This one time, at band camp, Niko Tyni said: > I'm a bit uneasy of testing this out on agnesi.d.o. Could somebody on > the debian-arm list look into it a bit to see how often it fails and how? > > Just running the test (attached for convenience) with the system perl > on lenny or sid should do, no need for the full build tree. > > Lowering the severity to 'important' for now because we have seen ten > or so successful arm builds in in unstable and just one failure. I have not yet run it on an arm machine to check, but one thing occurs to me immediately. This test thinks it is testing how well threads work, but a side effect of it's algorithm is that it actually tests how quickly threads can do loops of up to a million times. This may just be too much for arm - running 50 simultaneous threads doing a million loops may take more than the 30 seconds allowed on some of the older boxes. If you want to test that the threads succeeded, you could up the timeout (still racy) or change from cond_timedwait to cond_wait (of course, that can lock up a build if something goes wrong). Cheers, -- ----------------------------------------------------------------- | ,''`. Stephen Gran | | : :' : sgran@debian.org | | `. `' Debian user, admin, and developer | | `- http://www.debian.org | -----------------------------------------------------------------
Attachment:
signature.asc
Description: Digital signature