After trying different values for the TIMEOUT, I have found that the 300 second is partly to blame for the running out of resources on OS X. Apparently the regression tests spin off a bunch of processes/threads that time out after 300 seconds. This eventually leads to the resources depletion. Lowering this value to 10 allows me to run the regression tests to completion (a larger number might also work).
Is 300 seconds a resonable value? Obviously changing the timeout will not fix the problem, but at least it won't run out of resource or take hours to run test that are going to fail. On Sat, 04 Sep 2004 10:53:50 +0200, Dalibor Topic <[EMAIL PROTECTED]> wrote: > Michael Franz wrote: > > Taking a closer look at the failures (and reading Ricardo's sparc > > failures) darwin seems to be having a similar failure. I have 119 > > failure files and all of them have > > > > Timeout - 300 seconds - killed > > > > Is there a way to bump up the timeout? The intrp does run slow and > > the box I am using is not that fast (550MHz P3). > > You can bump up the timeout to a different value by setting the TIMEOUT > environment variable. I'd recommend testing whether the hanging tests > terminate at all, though. You might have just run into some threading > issues ;) > > >> > >>Just a note to let everyone know that kaffe darwin/x86 actually runs, > >>although not well. May not be great news but it makes me happy! > >> > >>Michael > >> > > > > Still happy it is running! > > > > Me too! > > cheers, > dalibor topic > _______________________________________________ kaffe mailing list [EMAIL PROTECTED] http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
