Ticket #116 (closed bug: fixed)
intermittent failure of t/pmc/freeze_25.pir
Reported by: | Infinoid | Owned by: | |
---|---|---|---|
Priority: | normal | Milestone: | |
Component: | none | Version: | |
Severity: | medium | Keywords: | |
Cc: | Language: | ||
Patch status: | Platform: |
Description
I am getting an intermittant failure in this test, and others in #parrot have confirmed it. The failure is in a Hash PMC, which, after being frozen/thawed, in some cases does not return the data that was put into the original hash.
The test is supposed to output two lines of similar numbers, as follows: 1 2 3 6 1 2 3 6
There are apparently 3 possible outcomes. The first is that the hash returns the same data that was put into it, and thus the second line of output is "1 2 3 6". This happens about 80% of the time. The second and third outcomes are failures, which look like:
1 2 3 6 3 6 Null PMC access in get_integer_keyed() current instr.: 'test' pc 125 (t/pmc/freeze_25.pir:45) called from Sub 'main' pc 42 (t/pmc/freeze_25.pir:13)
and
1 2 3 6 Null PMC access in get_integer_keyed() current instr.: 'test' pc 105 (t/pmc/freeze_25.pir:38) called from Sub 'main' pc 42 (t/pmc/freeze_25.pir:13)
(the difference being whether the "3 6" is printed or not.)
I have bisected this issue, and it was introduced in r34127.