avoid this, but that change got lost. of course, this case should be considered a failure, but to me the important part was to avoid the testsuite itself hanging forever. .mrg.
