From @moritz

in r31436 the test t/spec/S29-conversions/ord_and_chr.t fails on my
computer when called through make (ie 'make spectest_regression' or
'make t/spec/S29-conversions/ord_and_chr.t', but succeeds if the
corresponding .rakudo file is called manually, or though
tools/test_summary.pl.

output through make​:

$ TEST_JOBS=1 make t/spec/S29-conversions/ord_and_chr.t
t/spec/S29-conversions/ord_and_chr...........
1..240
ok 1 - ord() works for \32 == ' '
ok 2 - chr() works for \32 == ' '
...
ok 172 - chr() works for \117 == 'u'
ok 173 - ord() works for \118 == 'v'
Failed 67/240 subtests

Test Summary Report

t/spec/S29-conversions/ord_and_chr.rakudo (Wstat​: 11 Tests​: 173 Failed​: 0)
  Parse errors​: Bad plan. You planned 240 tests but ran 173.
Files=1, Tests=173, 3 wallclock secs ( 0.03 usr 0.00 sys + 2.55 cusr
0.04 csys = 2.62 CPU)
Result​: FAIL

When executed directly it passes all 240 tests (minus two skipped
tests), and everything is fine.

This is on Debian GNU/Linux "Etch", i386 32bit.

Moritz

--
Moritz Lenz
http://moritz.faui2k3.org/ | http://perl-6.de/

From @moritz

A few more details​:

This happens with both the system perl (5.8.8) and a vanilla 5.10.0
perl, and is also reproducible from a clean 'svn export' copy, and
without ccache (which I usually use for compiling parrot).

On another 32bit Debian system (with Testing this time) everything works
fine.

Still no clue what's going on...

The RT System itself - Status changed from 'new' to 'open'

From @moritz

On Fri Sep 26 09​:05​:32 2008, moritz@​casella.verplant.org wrote​:

in r31436 the test t/spec/S29-conversions/ord_and_chr.t fails on my
computer when called through make (ie 'make spectest_regression' or
'make t/spec/S29-conversions/ord_and_chr.t', but succeeds if the
corresponding .rakudo file is called manually, or though
tools/test_summary.pl.

weird things, today all is fine, no failures any more. Marking this as
rejected since I can't reproduce it anymore.

@moritz - Status changed from 'open' to 'rejected'

From @pmichaud

On Fri, Sep 26, 2008 at 09​:05​:32AM -0700, Moritz Lenz wrote​:

in r31436 the test t/spec/S29-conversions/ord_and_chr.t fails on my
computer when called through make (ie 'make spectest_regression' or
'make t/spec/S29-conversions/ord_and_chr.t', but succeeds if the
corresponding .rakudo file is called manually, or though
tools/test_summary.pl.

This has to be a problem with the test harness or Parrot somehow --
we've noted many times in the past that running things from the
harness sometimes causes things to fail that otherwise succeed.

At this point I'm not sure if this ticket belongs in the perl6
RT queue or the parrot RT queue. There's not much that Rakudo
can do to resolve the issue, so I'm thinking we may need to move
it to the Parrot queue.

Pm

From @moritz

re-opened because others (rindolf) experienced the same error.

@moritz - Status changed from 'rejected' to 'open'

From @moritz

We haven't seen a manifestation of that bug for quite some time; hope
it's gone...

@moritz - Status changed from 'open' to 'resolved'