t/src/extend_vtable.t: test failures on Darwin/PPC with --gc=ms2

[jkeenan]

As reported in TT #2097, Parrot does not currently build for me on Darwin/PPC in 'master', now that the default GC has been switched to gsm. So I tried to build with what was formerly the default: --gc=ms2. The build was successful, but I got new test failures in t/src/extend_vtable.t -- a file whose failures on this platform and others we had just cleaned up in TT #2084!

[parrot] 509 $ git show | head -1
commit 61176bb71505dd4d2a5d9c53adc1ba139698c0b2
[parrot] 511 $ prove -v t/src/extend_vtable.t 
t/src/extend_vtable.t .. 
1..100
ok 1 - Parrot_PMC_get_repr
...
not ok 7 - Parrot_PMC_get_number_keyed_str

#   Failed test 'Parrot_PMC_get_number_keyed_str'
#   at t/src/extend_vtable.t line 197.
# Exited with error code: [SIGNAL 6]
# Received:
# src/hash.c:1704: failed assertion 'key'
# 
# Expected:
# 0.000000
# Done!
# 
not ok 8 - Parrot_PMC_get_integer_keyed_str

#   Failed test 'Parrot_PMC_get_integer_keyed_str'
#   at t/src/extend_vtable.t line 205.
# Exited with error code: [SIGNAL 6]
# Received:
# src/hash.c:1704: failed assertion 'key'
# 
# Expected:
# 0
# Done!
# 
not ok 9 - Parrot_PMC_exists_keyed_str

#   Failed test 'Parrot_PMC_exists_keyed_str'
#   at t/src/extend_vtable.t line 213.
# Exited with error code: [SIGNAL 6]
# Received:
# src/hash.c:1704: failed assertion 'key'
# 
# Expected:
# 0
# Done!
# 
ok 10 - Parrot_PMC_delete_keyed
...
ok 22 - Parrot_PMC_(set|get)_integer_keyed_int
not ok 23 - Parrot_PMC_delete_keyed_str

#   Failed test 'Parrot_PMC_delete_keyed_str'
#   at t/src/extend_vtable.t line 343.
# Exited with error code: [SIGNAL 6]
# Received:
# src/hash.c:1704: failed assertion 'key'
# 
# Expected:
# Done!
# 
ok 24 - Parrot_PMC_get_string_keyed_str
...
ok 100 - Parrot_PMC_assign_pmc
# Looks like you failed 4 tests of 100.
Dubious, test returned 4 (wstat 1024, 0x400)
Failed 4/100 subtests 

Test Summary Report
-------------------
t/src/extend_vtable.t (Wstat: 1024 Tests: 100 Failed: 4)
  Failed tests:  7-9, 23
  Non-zero exit status: 4
Files=1, Tests=100, 131 wallclock secs 
  ( 0.24 usr  0.08 sys + 23.03 cusr 24.85 csys = 48.20 CPU)
Result: FAIL

[jkeenan]

As noted in TT #2097, when I ran the test that failed above I had a patch in position for src/platform/darwin/sysmem.c. The patch used HW_MEMSIZE instead of HW_PHYSMEM. When I reverted sysmem.c to master and re-tested, I got exactly the same failures. So the failures cannot be attributed to that patch (which we're considering for adoption in TT #2090).

Thank you very much.

kid51

[jkeenan]

Replying to jkeenan:

As reported in TT #2097, Parrot does not currently build for me on Darwin/PPC in 'master', now that the default GC has been switched to gsm.

That situation changed, so TT #2097 is now closed. But I still get the 4 test failures in t/src/extend_vtable.t reported in this ticket.

Thank you very much.

kid51

[dukeleto]

This should be fixed in 75e32d3 . If that works for you, please close this ticket.

[jkeenan]

Replying to dukeleto:

This should be fixed in 75e32d3 . If that works for you, please close this ticket.

duke,

It appears commit 75e32d3 was in your leto/embed_grant branch. Once I figured that out, I diffed the two branches and applied the resulting patch to master.

I then re-ran the tests ... only to discover that while the 4 failing tests now passed, one new test did not:

not ok 30 - Parrot_PMC_defined_keyed_str

#   Failed test 'Parrot_PMC_defined_keyed_str'
#   at t/src/extend_vtable.t line 403.
# Exited with error code: [SIGNAL 6]
# Received:
# src/hash.c:1704: failed assertion 'key'
# 
# Expected:
# 0
# Done!
# 

I then added one line to that test in the same way that you had added it to other tests. That finally enabled all tests to PASS on Darwin/PPC. See 7413e3082f.

Thanks for keeping plugging away at these tests. Closing.

kid51