t/pmc/resizablestringarray.t FAILS in make corevm/make coretest

[mikehh]

at r46158 on Ubuntu 10.04 amd64 building with g++

after make corevm/make coretest:

Test Summary Report
-------------------
t/pmc/resizablestringarray.t       (Wstat: 11 Tests: 211 Failed: 0)
  Non-zero wait status: 11
  Parse errors: Bad plan.  You planned 261 tests but ran 211.
Files=231, Tests=7253, 36 wallclock secs ( 3.43 usr  1.23 sys + 72.29 cusr 24.11 csys = 101.06 CPU)
Result: FAIL
make: *** [coretest] Error 1

prove t/pmc/resizablestringarray.t
t/pmc/resizablestringarray.t .. Failed 50/261 subtests 

Test Summary Report
-------------------
t/pmc/resizablestringarray.t (Wstat: 11 Tests: 211 Failed: 0)
  Non-zero wait status: 11
  Parse errors: Bad plan.  You planned 261 tests but ran 211.
Files=1, Tests=211,  0 wallclock secs ( 0.06 usr  0.02 sys +  0.11 cusr  0.01 csys =  0.20 CPU)
Result: FAIL

status 11 is a Segmentation fault

after then running make world/make test:

All tests successful.
Files=342, Tests=12510, 53 wallclock secs ( 5.94 usr  2.25 sys + 125.34 cusr 31.43 csys = 164.96 CPU)
Result: PASS

prove t/pmc/resizablestringarray.t
t/pmc/resizablestringarray.t .. ok       
All tests successful.
Files=1, Tests=261,  1 wallclock secs ( 0.14 usr  0.01 sys +  0.18 cusr  0.02 csys =  0.35 CPU)                                                                                                                                            
Result: PASS

I ran this a few times (makeclean and rebuild) and got the same result.

I extracted the two methods from the test before and after the failure and the test passes (after make corevm) with just those tests.

I got the same results building with g++ with --optimize and gcc with --optimize.

However, the test did not fail on the gcc build.

Looks to me like this is a problem with gc, but might be a little difficult to track down.

[mikehh]

The problem has not arisen again and seems to be fixed. Closing ticket.