Bug #2244 for Test-Unit: one test failure on Solaris 7, perl 5.8.0

Mon Mar 17 08:55:21 2003 Guest - Ticket created

Subject:

one test failure on Solaris 7, perl 5.8.0

I get a test failure in try_examples, related to examples/fail_example: The output differs VERY slightly (pseudo-diff syntax below ;-): < real output Show quoted text

> expected putput

< # ..FSuite teardown Show quoted text

> # .F.Suite teardown

I'm not sure whether this indicates a problem or not - let me know if you need more info. All other tests run ok. I got all the latest&greatest modules from CPAN (i.e. the "r" operation of the CPAN modules does not show any new modules available). My perl is 5.8.0 with ithreads and a shared libperl, compiled for Solaris 7 (Sparc) with Forte 6.2 Cheers, Marek

Mon Mar 17 11:19:14 2003 cpan [...] adamspiers.org - Correspondence added

[guest - Mon Mar 17 08:55:21 2003]: Show quoted text

> I get a test failure in try_examples, related to > examples/fail_example: > The output differs VERY slightly (pseudo-diff syntax below ;-): > < real output

> > expected putput

> < # ..FSuite teardown

> > # .F.Suite teardown

> > I'm not sure whether this indicates a problem or not - let me > know if you need more info. All other tests run ok. I got all the > latest&greatest modules from CPAN (i.e. the "r" operation of the > CPAN modules does not show any new modules available). > My perl is 5.8.0 with ithreads and a shared libperl, compiled for > Solaris 7 (Sparc) with Forte 6.2

Off the top of my head, it's probably just because the order the tests are run in is not deterministic. Should be easy to fix when I get a moment (no guarantees when though).

Wed Mar 31 06:27:59 2004 Guest - Correspondence added

From:

aa29 [...] mail.ru

Show quoted text

> Off the top of my head, it's probably just because the order > the tests are run in is not deterministic. Should be easy > to fix when I get a moment (no guarantees when though).

There is an attempt to prevent this in TU 0.24 by 's/^[\.F]+$/TEST-RUN- SUMMARY/sm', but corresponding strings (like '.F.Suite teardown') doesn't contain only '.' and 'F'. IMHO, regexp should be s/^[.F]+/TEST-RUN-SUMMARY/sm (there's no need no escape '.' in character calss). Thanks.

--- try_examples.t.orig 2004-03-31 14:52:29.570914800 +0400 +++ try_examples.t 2004-03-31 15:24:30.722535100 +0400 @@ -92,7 +92,7 @@ # bin the naughty carriage returns s/\r//g; # we can't assume the order of tests will be the same - s/^[\.F]+$/TEST-RUN-SUMMARY/sm; + s/^[.F]+/TEST-RUN-SUMMARY/sm; s/::Load[0-9_]+Anonymous[0-9_]+/::LOAD_ANONYMOUS_CLASSNAME/; # indent lines with '# ' so they're comments if the test fails s/\n/\n# /g;

Sat Jul 30 18:04:13 2005 Guest - Correspondence added

Subject:	fixed in CVS...
From:	mca1001 [...] users.sourceforge.net

[guest - Wed Mar 31 06:27:59 2004]: Show quoted text

> There is an attempt to prevent this in TU 0.24 by 's/^[\.F]+$/TEST- > RUN-SUMMARY/sm', but corresponding strings (like '.F.Suite teardown') > doesn't contain only '.' and 'F'.

Fixed in CVS, by including "\n?Suit teardown" before "$". Show quoted text

> IMHO, regexp should be s/^[.F]+/TEST-RUN-SUMMARY/sm (there's no need > no escape '.' in character calss).

Yes. Thanks. Project release & CPAN upload will follow eventually.

Wed Aug 03 13:39:39 2005 Guest - Correspondence added

[ASPIERS - Mon Mar 17 11:19:14 2003]: Show quoted text

> [guest - Mon Mar 17 08:55:21 2003]: >

> > I get a test failure in try_examples, related to > > examples/fail_example: > > The output differs VERY slightly (pseudo-diff syntax below ;-): > > < real output

> > > expected putput

> > < # ..FSuite teardown

> > > # .F.Suite teardown

> > > > I'm not sure whether this indicates a problem or not - let me > > know if you need more info. All other tests run ok. I got all the > > latest&greatest modules from CPAN (i.e. the "r" operation of the > > CPAN modules does not show any new modules available). > > My perl is 5.8.0 with ithreads and a shared libperl, compiled for > > Solaris 7 (Sparc) with Forte 6.2

> > Off the top of my head, it's probably just because the order > the tests are run in is not deterministic. Should be easy > to fix when I get a moment (no guarantees when though).

Test::Unit::Loader.pm Test::Unit::TestCase.pm To make the test cases deterministic, either update the list_tests method in Test::Unit::TestCase.pm or the list_tests call in Test::Unit::Loader.pm. A sort will do, and just happens to order the tests such that the unit test work.

Sat Oct 22 11:03:44 2005 mca1001 [...] users.sourceforge.net - Given to MCAST

Sat Oct 22 11:03:44 2005 mca1001 [...] users.sourceforge.net - Status changed from 'new' to 'resolved'

Sat Oct 22 11:03:44 2005 The RT System itself - Status changed from 'resolved' to 'open'

Sat Oct 22 11:03:44 2005 mca1001 [...] users.sourceforge.net - Correspondence added

Subject:

installation tests pass now

The release last week included the bug fix. http://testers.cpan.org/show/Test-Unit.html#Test-Unit-0.25 shows passes on various platforms, but not yet Solaris 7. I believe it should work though.

Sat Oct 22 11:03:44 2005 mca1001 [...] users.sourceforge.net - Fixed in 0.25 added

Sat Oct 22 11:05:04 2005 mca1001 [...] users.sourceforge.net - Status changed from 'open' to 'resolved'