Markus Wichitill wrote:Stas Bekman wrote:Randy Kobes wrote:For the protocol/echo_nonblock.t test, as I mentioned, the failure of the last test (which is expected to timeout) only occurred when running the full test suite. I'll try to narrow it down - one combination that does fail is perl t/TEST -v apr protocol in which test 3 in protocol/echo_nonblock receives "should timeout", rather than "TIMEUP".
For both of these tests, nothing was in the error log that could offer a clue.
OK, looking forward to the SMOKE's shortest sequence.
I'm seeing this issue, too, now, after having built from SVN the first time (probably just coincidence, since I haven't built often during the last few weeks). It's not always failing, but it's also failing if the test is run on its own:
Markus, any difference with this patch?
Index: t/protocol/echo_nonblock.t =================================================================== RCS file: /home/cvs/modperl-2.0/t/protocol/echo_nonblock.t,v retrieving revision 1.2 diff -u -r1.2 echo_nonblock.t --- t/protocol/echo_nonblock.t 28 Sep 2004 22:47:22 -0000 1.2 +++ t/protocol/echo_nonblock.t 21 Nov 2004 16:53:11 -0000 @@ -21,6 +21,7 @@
# now get a timed out request $expected = "TIMEUP"; +sleep 2; print $socket "should timeout\n"; chomp($received = <$socket> || ''); ok t_cmp $received, $expected, "timed out";
If it works, try to lower it to 1 sec if possible.
With "sleep 1", I didn't get any failures running "perl t\TEST -times=100 t\protocol\echo_nonblock.t" and running the whole test suite three times.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
