Steve Peters [Tue, 6 Mar 2007 16:49:47 +0000]
Upgrade to IO-Zlib-1.05. There are updates to the core that have
not yet been included in the CPAN distribution, so the core
module has a version bump.
p4raw-id: //depot/perl@30489
Rafael Garcia-Suarez [Tue, 6 Mar 2007 15:51:06 +0000]
Fix error messages returned by S_require_tie_mod. Fix small leaks
happening in there too. More importantly, call it when we load both
a hash or a glob.
p4raw-id: //depot/perl@30488
Rafael Garcia-Suarez [Tue, 6 Mar 2007 15:04:03 +0000]
Fix a small leak on loading *+ and *- globs
p4raw-id: //depot/perl@30487
Steve Peters [Tue, 6 Mar 2007 13:52:56 +0000]
Upgrade to CGI.pm-3.27
p4raw-id: //depot/perl@30486
Jerry D. Hedden [Fri, 2 Mar 2007 06:59:41 +0000]
'Unknown PerlIO layer "encoding"...' message
From: "Jerry D. Hedden" <jdhedden@yahoo.com>
Message-ID: <794180.55454.qm@web30208.mail.mud.yahoo.com>
p4raw-id: //depot/perl@30485
Tels [Fri, 2 Mar 2007 11:31:39 +0000]
Doc patch about PerlIO::encoding:
Subject: [PATCH] Re: 'Unknown PerlIO layer "encoding"...' message
Message-Id: <
200703021131.41085@bloodgate.com>
p4raw-id: //depot/perl@30484
Rafael Garcia-Suarez [Tue, 6 Mar 2007 10:32:01 +0000]
dumpvar.pl must load re::Tie::Hash::NamedCapture if it wants
to dump %+ and %-.
p4raw-id: //depot/perl@30483
Steve Hay [Tue, 6 Mar 2007 09:20:51 +0000]
Reinstate part of #28744 that was accidentally reverted in #30454
to suppress a couple of warnings (at least on Win32)
p4raw-id: //depot/perl@30482
Nicholas Clark [Mon, 5 Mar 2007 23:52:09 +0000]
As SvUPGRADE() is a macro wrapping a call to sv_upgrade() inside
a check on the existing SvTYPE(), there's no need to wrap it inside
another explcit check of SvTYPE(). This won't make any difference to
the output of an optimising compiler, but it makes the source clearer.
p4raw-id: //depot/perl@30481
Nicholas Clark [Mon, 5 Mar 2007 12:26:46 +0000]
Change 30461 was wrong. As ext/re (re)builds the regexp engine with
-DDEBUGGING, it's going to need PL_reg_name even if core perl doesn't.
So something is always going to use it, so always define it, and always
export it. (But only define it once, so that static builds work.)
p4raw-id: //depot/perl@30464
Craig A. Berry [Mon, 5 Mar 2007 00:13:50 +0000]
In vms/gen_shrfls.pl, no more special handling of vars from
regcomp.h -- just process the file. (Needed for PL_reg_name.)
p4raw-id: //depot/perl@30463
Craig A. Berry [Sun, 4 Mar 2007 22:58:29 +0000]
Use linker flags rather than compiler flags in ExtUtils::Liblist::Kid::_vms_ext
when determining whether the perlshr library was linked debug. Also submitted
to CPAN RT queue at <http://rt.cpan.org/Ticket/Display.html?id=25268>.
p4raw-id: //depot/perl@30462
Nicholas Clark [Sun, 4 Mar 2007 19:22:45 +0000]
It seems that change 30451 should also have arranged to export
PL_reg_name properly. Hopefuly this will fix it, but I don't have
access to any platform where I can test this directly.
p4raw-id: //depot/perl@30461
Tels [Sun, 4 Mar 2007 15:57:01 +0000]
Math::BigInt 1.80
Message-Id: <
200703041557.02996@bloodgate.com>
p4raw-id: //depot/perl@30460
Nicholas Clark [Sun, 4 Mar 2007 11:51:08 +0000]
Rename exception() to apitest_exception() in XS/APItest, so that it
doesn't clash with the exception() in Devel/PPPort. This way, all
tests pass when both modules are staticly linked.
p4raw-id: //depot/perl@30459
Steve Peters [Sun, 4 Mar 2007 07:01:34 +0000]
Upgrade to IO-Compress-Base-2.004, Compress-Raw-Zlib-2.004,
IO-Compress-Zlib-2.004, and Compress-Zlib-2.004.
p4raw-id: //depot/perl@30458
Steve Peters [Sun, 4 Mar 2007 04:51:28 +0000]
Upgrade to Time-HiRes-1.9707
p4raw-id: //depot/perl@30457
Craig A. Berry [Sat, 3 Mar 2007 23:26:33 +0000]
Clean up the right file in vms/test.com (following #30414).
p4raw-id: //depot/perl@30456
Rafael Garcia-Suarez [Sat, 3 Mar 2007 18:19:01 +0000]
Add missing file from change #30454
p4raw-link: @30454 on //depot/perl:
caa6400115a38c0669db5e1b28eaa0eb8d1aad26
p4raw-id: //depot/perl@30455
Tels [Sat, 3 Mar 2007 17:09:58 +0000]
Math::BigInt::FastCalc 0.12 (plug leaks)
Message-Id: <
200703031710.00203@bloodgate.com>
p4raw-id: //depot/perl@30454
Nicholas Clark [Sat, 3 Mar 2007 15:50:33 +0000]
Must only use dynamic extensions in the Dynaloader test. If the
extensions are static, it doesn't work :-)
p4raw-id: //depot/perl@30453
Nicholas Clark [Sat, 3 Mar 2007 15:28:39 +0000]
As the test is about the parser, not actually running the code, better
to avoid running <STDOUT> rather than run it with warnings disabled.
p4raw-id: //depot/perl@30452
Nicholas Clark [Sat, 3 Mar 2007 12:22:19 +0000]
Define and initialise reg_name only once.
This allows re to be a static extension.
As it's now no-longer a static variable in regcomp.c, it needs a PL_
prefix.
p4raw-id: //depot/perl@30451
Rafael Garcia-Suarez [Sat, 3 Mar 2007 10:09:49 +0000]
Don't let feature.pm interfere with the tests for use VERSION.
That probably needs to be refined.
p4raw-id: //depot/perl@30450
Jerry D. Hedden [Thu, 1 Mar 2007 08:47:30 +0000]
Don't copy DynaLoader.o unnecessarily
From: "Jerry D. Hedden" <jdhedden@yahoo.com>
Message-ID: <566598.96493.qm@web30204.mail.mud.yahoo.com>
p4raw-id: //depot/perl@30449
Rafael Garcia-Suarez [Sat, 3 Mar 2007 09:36:54 +0000]
Add a volatile modifier to avoid possible cloberring by longjmp,
as the compiler used by Jerry D. Hedden warns.
p4raw-id: //depot/perl@30448
Rafael Garcia-Suarez [Sat, 3 Mar 2007 09:34:23 +0000]
Compilation warnings fixes by Jerry D. Hedden
p4raw-id: //depot/perl@30447
Rafael Garcia-Suarez [Sat, 3 Mar 2007 09:29:37 +0000]
Make use VERSION also load feature.pm
p4raw-id: //depot/perl@30446
Craig A. Berry [Sat, 3 Mar 2007 04:35:24 +0000]
Fix a few errors and ambiguities in configure.com caught by
Charlie Hammond's DCL Checker.
p4raw-id: //depot/perl@30445
Nicholas Clark [Fri, 2 Mar 2007 15:02:43 +0000]
Avoid an assertion failure when overloading readpipe.
p4raw-id: //depot/perl@30444
Nicholas Clark [Fri, 2 Mar 2007 14:42:27 +0000]
More assertion failures, found by auditing the code.
p4raw-id: //depot/perl@30443
Nicholas Clark [Fri, 2 Mar 2007 13:23:22 +0000]
Don't crash if the symbol table entry for ISA isn't a typeglob.
p4raw-id: //depot/perl@30442
Rafael Garcia-Suarez [Fri, 2 Mar 2007 12:58:04 +0000]
More work on perldelta
p4raw-id: //depot/perl@30441
Nicholas Clark [Thu, 1 Mar 2007 17:46:49 +0000]
The bug fixed by change 29610 would have been avoided with this
assertion - ensure that the hek argument to Perl_unshare_hek() is not
NULL, because if it is, then S_unshare_hek_or_pvn() will merrily
unshare the string "" instead.
p4raw-id: //depot/perl@30440
Nicholas Clark [Thu, 1 Mar 2007 16:09:31 +0000]
Don't SEGV when dumping an undefined typeglob. The HEK used to store
the GV's name can be NULL.
p4raw-id: //depot/perl@30439
Rafael Garcia-Suarez [Thu, 1 Mar 2007 15:22:52 +0000]
Mention the syntax C<use feature ':5.10'> in feature.pm's synopsis
p4raw-id: //depot/perl@30438
Rafael Garcia-Suarez [Thu, 1 Mar 2007 11:20:14 +0000]
Upgrade to Params::Check 0.26
p4raw-id: //depot/perl@30437
Yves Orton [Wed, 28 Feb 2007 17:45:33 +0000]
Re: New file: t/op/regexp_email.t
Message-ID: <
9b18b3110702280845p7860ca08taf1aead39a178aa4@mail.gmail.com>
p4raw-id: //depot/perl@30436
Rafael Garcia-Suarez [Thu, 1 Mar 2007 10:30:04 +0000]
A new test by Abigail: Check that certain modules don't
get loaded when other modules are used.
p4raw-id: //depot/perl@30435
Jerry D. Hedden [Wed, 28 Feb 2007 10:59:49 +0000]
Re: Revisits in 'make install'
From: "Jerry D. Hedden" <jdhedden@yahoo.com>
Message-ID: <95346.11233.qm@web30207.mail.mud.yahoo.com>
p4raw-id: //depot/perl@30434
Rafael Garcia-Suarez [Thu, 1 Mar 2007 10:10:41 +0000]
A couple of casting nits by Jarkko
p4raw-id: //depot/perl@30433
Rafael Garcia-Suarez [Thu, 1 Mar 2007 09:19:07 +0000]
Teach Module::Build about DragonflyBSD
p4raw-id: //depot/perl@30432
Rafael Garcia-Suarez [Wed, 28 Feb 2007 15:27:23 +0000]
Missing dual-lived modules in maintainers list
p4raw-id: //depot/perl@30431
Rafael Garcia-Suarez [Wed, 28 Feb 2007 14:51:41 +0000]
Update my email address
p4raw-id: //depot/perl@30430
Yitzchak Scott-Thoennes [Mon, 26 Feb 2007 11:28:00 +0000]
Win32 doc update for Win32CORE changes
From: "Yitzchak Scott-Thoennes" <sthoenna@efn.org>
Message-ID: <13678.68.166.55.187.
1172518080.squirrel@68.166.55.187>
p4raw-id: //depot/perl@30429
Abigail [Tue, 27 Feb 2007 23:00:28 +0000]
Unexpected output
Message-ID: <
20070227220028.GA460@abigail.nl>
p4raw-id: //depot/perl@30428
Rafael Garcia-Suarez [Wed, 28 Feb 2007 09:22:56 +0000]
Remove obsolete test for the deprecation warning I just removed.
p4raw-id: //depot/perl@30427
Jerry D. Hedden [Tue, 27 Feb 2007 09:38:15 +0000]
Fix ext/Win32API/File/t/file.t under test harness
From: "Jerry D. Hedden" <jdhedden@yahoo.com>
Message-ID: <994673.36516.qm@web30215.mail.mud.yahoo.com>
p4raw-id: //depot/perl@30426
Jerry D. Hedden [Tue, 27 Feb 2007 10:33:36 +0000]
Re: [PATCH] Re: Math::BigInt and Win32API::File whines a lot during make test...
From: "Jerry D. Hedden" <jdhedden@yahoo.com>
Message-ID: <738238.57355.qm@web30201.mail.mud.yahoo.com>
(Silences warnings about math lib(s), avoids using Math::BigInt
when using 64-bit ints and changes "new" to "->new()")
p4raw-id: //depot/perl@30425
Rafael Garcia-Suarez [Wed, 28 Feb 2007 08:54:05 +0000]
(?p{}) has been deprecated for a long time.
Remove it, so people are not confused with regard to the
new /p regexp modifier.
p4raw-id: //depot/perl@30424
Rafael Garcia-Suarez [Wed, 28 Feb 2007 08:45:55 +0000]
Add new tests for keys in %+ and %-
p4raw-id: //depot/perl@30423
Rafael Garcia-Suarez [Wed, 28 Feb 2007 08:45:13 +0000]
perltodo nit for SelfLoader
p4raw-id: //depot/perl@30422
Rafael Garcia-Suarez [Wed, 28 Feb 2007 08:44:16 +0000]
Fix a couple of typos in internal docs
p4raw-id: //depot/perl@30421
Abigail [Tue, 27 Feb 2007 23:54:56 +0000]
Change of email address.
Message-ID: <
20070227225456.GA1808@abigail.nl>
Date: Tue, 27 Feb 2007 23:54:56 +0100
p4raw-id: //depot/perl@30420
Jerry D. Hedden [Tue, 27 Feb 2007 11:27:58 +0000]
Skip tests if Encode not built
From: "Jerry D. Hedden" <jdhedden@yahoo.com>
Message-ID: <176638.9767.qm@web30207.mail.mud.yahoo.com>
p4raw-id: //depot/perl@30419
Rafael Garcia-Suarez [Tue, 27 Feb 2007 14:07:28 +0000]
Disable the strict pragma before eval'ing selfloaded code.
p4raw-id: //depot/perl@30418
Rafael Garcia-Suarez [Tue, 27 Feb 2007 13:39:03 +0000]
dmq points out that %+ and %- are not arrays. Fix the doc.
p4raw-id: //depot/perl@30417
Rafael Garcia-Suarez [Tue, 27 Feb 2007 13:16:46 +0000]
Update Changes
p4raw-id: //depot/perl@30416
Rafael Garcia-Suarez [Tue, 27 Feb 2007 10:56:17 +0000]
Rework and fix docs for %+, %- and re::Tie::Hash::NamedCapture.
p4raw-id: //depot/perl@30415
Craig A. Berry [Tue, 27 Feb 2007 01:41:31 +0000]
On VMS, use the normal executable file extension for the copy of
Perl used to run the test suite. Not doing so was confusing
some of the Module::Build tests.
p4raw-id: //depot/perl@30414
Yves Orton [Mon, 26 Feb 2007 16:45:23 +0000]
Re: [perl #41565] qr// memory corruption
Message-ID: <
9b18b3110702260745u373980e8x19ba3afdf33bd424@mail.gmail.com>
p4raw-id: //depot/perl@30413
Yves Orton [Thu, 15 Feb 2007 17:22:16 +0000]
Re: Regexp recursion limit too low?
Message-ID: <
9b18b3110702150822o13a4f240g86463c60e625fb8f@mail.gmail.com>
p4raw-id: //depot/perl@30412
Steve Hay [Mon, 26 Feb 2007 13:45:23 +0000]
Bump $Win32API::File::VERSION for change #30409
p4raw-link: @30409 on //depot/perl:
57d226e4c036539e44197f78a5d0e896298c6424
p4raw-id: //depot/perl@30411
Craig A. Berry [Mon, 26 Feb 2007 13:36:41 +0000]
In configure.com, include the Win32 extensions in $Config{known_extensions},
but not it $Config{extensions}. This way t/TEST knows to exclude them.
p4raw-id: //depot/perl@30410
Yitzchak Scott-Thoennes [Mon, 26 Feb 2007 00:45:10 +0000]
Re: [PATCH] Re: Win32 modules & cygwin
From: "Yitzchak Scott-Thoennes" <sthoenna@efn.org>
Message-ID: <1300.63.226.247.142.
1172479510.squirrel@63.226.247.142>
(Stops Win32* extension tests from being run when the
corresponding extensions haven't been built)
p4raw-id: //depot/perl@30409
Nicholas Clark [Mon, 26 Feb 2007 11:59:27 +0000]
Update the FSF's address. Plus some other whitespace and NON-WHITESPACE
changes that have crept into the GPL v2.
p4raw-id: //depot/perl@30408
Nicholas Clark [Mon, 26 Feb 2007 11:07:06 +0000]
Fix bug #41550 - AUTOLOAD :lvalue not working the same in blead as in
5.8.8 (a code example from "Extending and Embedding Perl")
p4raw-id: //depot/perl@30407
Nicholas Clark [Mon, 26 Feb 2007 09:33:15 +0000]
Convert t/op/sub_lval.t to test.pl
p4raw-id: //depot/perl@30406
Rafael Garcia-Suarez [Mon, 26 Feb 2007 06:59:44 +0000]
Module::Pluggable wasn't in the maintainers list
p4raw-id: //depot/perl@30405
Rafael Garcia-Suarez [Mon, 26 Feb 2007 06:57:22 +0000]
Add a version number to Module::Pluggable::Object and
bump the version number of Module::Pluggable
p4raw-id: //depot/perl@30404
Craig A. Berry [Mon, 26 Feb 2007 04:35:18 +0000]
In perlport, refine description of eight-level directory depth
limitation on older VMS systems and add example of how to tell
when we're in trouble.
p4raw-id: //depot/perl@30403
Craig A. Berry [Mon, 26 Feb 2007 01:08:07 +0000]
Assorted updates to vms/vms.c to fix compiler errors on VAX.
p4raw-id: //depot/perl@30402
Craig A. Berry [Mon, 26 Feb 2007 01:06:30 +0000]
One more longdouble nit for configure.com, following 30217, 30325, and 30329.
p4raw-id: //depot/perl@30401
Craig A. Berry [Mon, 26 Feb 2007 00:42:30 +0000]
Module::Pluggable::Object::search_paths portability update prompted by
VMS test failures. Patch also submitted to CPAN RT queue at
<http://rt.cpan.org/Public/Bug/Display.html?id=13607>.
p4raw-id: //depot/perl@30400
Alexey Tourbin [Sat, 24 Feb 2007 21:24:32 +0000]
sigtrap.pm (handler_traceback): do not clobber $_
Message-ID: <
20070224182432.GA14446@localhost.localdomain>
p4raw-id: //depot/perl@30399
Devin Heitmueller [Tue, 20 Feb 2007 16:38:20 +0000]
[perl #41560] [PATCH] crash in Perl_vmess when GvIOp is null
From: "Devin Heitmueller" (via RT) <perlbug-followup@perl.org>
Message-ID: <rt-3.6.HEAD-24617-
1172018300-131.41560-75-0@perl.org>
p4raw-id: //depot/perl@30398
Wolfgang Laun [Sat, 24 Feb 2007 15:08:44 +0000]
Re: [perl #41513] (parsing?)problem when using a '/' followed by a Switch statement
From: "Wolfgang Laun" <wolfgang.laun@gmail.com>
Message-ID: <
17de7ee80702240608n1411ef67xcdd1ffb238742bc1@mail.gmail.com>
p4raw-id: //depot/perl@30397
Rafael Garcia-Suarez [Sun, 25 Feb 2007 15:29:20 +0000]
- Use the same regexp as in change #30395 to parse subroutine
declarations
- Make SelfLoader strict-compliant
- Constant fold debug code out
p4raw-link: @30395 on //depot/perl:
91798d18643cb4539dc3cf0a83f879bbe84d17c2
p4raw-id: //depot/perl@30396
Rafael Garcia-Suarez [Sun, 25 Feb 2007 15:04:17 +0000]
Factorize three regexps into one, using new regexp features
(with some help from Yves Orton). (Now the compatibilty of this
module with 5.8.x perls is doomed.)
p4raw-id: //depot/perl@30395
Rafael Garcia-Suarez [Sun, 25 Feb 2007 14:42:11 +0000]
Don't clobber $_ by using "my $_". Potential bug found by Alexey Tourbin.
p4raw-id: //depot/perl@30394
Craig A. Berry [Sun, 25 Feb 2007 00:54:14 +0000]
Add a comment to mg.c to clarify that words like "raise" and
"pending" in the context of Perl's deferred signals bear only
an analogous relationship to the underlying signal mechanism.
p4raw-id: //depot/perl@30393
Craig A. Berry [Sat, 24 Feb 2007 21:52:23 +0000]
Update descriptions of long-running ops and non-deferrable
signals in perlipc.pod.
p4raw-id: //depot/perl@30392
Alexey Tourbin [Sat, 24 Feb 2007 15:52:23 +0000]
ExtUtils/Constant/Base.pm: use constant folding to eliminated dead code
Message-ID: <
20070224125223.GA17227@localhost.localdomain>
p4raw-id: //depot/perl@30391
Alexey Tourbin [Sat, 24 Feb 2007 15:38:58 +0000]
ExtUtils/MM_Unix.pm (fixin): do not clobber $/
Message-ID: <
20070224123858.GA11434@localhost.localdomain>
p4raw-id: //depot/perl@30390
Alexey Tourbin [Sat, 24 Feb 2007 14:47:35 +0000]
perlio.c (PerlIO_tmpfile): fix memory leak
Message-ID: <
20070224114735.GA3454@localhost.localdomain>
p4raw-id: //depot/perl@30389
Rafael Garcia-Suarez [Sat, 24 Feb 2007 09:53:56 +0000]
Add a new test for undef and delete on stash entries that
are bound to subroutines or methods. Based on a test by
Robert 'phaylon' Sedlacek.
p4raw-id: //depot/perl@30388
Rafael Garcia-Suarez [Sat, 24 Feb 2007 09:29:20 +0000]
C<keys %+> should not warn. Don't make scan_ident test for
ambiguous unary ops when the following token is a hash identifier.
(like with scalars and arrays)
p4raw-id: //depot/perl@30387
Rafael Garcia-Suarez [Sat, 24 Feb 2007 09:04:10 +0000]
Indexing and POD fixes
p4raw-id: //depot/perl@30386
Wolfgang Laun [Thu, 22 Feb 2007 09:36:19 +0000]
perlretut.pod, 2nd version
Message-ID: <
45DD5603.1040402@thalesgroup.com>
p4raw-id: //depot/perl@30385
Rafael Garcia-Suarez [Sat, 24 Feb 2007 07:22:05 +0000]
At least, fix the MakeMaker build_man tests.
p4raw-id: //depot/perl@30384
Rick Delaney [Wed, 21 Feb 2007 16:53:16 +0000]
Re: [perl #41546] perl 5.8.x bug: overloaded 'eq' does not work with 'nomethod'
Message-ID: <
20070221215316.GF5646@bort.ca>
p4raw-id: //depot/perl@30383
Rafael Garcia-Suarez [Thu, 22 Feb 2007 18:29:18 +0000]
Reinstate tests removed by change #30381, but don't
"use Config" in the test, because that will overwrite
%main::Config with the real Config hash, not the fake
one from MakeMaker.
p4raw-link: @30381 on //depot/perl:
cb7d39f54ca678d4ba0e91eaea2d10f4ab6317d7
p4raw-id: //depot/perl@30382
Steve Peters [Thu, 22 Feb 2007 18:09:06 +0000]
Revert changes to lib/ExtUtils/t/build_man.t from change
#30380. The changes in that test write to the now
readonly %Config.
p4raw-id: //depot/perl@30381
Steve Peters [Thu, 22 Feb 2007 14:36:27 +0000]
Upgrade to ExtUtils-MakeMaker-6.32. Included a version
bump to ExtUtils::MakeMaker to 6.32_01 because changes
#29266, #29583, #29580, #29581 and #29583 were missing from the new
release.
p4raw-id: //depot/perl@30380
Yitzchak Scott-Thoennes [Tue, 20 Feb 2007 01:50:18 +0000]
Re: Win32 modules & cygwin
From: "Yitzchak Scott-Thoennes" <sthoenna@efn.org>
Message-ID: <1254.67.42.109.122.
1171965018.squirrel@67.42.109.122>
Also includes integration & deletion of win32/ext/Win32API to
ext/Win32API, and addition of copyright message and corrections to
spacing/tabbing as suggested by Jan Dubois.
p4raw-id: //depot/perl@30379
Steve Peters [Wed, 21 Feb 2007 17:50:37 +0000]
Upgrade to CPAN-1.88_77.
p4raw-id: //depot/perl@30378
Jerry D. Hedden [Wed, 21 Feb 2007 09:27:20 +0000]
Re: [PATCH] Test with Errno in autouse.t
From: "Jerry D. Hedden" <jdhedden@yahoo.com>
Message-ID: <619324.6794.qm@web30207.mail.mud.yahoo.com>
p4raw-id: //depot/perl@30376
Steve Hay [Wed, 21 Feb 2007 13:36:14 +0000]
Remove unreferenced local variable
p4raw-id: //depot/perl@30375
Steve Hay [Wed, 21 Feb 2007 08:40:18 +0000]
Fix Win32 compilation error following #30372
p4raw-id: //depot/perl@30374
Craig A. Berry [Wed, 21 Feb 2007 01:55:37 +0000]
Fix VMS exit handling broken in Test::Harness 2.64. Also
submitted as http://rt.cpan.org/Public/Bug/Display.html?id=24985
on CPAN RT queue.
p4raw-id: //depot/perl@30373