Chris Pepper [Tue, 22 Apr 2003 04:31:17 +0000]
[perl #22011] [PATCH] pod/perlmod.pod (v5.8.0)
From: Chris Pepper (via RT) <perlbug-followup@perl.org>
Message-Id: <rt-22011-55473.11.
0523590303906@bugs6.perl.org>
(with minor further corrections)
p4raw-id: //depot/perl@19319
Dave Mitchell [Mon, 21 Apr 2003 13:19:50 +0000]
A new fatal error :
Subject: [PATCH] Perl_croak("Use of freed value in iteration")
Message-ID: <
20030421121950.GB18189@fdgroup.com>
Message-ID: <
20030421125433.GC18189@fdgroup.com>
p4raw-id: //depot/perl@19316
Jarkko Hietaniemi [Wed, 23 Apr 2003 18:53:46 +0000]
Introduce 'fast stdio' symbols to non-Configure lands.
p4raw-id: //depot/perl@19314
Jarkko Hietaniemi [Wed, 23 Apr 2003 18:21:47 +0000]
Regen Configure, Glossary, et alia.
p4raw-id: //depot/perl@19312
Jarkko Hietaniemi [Wed, 23 Apr 2003 08:34:33 +0000]
Handle the combination of dosish and non-faststdio.
p4raw-id: //depot/perl@19310
Ilya Zakharevich [Tue, 22 Apr 2003 11:43:17 +0000]
5.8.0: two problems
Message-ID: <
20030422184316.GA8602@math.berkeley.edu>
p4raw-id: //depot/perl@19308
Casey West [Tue, 22 Apr 2003 11:19:09 +0000]
Doc pointers in perlbot.
Message-Id: <
20030422151909.GC11119@geeknest.com>
(Applied by hand.)
p4raw-id: //depot/perl@19307
Jarkko Hietaniemi [Tue, 22 Apr 2003 13:09:21 +0000]
z/OS: changes for building threaded from "Brian De Pradine"
<PRADINE@uk.ibm.com>.
p4raw-id: //depot/perl@19306
Jarkko Hietaniemi [Tue, 22 Apr 2003 11:42:35 +0000]
Enache's turn.
p4raw-id: //depot/perl@19305
Jarkko Hietaniemi [Tue, 22 Apr 2003 10:48:47 +0000]
Re-merge Encode 1.92.
p4raw-id: //depot/perl@19304
Jarkko Hietaniemi [Tue, 22 Apr 2003 05:26:21 +0000]
Argh, cut-and-pasto.
p4raw-id: //depot/perl@19302
Artur Bergman [Mon, 21 Apr 2003 21:33:14 +0000]
Fix bug 17964 by properly refcounting the gv top.
Seems like the entries in rsfp are way different than
normal PVIO objects.
p4raw-id: //depot/perl@19301
Adrian M. Enache [Sun, 20 Apr 2003 02:45:48 +0000]
Fix another segfault case (warn called from UNIVERSAL::DESTROY).
Subject: Re: [perl #21347] segfault in UNIVERSAL::AUTOLOAD
Date: Sun, 20 Apr 2003 02:45:48 +0300
Message-ID: <
20030419234548.GA849@ratsnest.hole>
and
Date: Wed, 2 Apr 2003 07:52:28 +0300
Message-ID: <
20030402045227.GA1023@ratsnest.hole>
p4raw-id: //depot/perl@19300
Artur Bergman [Mon, 21 Apr 2003 17:05:31 +0000]
Croak on non ref to shared methods instead of segfault.
This can really only happen for &share("string").
p4raw-id: //depot/perl@19299
Jarkko Hietaniemi [Mon, 21 Apr 2003 16:24:36 +0000]
Add OpenZaurus cross-compilation patches.
p4raw-id: //depot/perl@19297
Artur Bergman [Mon, 21 Apr 2003 13:24:40 +0000]
Fixes bug #17043, resets PL_srand_called before the cloning.
Allows people to override the resetting by using srand() in CLONE.
p4raw-id: //depot/perl@19289
Jarkko Hietaniemi [Mon, 21 Apr 2003 06:50:42 +0000]
Introduce two new Configure symbols:
[1] d_faststdio = d_stdstdio && d_stdio_ptr_lval &&
(d_stdio_cnt_lval || d_stdio_ptr_lval_sets_cnt)
[2] usefaststdio = do we use fast stdio if we have it?
For 5.[68], we do. For anything else, we don't.
(At least, unless otherwise instructed by -Dusefaststdio.)
This means that for bleadperl we no more use stdio, but instead
default to perlio: the effect of PERLIO=perlio, in other words.
(PERLIO=stdio will still switch to using stdio.)
This change may endanger extensions using FILE*-- but if we are
to migrate fully to perlio, better start swallowing the poison now.
For maintperl, the usefaststdio still defaults to yes.
p4raw-id: //depot/perl@19286
Jarkko Hietaniemi [Mon, 21 Apr 2003 05:43:55 +0000]
Upgrade to Locale::Maketext 1.05.
p4raw-id: //depot/perl@19285
Jarkko Hietaniemi [Mon, 21 Apr 2003 05:40:34 +0000]
Upgrade to Test 1.24.
p4raw-id: //depot/perl@19284
Dave Mitchell [Sun, 20 Apr 2003 21:46:51 +0000]
goto &sub doesn't empty @_
Message-ID: <
20030420204651.GE15591@fdgroup.com>
p4raw-id: //depot/perl@19283
Adrian M. Enache [Fri, 18 Apr 2003 19:06:07 +0000]
Remove, in the main Makefile, the $(LDLIBPTH) variable from
the commands that invoke $(CC).
Subject: Re: Smoke failure
Message-ID: <
20030418160606.GA797@ratsnest.hole>
p4raw-id: //depot/perl@19282
Adrian M. Enache [Sat, 19 Apr 2003 20:12:18 +0000]
Re: [perl #22000] AutoReply: [PATCH] Naughty sv_gets/RsRECORD bug
Message-ID: <
20030419171218.GA974@ratsnest.hole>
p4raw-id: //depot/perl@19281
Steve Hay [Wed, 16 Apr 2003 11:12:22 +0000]
Document PERL_DEBUGGING_MSTATS in INSTALL and in the Win32 Makefile.
Subject: [DOC PATCH 5.8.0] Re: Problem with Devel::Peek on Perl 5.8.0
Message-ID: <
3E9D2C86.8040104@uk.radan.com>
p4raw-id: //depot/perl@19280
Jarkko Hietaniemi [Sun, 20 Apr 2003 07:42:17 +0000]
Give more information of what to do when -Dusexxx fail.
p4raw-id: //depot/perl@19278
Adrian M. Enache [Wed, 2 Apr 2003 05:02:42 +0000]
Fix bug #21347 (segfault in UNIVERSAL::AUTOLOAD with qr//)
by adding a dummy destructor method Regexp::DESTROY.
This prevents infinite recursion, since Regexp::DESTROY
is no more autoloaded.
Subject: Re: [perl #21347] segfault in UNIVERSAL::AUTOLOAD
Message-ID: <
20030402020242.GA2966@ratsnest.hole>
p4raw-id: //depot/perl@19277
Dave Mitchell [Sat, 19 Apr 2003 20:06:51 +0000]
Re: [PATCH] [perl #21875] Hash ref transformed as a list
Message-ID: <
20030419190651.GD13333@fdgroup.com>
p4raw-id: //depot/perl@19276
Jarkko Hietaniemi [Sat, 19 Apr 2003 18:52:24 +0000]
Restore the two variables retired by the change #19268
(for binary backward compatibility)
p4raw-link: @19268 on //depot/perl:
dd28f7bb7eebdb0b562c940b3c4f89457e829ea6
p4raw-id: //depot/perl@19275
Jarkko Hietaniemi [Sat, 19 Apr 2003 17:36:33 +0000]
Fail if the specified -Duse64xxx cannot be fulfilled.
(And undo a doublesize probe glitch from change #19272.)
p4raw-link: @19272 on //depot/perl:
fe63a0b4487686a4d99ab713bc73995cff5ffdc0
p4raw-id: //depot/perl@19273
Jarkko Hietaniemi [Sat, 19 Apr 2003 16:38:12 +0000]
If long doubles fail, we really cannot sensibly
drop back to doubles anymore, therefore abort.
Pointed out by Slaven Rezic. (I tried to fix this
via metaconfig but ended up with a vicious cyclic
dependency.)
p4raw-id: //depot/perl@19272
Jarkko Hietaniemi [Sat, 19 Apr 2003 12:09:21 +0000]
Upgrade to Net::Ping 2.30.
p4raw-id: //depot/perl@19270
Jarkko Hietaniemi [Sat, 19 Apr 2003 07:10:57 +0000]
Bump the test count.
p4raw-id: //depot/perl@19269
Dave Mitchell [Mon, 7 Apr 2003 10:00:41 +0000]
allow recursive FETCHes
Message-ID: <
20030407100041.A1617@fdgroup.com>
p4raw-id: //depot/perl@19268
Craig A. Berry [Thu, 17 Apr 2003 17:18:19 +0000]
readline.t tweak for VMS
From: "Craig A. Berry" <craigberry@mac.com>
Message-ID: <
3E9F282B.6090603@mac.com>
p4raw-id: //depot/perl@19267
Adrian M. Enache [Wed, 16 Apr 2003 21:11:11 +0000]
Re: any takers for this Storable bug?
Message-ID: <
20030416181111.GA6687@ratsnest.hole>
ams had already fixed the bug by #19227, but take the
test case.
p4raw-id: //depot/perl@19266
Artur Bergman [Sat, 19 Apr 2003 00:40:42 +0000]
Don't use the stack to allocate a SV that we hand out to another
function, only badness can ensure.
Message-Id: <
20030418180131.GD10086@fdgroup.com>
Thanks go to Dave Mitchell for help in tracking down the bug!
p4raw-id: //depot/perl@19265
Jarkko Hietaniemi [Fri, 18 Apr 2003 19:28:04 +0000]
UTF8 regexp patch from Inaba Hiroto.
p4raw-id: //depot/perl@19264
Jarkko Hietaniemi [Fri, 18 Apr 2003 19:16:43 +0000]
Unused variables.
p4raw-id: //depot/perl@19263
Artur Bergman [Thu, 17 Apr 2003 20:02:33 +0000]
Another needle in the voodoo doll.
Now I think it should work with the shared
strings.
p4raw-id: //depot/perl@19262
Artur Bergman [Thu, 17 Apr 2003 17:31:52 +0000]
Extreme Voodoo programming, by not pinching the SV if it is a
PADTMP it happily keeps it's value later on!
p4raw-id: //depot/perl@19254
Jarkko Hietaniemi [Thu, 17 Apr 2003 17:09:16 +0000]
Implement -Dnoextensions=... and -Donlyextensions=...
p4raw-id: //depot/perl@19253
Abhijit Menon-Sen [Thu, 17 Apr 2003 16:19:31 +0000]
[#21962] Speculative fix: don't set GvNAMELEN to 1 if GvNAME is "".
Anything that's broken by this change should be fixed.
p4raw-id: //depot/perl@19252
Jarkko Hietaniemi [Thu, 17 Apr 2003 15:39:56 +0000]
Introduce the new *l APIs, undefined, to non-Configure lands.
p4raw-id: //depot/perl@19251
Jarkko Hietaniemi [Thu, 17 Apr 2003 15:07:08 +0000]
Regen Glossary et alia.
p4raw-id: //depot/perl@19250
Jarkko Hietaniemi [Thu, 17 Apr 2003 14:53:41 +0000]
Small tidying up of the selectminbits code, but most
importantly allocate the buffer from the heap using
malloc, not from the stack: using stack causes unaligned
access in LP64 Solaris since stack data is not guaranteed
to be 64-bit aligned, while a fd_set is (it has 1024 longs).
(via Alan Burlison)
p4raw-id: //depot/perl@19248
Adrian M. Enache [Thu, 17 Apr 2003 14:35:36 +0000]
Re: [perl #21950] 5.8.x -- :scalar and magic strings
Message-Id: <
20030417113536.GA898@ratsnest.hole>
p4raw-id: //depot/perl@19247
Jarkko Hietaniemi [Thu, 17 Apr 2003 09:20:27 +0000]
Document that the open pragma does not affect
the three-headed hydra, ummm, open.
p4raw-id: //depot/perl@19245
Jarkko Hietaniemi [Thu, 17 Apr 2003 06:45:26 +0000]
More maneuvering closer to proposed POSIX classes
mapping to Unicode: (1) The definitions of "graph" and
"print" are still in flux: move closer back to how Perl
5.8.0 defined it (though drop Co, that did not make sense)
(2) Add U+0085 (NEL) to "space".
p4raw-id: //depot/perl@19244
Jarkko Hietaniemi [Thu, 17 Apr 2003 05:29:24 +0000]
Detect the 'fast stdio' case.
p4raw-id: //depot/perl@19243
Jarkko Hietaniemi [Wed, 16 Apr 2003 20:14:01 +0000]
Fix up Larry's copyright statements to my best knowledge.
(Lots of Perl 5 source code archaeology was involved.)
Larry didn't make strangled noises when I showed him
the patch, either :-)
p4raw-id: //depot/perl@19242
Jarkko Hietaniemi [Wed, 16 Apr 2003 17:24:24 +0000]
Handle PERLIO= and document a bit.
p4raw-id: //depot/perl@19241
Artur Bergman [Wed, 16 Apr 2003 16:27:43 +0000]
Fix double join problem for those places where free
actually means free (solaris for example :)
Fixes bug #16681
p4raw-id: //depot/perl@19240
Jarkko Hietaniemi [Wed, 16 Apr 2003 14:57:47 +0000]
Try to get the layers.t working also for dosish platforms.
p4raw-id: //depot/perl@19239
Jarkko Hietaniemi [Wed, 16 Apr 2003 13:27:24 +0000]
A -C parsed from the #! line can still have line endings in it.
p4raw-id: //depot/perl@19238
Jarkko Hietaniemi [Wed, 16 Apr 2003 13:23:37 +0000]
What does an empty PERLIO= mean, anyway?
p4raw-id: //depot/perl@19237
Jarkko Hietaniemi [Wed, 16 Apr 2003 13:09:42 +0000]
Document the effect of thread creation context.
p4raw-id: //depot/perl@19236
Jarkko Hietaniemi [Wed, 16 Apr 2003 12:13:07 +0000]
Use the right prototype and a glob is fine (from Rafael).
p4raw-id: //depot/perl@19235
Jarkko Hietaniemi [Wed, 16 Apr 2003 10:20:46 +0000]
PerlIO::get_layers() usage tip.
p4raw-id: //depot/perl@19234
Jarkko Hietaniemi [Wed, 16 Apr 2003 10:07:48 +0000]
Syntax terror.
p4raw-id: //depot/perl@19233
Jarkko Hietaniemi [Wed, 16 Apr 2003 09:45:51 +0000]
Redundant test.
p4raw-id: //depot/perl@19232
Jarkko Hietaniemi [Wed, 16 Apr 2003 09:21:25 +0000]
Synchronize the specifications of the POSIX character
classes alnum, graph, and print closer to the planned
Unicode proposal.
p4raw-id: //depot/perl@19231
Jarkko Hietaniemi [Wed, 16 Apr 2003 08:57:01 +0000]
Patch from Merijn.
p4raw-id: //depot/perl@19230
Jarkko Hietaniemi [Wed, 16 Apr 2003 08:27:26 +0000]
Thinko in modfl detection.
p4raw-id: //depot/perl@19228
Abhijit Menon-Sen [Wed, 16 Apr 2003 02:32:11 +0000]
[#17040] Storable now handles self-tied scalars with NULL mg_obj.
p4raw-id: //depot/perl@19227
Jarkko Hietaniemi [Tue, 15 Apr 2003 18:00:21 +0000]
Upgrade to Pod::LaTeX 0.55.
p4raw-id: //depot/perl@19226
Jarkko Hietaniemi [Tue, 15 Apr 2003 17:01:27 +0000]
The 'disable-longdouble-if-no-modfl-sqrtl' part didn't stick.
p4raw-id: //depot/perl@19224
Jarkko Hietaniemi [Tue, 15 Apr 2003 16:36:32 +0000]
Handle PERLIO=stdio, PERLIO=perlio, PERLIO=mmap, and no PERLIO.
p4raw-id: //depot/perl@19223
Jarkko Hietaniemi [Tue, 15 Apr 2003 16:27:49 +0000]
False alarm: no cyclic dependency.
p4raw-id: //depot/perl@19221
Jarkko Hietaniemi [Tue, 15 Apr 2003 16:08:11 +0000]
Backport of _most_ of the changes in #19218: not one big
cyclic dependency, though, it must be broken up somehow.
p4raw-id: //depot/perl@19219
Jarkko Hietaniemi [Tue, 15 Apr 2003 14:09:31 +0000]
Big patch from Alan Burlison to make Solaris long doubles work.
(Also spreads more <stdlib.h> joy.)
Needs backporting to metaconfig units and introducing the new
symbols to non-Configure lands.
p4raw-id: //depot/perl@19218
Jarkko Hietaniemi [Tue, 15 Apr 2003 14:01:03 +0000]
Don't skip the whole test for PERLIO=perlio.
p4raw-id: //depot/perl@19217
Jarkko Hietaniemi [Tue, 15 Apr 2003 05:54:17 +0000]
Add a test for [perl #17753].
p4raw-id: //depot/perl@19215
Jarkko Hietaniemi [Tue, 15 Apr 2003 05:40:52 +0000]
[perl #21845] "proto.h": 'id' is a reserved word in Objective-C.
p4raw-id: //depot/perl@19214
Artur Bergman [Mon, 14 Apr 2003 21:42:10 +0000]
Ook, run tests on correct build-dir next time.
We must check if the hv is a stash, otherwise we try to delete
something that is a HvNAME of null..... tests pass better now
p4raw-id: //depot/perl@19213
Artur Bergman [Mon, 14 Apr 2003 21:15:00 +0000]
Fixes to bugs introduced by PL_stashcache
A) Follow suggestion by Benjamin Goldberg to use hv_delete
instead of hv_delete_ent to avoid creating a temporary SV
B) Don't increment the refcount, sneak it into an IV instead
C) When a GP is a deleted that contains a stash, remove the
corresponding entry since hv might be in use in other places.
D) Note that no test cases test the deletion of packages to catch
this bug.
p4raw-id: //depot/perl@19212
Rafael Garcia-Suarez [Mon, 14 Apr 2003 21:11:11 +0000]
Mention that "make test_harness" lets messages sent through
standard error out. Complements change #19206.
p4raw-link: @19206 on //depot/perl:
244d9cb7d1d2e3a6aaa13f677e17b42673b10de9
p4raw-id: //depot/perl@19211
Inaba Hiroto [Fri, 21 Mar 2003 17:35:21 +0000]
Re: Parse::RecDescent triggers infinete loop in perl5.9.0 and 5.8.1 [perl #17757]
Date: Fri, 21 Mar 2003 17:35:21 +0900
Message-ID: <
3E7ACEC8.E0D57221@st.rim.or.jp>
Date: Sat, 22 Mar 2003 11:28:05 +0900
Message-ID: <
3E7BCA34.6824269F@st.rim.or.jp>
p4raw-id: //depot/perl@19210
Jarkko Hietaniemi [Mon, 14 Apr 2003 20:36:55 +0000]
PERLIO=perlio requires extra steps.
p4raw-id: //depot/perl@19209
Jarkko Hietaniemi [Mon, 14 Apr 2003 20:02:04 +0000]
Skip as early as possible if no perlio since the
use open would be compile-time.
p4raw-id: //depot/perl@19208
Nathan Torkington [Sat, 12 Apr 2003 12:40:45 +0000]
Document and test autovivified dirhandles.
Subject: [perl #21952] [patch] t/op/readdir.t and perlfunc.pod
From: Nathan Torkington (via RT) <perlbug-followup@perl.org>
Message-Id: <rt-21952-55079.8.
08945458828887@bugs6.perl.org>
p4raw-id: //depot/perl@19207
Rafael Garcia-Suarez [Mon, 14 Apr 2003 19:13:26 +0000]
More docs about running tests and the test suite.
Update perlhack.pod in various places.
Add 'test_harness' as a .PHONY dependency in the Makefile.
p4raw-id: //depot/perl@19206
Jarkko Hietaniemi [Mon, 14 Apr 2003 18:58:01 +0000]
Build get_layers() also under -Uuseperlio.
p4raw-id: //depot/perl@19205
Jarkko Hietaniemi [Mon, 14 Apr 2003 18:48:04 +0000]
Bump up the version and admonish against aspiring (ab)use.
p4raw-id: //depot/perl@19204
Jarkko Hietaniemi [Mon, 14 Apr 2003 17:35:51 +0000]
Introduce PerlIO::get_layers() to allow people to peek
at the PerlIO layer stack.
p4raw-id: //depot/perl@19203
Craig A. Berry [Mon, 14 Apr 2003 00:43:00 +0000]
vms/vms.c waitpid tweak
From: "Craig A. Berry" <craigberry@mac.com>
Message-ID: <
3E9A4A64.6030205@mac.com>
p4raw-id: //depot/perl@19202
Artur Bergman [Sun, 13 Apr 2003 19:26:46 +0000]
Fix bug #16253, use all the CPU that we can :-)
p4raw-id: //depot/perl@19201
Artur Bergman [Sun, 13 Apr 2003 18:38:58 +0000]
Fix bug #15893 by creating shared scalars with a refcount of 1
instead of 2 and make push and pop increase the refcount.
p4raw-id: //depot/perl@19200
Jarkko Hietaniemi [Sun, 13 Apr 2003 11:45:42 +0000]
Because we chdir to 'Big-Dummy' the lib/utf8.pm may be at
../../lib for the lc() at line 66 (UTF-8 data possible
if using UTF-8 locales).
p4raw-id: //depot/perl@19199
Jarkko Hietaniemi [Sun, 13 Apr 2003 09:43:09 +0000]
The lib/ipc_sysv.t test has a new name.
p4raw-id: //depot/perl@19198
Jarkko Hietaniemi [Sun, 13 Apr 2003 09:32:05 +0000]
Continue hunting for the cause of the Tru64 failed threaded smokes
most probably introduced by the change #19157. The tweak suggested
by Stephen McCamant trades a possible dangling pointer (CopFILE()
is tricky like that with threads) to a memory leak. (This means
that this tweak shouldn't be left in as-is, but if this helps,
we at least know that the failure was caused by the #19157.)
p4raw-link: @19157 on //depot/perl:
3871c2ef9aa1dfeba4631a17efdbf486012d47a6
p4raw-id: //depot/perl@19197
Jarkko Hietaniemi [Sun, 13 Apr 2003 09:20:02 +0000]
Upgrade to Net::Ping 2.29.
p4raw-id: //depot/perl@19196
Jarkko Hietaniemi [Sun, 13 Apr 2003 09:02:05 +0000]
Upgrade to Test::Harness 2.27_04.
p4raw-id: //depot/perl@19195
Rafael Garcia-Suarez [Sat, 12 Apr 2003 21:21:04 +0000]
Darwin/Rhapsody hints files updates by Wilfredo Sánchez
p4raw-id: //depot/perl@19194
Jarkko Hietaniemi [Sat, 12 Apr 2003 18:23:40 +0000]
The debugging aid #19182 didn't.
p4raw-id: //depot/perl@19192
Jarkko Hietaniemi [Sat, 12 Apr 2003 16:51:12 +0000]
Benchmark nit from Nicholas Clark (in OpenBSD the result could
be a negative zero, -0).
p4raw-id: //depot/perl@19191
Michael G. Schwern [Fri, 11 Apr 2003 00:32:02 +0000]
MakeMaker 6.10_02 -> 6.10_03
Message-ID: <
20030411073202.GC5916@windhund.schwern.org>
p4raw-id: //depot/perl@19190
Rafael Garcia-Suarez [Thu, 10 Apr 2003 20:40:14 +0000]
Peter Scott suggests that the docs for base.pm should
state explicitely that base.pm doesn't call import()
on base classes. Plus a few typos.
p4raw-id: //depot/perl@19189
Rafael Garcia-Suarez [Thu, 10 Apr 2003 19:45:08 +0000]
PerlFAQ sync.
p4raw-id: //depot/perl@19188
Rafael Garcia-Suarez [Thu, 10 Apr 2003 19:25:18 +0000]
Fix bug #21914, 3-arg open + local $fh + strict coredump.
There used to be a nullstash (%{"<none>::"}) to hold illegal
variables under strict 'vars'. It has been removed by change #17942.
This can lead to segfaults during the parsing, because some illegal
variables might have been discarded before the compilation of the
current statement had completed.
Bug analysis and test case by Enache Adrian.
p4raw-link: @17942 on //depot/perl:
de11ba31bba9f0eef0f76239d1d93a010926d6cf
p4raw-id: //depot/perl@19187
Jarkko Hietaniemi [Thu, 10 Apr 2003 19:13:40 +0000]
Be more descriptive.
p4raw-id: //depot/perl@19186
Jarkko Hietaniemi [Thu, 10 Apr 2003 19:06:02 +0000]
Noted by Nat: -0 didn't work that well with Unicode.
p4raw-id: //depot/perl@19185