On Jul 10, 2014, at 11:39 AM, Ben Pfaff <b...@nicira.com> wrote: > On Mon, Jun 09, 2014 at 11:13:52AM -0700, Ben Pfaff wrote: >> On Thu, May 29, 2014 at 05:16:09PM -0700, Jarno Rajahalme wrote: >>> (snip) >>> I made a deliberate core-producing error to test this out. As a result I >>> have three core files: >>> >>> $ ls -l tests/testsuite.dir/*/core >>> -rw------- 1 jrajahalme jrajahalme 20344832 May 29 16:53 >>> tests/testsuite.dir/0352/core >>> -rw------- 1 jrajahalme jrajahalme 20353024 May 29 16:53 >>> tests/testsuite.dir/0353/core >>> -rw------- 1 jrajahalme jrajahalme 20357120 May 29 16:53 >>> tests/testsuite.dir/0764/core >>> >>> But I don?t seem to find the string ?core dumped? from the make output, nor >>> from testsuite.log. Where should it be? >> >> I believe that this is not being reported because the test failed >> before we got to the point of the log scan or checking for cores. >> That is OK: the cores will be found when the developer investigates >> the test failure. We only need to report any cores that appear when >> the test would otherwise pass. > > Any further thoughts on that?
When I run multiple tests in parallel, it is typical that some tests fail, but when I run —recheck to check them again one at a time. My concern is that even if that succeeds, it is possible that there was a core, which would indicate a more serious problem than just a simple test script race. Therefore it would be nice to have a clearly visible note/warning if any cores were produced. Alternatively, maybe —recheck should decline rerunning a test that produced a core. (I just recently learned about ‘—recheck’, I previously re-run the failed tests manually. So I don’t know if ‘—recheck’ already checks for cores or not.) Jarno _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
