#1393: t/compilers/pge/pge_examples.t: PASS by itself, FAIL during 'make smoke'
---------------------+------------------------------------------------------
Reporter: jkeenan | Owner:
Type: bug | Status: new
Priority: normal | Milestone:
Component: testing | Version: 1.9.0
Severity: medium | Keywords:
Lang: | Patch:
Platform: |
---------------------+------------------------------------------------------
Comment(by mikehh):
I agree completely that we have a problem here. I was merely indicating
that the problem did not occur with various options, not saying that we
need not fix it because we get it to work under certain circumstances.
I normally build with gcc using the following configure options:
perl Configure.pl --optimize --test --maintainer --configure_trace
and try with and without the --optimize.
On Ubuntu 9.10 i386 we are having a problem with perl Configure.pl --test
--maintainer --configure_trace
I ran a series of builds using different variations of this, at r43380,
after a make realclean each time.
The --maintainer option appears to give different results (with or without
--test and --configure_trace).
With the --maintainer option I get t/compilers/pge/pge_examples.t failing
test 2 as covered in this ticket. prove passes and perl t/harness fails
with the --gc-debug option but passes without.
If I drop the --maintainer option on the build the test passes, however
the test t/library/test_more.t fails after test 60 (see TT #473, comments
by jkeenan).
this test fails prove and perl t/harness with or without --gc-debug
but t/compilers/pge/pge_examples.t passes.
I tried this quite a few times and got the same results.
In both situations the test fails with:
src/gc/api.c:245: failed assertion 'PObj_is_PMC_TEST(obj)'
this is part of the function Parrot_gc_mark_PMC_alive_fun
and seems to be failing the assertion that it is an object.
Why this should happen and why the tests are ok in different builds and
not others is what needs investigation and resolution.
I don't have any ideas at the moment but will continue investigating.
Hopefully we can get others (Whiteknight, bacek, chromatic perhaps) to
have a look.
--
Ticket URL: <https://trac.parrot.org/parrot/ticket/1393#comment:15>
Parrot <https://trac.parrot.org/parrot/>
Parrot Development
_______________________________________________
parrot-tickets mailing list
[email protected]
http://lists.parrot.org/mailman/listinfo/parrot-tickets
