More adventures in toolchain zanyness

Alias on 2007-08-15T01:02:21

RedHat Enterprise Linux 4 is a deeply zany platform.

I've been discovering all sorts of unusual bugs in otherwise stable and mature modules.

For example, running prove.

Ignoring the failure for a second, I find it quite weird that the default readdir order is backwards. [adamk@dev1web05 k2]$ K2_TEST_WWW=XXXXXXXXX prove -r t t/www/02_login....NOK 7 # Failed test 'Page is loading' # at t/www/02_login.t line 52. # searched: "<html\x{0a}><head\x{0a}><meta content="text/html; charset=IS"... # can't find: "Page is loading" # Looks like you failed 1 test of 7. t/www/02_login....dubious Test returned status 1 (wstat 256, 0x100) DIED. FAILED test 7 Failed 1/7 tests, 85.71% okay t/www/01_index....ok t/02_compile......ok t/01_config.......ok Failed Test Stat Wstat Total Fail Failed List of Failed ------------------------------------------------------------------------------- t/www/02_login.t 1 256 7 1 14.29% 7 Failed 1/4 test scripts, 75.00% okay. 1/18 subtests failed, 94.44% okay.

Test::Simple is a very mature and stable module, at the core of Perl's ability to run unit tests and a dependency for more of CPAN... except that it fails tests on RedHat Enterprise Linux 4.

... t/skipall.................ok t/sort_bug................dubious Test returned status 255 (wstat -1, 0xffffffff) after all the subtests completed successfully t/tbt_01basic.............ok t/tbt_02fhrestore.........ok t/tbt_03die...............ok t/tbt_04line_num..........ok t/tbt_05faildiag..........ok t/tbt_06errormess.........ok t/tbt_07args..............ok t/thread_taint............ok t/threads.................dubious Test returned status 255 (wstat -1, 0xffffffff) after all the subtests completed successfully t/todo....................ok ...

For some reason threads are making the apparent return value to be 255, even when the test script has an explicit exit(0) in it.

And then there's Scalar::Util. I'm not sure if this is a problem in other places without the compiled version of Scalar::Util, but it appears that the "blessed" function has problems with one of the more evil tests in Params::Util, causing tests to fail for Params::Util.

The relevant test statements are

my $false_obj2 = bless \do { my $i } => "\0";

is( _INVOCANT($false_obj2), 1, 'Null-class blessed object is an invocant' );


I'm sure there's more oddities hidden in here I haven't found yet, so the Adventure in Toolchain Zanyness will surely continue.

Whatever it is that this OS is tickling, I'd love to see it fixed, so if anybody has a legal license of RHEL 4 I think it would be VERY useful to get a CPAN Testers installation running on it.


Test::Simple

Mr. Muskrat on 2007-08-16T14:30:39

Different results but still a failure.

>8 snip
t/skipall.................ok
t/sort_bug................dubious
        Test returned status 0 (wstat 11, 0xb)
DIED. FAILED tests 1-2
        Failed 2/2 tests, 0.00% okay
t/tbt_01basic.............ok
t/tbt_02fhrestore.........ok
t/tbt_03die... ............ok
t/tbt_04line_num..........ok
t/tbt_05faildiag..........ok
t/tb t_06errormess.........ok
t/tbt_07args..............ok
t/thread_taint.......... ..ok
t/threads.................ok
t/todo....................ok
t/undef....... ............ok
t/use_ok..................ok
t/useing..................ok
Fail ed Test  Stat Wstat Total Fail  Failed  List of Failed
------------------------------------------------------------------------ -------------------------------------------------------------------------------- ----
t/sort_bug.t    0    11     2    4 200.00%  1-2
4 tests and 10 subtests skipped.
Failed 1/66 test scripts, 98.48% okay. 2/491 subtests failed, 99.59% okay.
make: *** [test_dynamic] Error 255
  /usr/bin/make test -- NOT OK
Running make install
  make test had returned bad status, won't install without force