On Sunday 18 March 2007 10:06, Jerry Gay wrote: > at r17619, i'm unable to run make test, due to failures in p5rx tests. > actually, none of them are failing, but my system becomes > non-responsive. after approximately test 200, tests (and there are > over 900) take a few seconds each to complete, and each one uses more > and more memory. at over 125MB, and 100% cpu, i can do little but kill > the process. this means that the harness dies, and i can't complete > 'make test'. > > i suspect a string gc problem.
A GC problem might explain the STM test failure too. Do you see that? If so, is it possible to do a binary search to figure out which checkin caused (or revealed) these failures? -- c
