On 11/17/11 6:55 AM, Vasily Chekalkin wrote:
Hello.
Thanks a lot. Can you try again? I decided to go with smaller steps.
This smoke report indicates that the kill_constants branch now builds
and tests successfully on linux/i386:
http://smolder.parrot.org/app/projects/report_details/24470
However, this smoke report indicates that the kill_constants branch
fails one more test on Darwin/PPC than does master:
http://smolder.parrot.org/app/projects/report_details/24471
That smolder shows failures in t/pmc/nci.t -- the one also found in
master -- and in t/pmc/task.t. For the latter, see:
http://smolder.parrot.org/app/projects/tap_stream/24471/230
Now, what makes this really annoying is that this is one of these
failures that I have *only* observed during 'make test'. When I run
prove -v t/pmc/task.t, all tests pass:
$ prove -v t/pmc/task.t
t/pmc/task.t ..
1..10
ok 1 - Task ran in order (0)
ok 2 - Implicit task ran in order (1)
ok 3 - Task ran in order (2)
ok 4 - Got message after block
ok 5 - Got existing message
ok 6 - task_to_kill running
ok 7 - task_to_kill killed
ok 8 - in wait_sub1
ok 9 - After wait
ok 10 - Pre-empt and exit
ok
All tests successful.
Files=1, Tests=10, 3 wallclock secs ( 0.13 usr 0.05 sys + 0.06 cusr
0.13 csys = 0.37 CPU)
Result: PASS
Same thing with perl t/pmc/task.t; perl t/harness t/pmc/task.t.
I know we have encountered this sort of thing in the past, but at the
moment I can't recall how we dealt with it.
Thank you very much.
Jim Keenan
_______________________________________________
http://lists.parrot.org/mailman/listinfo/parrot-dev
