[ANNOUNCE] Test::More/Builder/Simple 0.90

This is a quick release to sync with perl so 5.10.1 can release with a stable
version number.  It does NOT include the subtest() code in 0.89_01 (don't
worry, it'll be back).

http://github.com/schwern/test-more/tree/v0.90


0.90  Thu Jul  2 13:18:25 PDT 2009
    Docs
    * Finally added a note about the "Wide character in print" warning and
      how to work around it.
    * Note the IO::Stringy license in our copy of it.
      [test-more.googlecode.com 47]

    Test Fixes
    * Small fixes for integration with the Perl core
      [bleadperl eaa0815147e13cd4ab5b3d6ca8f26544a9f0c3b4]
    * exit code tests could be effected by errno when PERLIO=stdio
      [bleadperl c76230386fc5e6fba9fdbeab473abbf4f4adcbe3]

    Other
    * This is a stable release for 5.10.1.  It does not include
      the subtest() work in 0.89_01.


0.88  Sat May 30 12:31:24 PDT 2009
    Turing 0.87_03 into a stable release.

On Thu, Jul 2, 2009 at 4:03 PM, Michael G Schwern<schwern@pobox.com> wrote:
> This is a quick release to sync with perl so 5.10.1 can release with a stable
> version number.  It does NOT include the subtest() code in 0.89_01 (don't
> worry, it'll be back).

The attached patch does two things:

1.)  Copy and paste the skip logic from t/fork.t to
t/Builder/fork_with_new_stdout.t for platforms that don't have
$Config{d_fork} or pseudo-fork.

2.)  Now works a bit harder to suppress (most of) the SS$_ABORT
messages on VMS.  Passing -Mvmsish=hushed on the command line works
for one-liners, but not for separate programs.  I should know why that
is, but it's not coming to mind right now.  In any case, there is
still some chatter [1], but fixing that would mean prepending the
pragma to all the .plx files.


Tests run with:

$ perl -v

This is perl, v5.11.0 (GitLive-blead-1508-g24054a1*) built for VMS_IA64-ithread

[1]
ok 256 - exit map test for 255
# Done.
ok 257 - death.plx exited with 4 (expected 4)
%SYSTEM-F-ABORT, abort
%SYSTEM-F-ABORT, abort
ok 258 - too_few.plx exited with 4 (expected 4)
ok 259 - require.plx exited with 0 (expected 0)
%SYSTEM-F-ABORT, abort
%SYSTEM-F-ABORT, abort
ok 260 - too_few_fail.plx exited with 4 (expected 4)
%SYSTEM-F-ABORT, abort
%SYSTEM-F-ABORT, abort
ok 261 - five_fail.plx exited with 4 (expected 4)
ok 262 - pre_plan_death.plx exited with 4 (expected non-zero)
ok 263 - death_with_handler.plx exited with 4 (expected 4)
%SYSTEM-F-ABORT, abort
%SYSTEM-F-ABORT, abort
ok 264 - extras.plx exited with 4 (expected 4)
ok 265 - success.plx exited with 0 (expected 0)
%SYSTEM-F-ABORT, abort
%SYSTEM-F-ABORT, abort
ok 266 - two_fail.plx exited with 4 (expected 4)
ok 267 - death_in_eval.plx exited with 0 (expected 0)
ok 268 - last_minute_death.plx exited with 4 (expected 4)
%SYSTEM-F-ABORT, abort
%SYSTEM-F-ABORT, abort
ok 269 - one_fail.plx exited with 4 (expected 4)
ok 270 - exit.plx exited with 4 (expected 4)

Attachment: vms_test_more.patch

Thanks.  I added that as d238a5d to the stable branch.

Does this mean that 0.90 isn't suitable for 5.10.1 and that I need to
wait for a 0.91 or whatever?

We're back in the same place as before.  Its just a couple of little test
fixes.  You can apply them or I can make another release.

Since its easy enough to make a release now days I'll just do it.  But I feel
like we're going to have this conversation again in a couple days.

Also, Craig, you now have a commit bit on the test-more repo to speed up VMS
patching.  Feel free to make small fixes right in stable or master.  If you're
doing anything large make a VMS branch.

Well, not really. The ideal "CPAN" goals for a maint release are
1) include distributions that actually match something on CPAN
2) ship things that we think pass tests

So unless some other problem comes up, then 0.91 should match both those
goals and everyone's happy.

Dave Mitchell wrote:
> So unless some other problem comes up, then 0.91 should match both those
> goals and everyone's happy.

0.92 is on its way to CPAN with Craig's fixes.

http://github.com/schwern/test-more/tree/v0.92

OK thanks!

Its now in blead with 3e887aae38ef79132cc7112942154a36f73033f5,
and the bolierplate for three of the new tests fixed up with
45d8b36ec02e88acbbf23bf2514e70f8acb71985

Thanks, got it.

The test reports for 0.92 are all green for 5.10.0 so you're clear to add it
into 5.10.1.  You'll understand if I don't make a new release just for the
boilerplate fixes.