platform/upstream/perl.git
27 years agoDocument IO::File::new_tmpfile
Chip Salzenberg [Mon, 20 Jan 1997 16:11:30 +0000 (04:11 +1200)]
Document IO::File::new_tmpfile

27 years agoFix quotewords
Hugo van der Sanden [Thu, 6 Mar 1997 07:55:25 +0000 (07:55 +0000)]
Fix quotewords

Subject: Re: Fix for ::SUPER with UNIVERSAL

In <199703052319.SAA32568@rio.atlantic.net>, Chip Salzenberg writes:
:Cute.  It turns out that FileHandle and IO::Handle were okay, but the
:handling of ::SUPER (which is used by IO::Handle::new) was broken in
:the face of UNIVERSAL.
:
:Here's a patch.
That fixes it.
Here's one back - I think Text::ParseWords->quotewords is more broken than
this, but it's a start.

p5p-msgid: 199703060755.HAA15060@crypt.compulink.co.uk

27 years agoBenchmark: using code refs
Hugo van der Sanden [Tue, 4 Mar 1997 11:32:11 +0000 (11:32 +0000)]
Benchmark: using code refs

It often irritates me that I can't supply a coderef to the Benchmark::
routines: here's a small patch to allow them.

p5p-msgid: 199703041132.LAA07613@tyree.iii.co.uk

27 years agoRefresh Getopt::Long to 2.9
Johan Vromans [Fri, 28 Feb 1997 18:11:34 +0000 (06:11 +1200)]
Refresh Getopt::Long to 2.9

27 years ago[dummy merge]
Chip Salzenberg [Sat, 1 Mar 1997 06:40:49 +0000 (18:40 +1200)]
[dummy merge]

This merge exists so that the p5p version of the patch and the applied
version are both in the history

27 years agoNewer ReadLine
Chip Salzenberg [Sat, 1 Mar 1997 06:40:49 +0000 (18:40 +1200)]
Newer ReadLine

(this is the same change as commit d064f897c983478165072b3747bdb05ee656e1e6, but as applied)

27 years agoNewer ReadLine
Ilya Zakharevich [Tue, 4 Mar 1997 06:34:28 +0000 (01:34 -0500)]
Newer ReadLine

Finally I could compile GNU ReadLine2.1, so I got some experience with
T::R::Gnu. Unfortunately, debugger could not use any advanced features
of T::R::Gnu since it was accessing the features of T::R::Perl via
backdoors.

I reworked the interface to use object-oriented methods so that it
should not know anything about the particular ReadLine package it uses
(as far as it conforms to some simple interface).

Below is that part of the patch which should go into standard
distribution

Consider it as mildly emergent bug fix, it is -w-safe,

Enjoy,

p5p-msgid: 199703040634.BAA19919@monk.mps.ohio-state.edu

27 years agoChange 'continuing anyway' to 'probably harmless'
Chip Salzenberg [Fri, 28 Feb 1997 17:28:12 +0000 (05:28 +1200)]
Change 'continuing anyway' to 'probably harmless'

27 years agonear-harmless bug in _91's Configure
Roderick Schertler [Sun, 2 Mar 1997 02:26:49 +0000 (21:26 -0500)]
near-harmless bug in _91's Configure

At line 8118 of _91's Configure is this:

    echo"(I can't seem to compile the test program...)"

note missing space after echo.

p5p-msgid: pzg1yfuiza.fsf@eeyore.ibcinc.com

27 years agoRe: The continuing MachTen saga
Tom Phoenix [Wed, 5 Mar 1997 17:47:22 +0000 (09:47 -0800)]
Re: The continuing MachTen saga

private-msgid: Pine.GSO.3.95q.970305091611.3572E-100000@kelly.teleport.com

27 years agoHP-UX hint update
Raphael Manfredi [Thu, 6 Mar 1997 13:10:38 +0000 (14:10 +0100)]
HP-UX hint update

Quoting Chip Salzenberg:
:If you have a second, please look them over and see if anything
:important is out or anything broken is in.

I don't see my hint/hpux.sh patch. Here it is again (in unified format
this time):

p5p-msgid: 1479.857653838@lyon.grenoble.hp.com

27 years ago[differences between cumulative patch application and perl-5.003_91]
Chip Salzenberg [Tue, 25 Feb 1997 01:12:02 +0000 (13:12 +1200)]
[differences between cumulative patch application and perl-5.003_91]

27 years ago[inseparable changes from match from perl-5.003_90 to perl-5.003_91]
Perl 5 Porters [Tue, 25 Feb 1997 01:12:02 +0000 (13:12 +1200)]
[inseparable changes from match from perl-5.003_90 to perl-5.003_91]

 BUILD PROCESS

Subject: Sanity check linking with $libs
Date: Tue, 25 Feb 1997 14:13:45 -0500 (EST)
From: Andy Dougherty <doughera@fractal.phys.lafayette.edu>
Files: Configure
Msg-ID: <Pine.SOL.3.95q.970225221634.2486A-100000@fractal.lafayette.edu>

    (applied based on p5p patch as commit 5c37e92e59bb92e49d5a21017cd6dc066a28ddea)

Subject: Flush stdout when printing $randbits guess
From: Chip Salzenberg <chip@perl.com>
Files: Configure

Subject: Configure changes for Irix nm
From: Helmut Jarausch <helmutjarausch@unknown>
Files: Configure

 CORE LANGUAGE CHANGES

Subject: Fix perl_call_*() when !G_EVAL
Date: Tue, 25 Feb 1997 02:25:56 -0500
From: Gurusamy Sarathy <gsar@engin.umich.edu>
Files: MANIFEST gv.c interp.sym perl.c perl.h pp_ctl.c pp_sys.c t/op/runlevel.t
Msg-ID: <199702250725.CAA09192@aatma.engin.umich.edu>, <199702251925.OAA15498@aatma.engin.umich.edu>, <199702252200.RAA16853@aatma.engin.umich.edu>

    (applied based on p5p patch as commits 40f788c454d994616342c409de5b5d181ad9b8af, and 907a881cde89c56bc61d3f314c0efb8754ca472a20efc0829f6564c44574762adb07e8865bc14026)

Subject: Fix taint tests for writeable dirs in $ENV{PATH}
From: Chip Salzenberg <chip@perl.com>
Files: mg.c mg.h pod/perlsec.pod taint.c

Subject: Forbid tainted parameters for truncate()
From: Chip Salzenberg <chip@perl.com>
Files: pp_sys.c

Subject: Don't taint magic hash keys unnecessarily
Date: Fri, 28 Feb 1997 02:11:26 -0500 (EST)
From: Charles Bailey <bailey@HMIVAX.HUMGEN.UPENN.EDU>
Files: hv.c

    private-msgid: <01IFXL9TY74Y00661G@hmivax.humgen.upenn.edu>

 CORE PORTABILITY

Subject: VMS patches post _90
Date: Fri, 28 Feb 1997 15:26:33 -0500 (EST)
From: Charles Bailey <bailey@HMIVAX.HUMGEN.UPENN.EDU>
Files: doio.c mg.c perl.h pp_hot.c t/op/rand.t t/op/taint.t taint.c vms/descrip.mms vms/vms.c

    private-msgid: <01IFYDE5ZT7O005A53@hmivax.humgen.upenn.edu>

Subject: Fix taint check in system() and exec() under VMS and OS/2
From: Chip Salzenberg <chip@perl.com>
Files: pp_sys.c

Subject: If _XOPEN_VERSION >= 4, socket length parameters are size_t
From: Michael H. Moran <mhm@austin.ibm.com>
Files: perl.h pp_sys.c

Subject: Make dooneliner() compile again
From: Chip Salzenberg <chip@perl.com>
Files: pp_sys.c

 DOCUMENTATION

Subject: Move ENVIRONMENT from perl.pod to perlrun.pod
From: Chip Salzenberg <chip@perl.com>
Files: pod/perl.pod pod/perlrun.pod

Subject: Describe PERL_DEBUG_MSTATS in perlrun.pod
From: Nat <gnat@frii.com>
Files: pod/perlrun.pod

Subject: Fix references to perlbug
From: Chip Salzenberg <chip@perl.com>
Files: pod/perl.pod pod/perldelta.pod pod/perllocale.pod pod/perltoc.pod

 OTHER CORE CHANGES

Subject: Short-circuit duplicate study() calls
From: Chip Salzenberg <chip@perl.com>
Files: pp.c

Subject: Call sv_set[iu]v() with [IU]V parameter, not [IU]32
From: Chip Salzenberg <chip@perl.com>
Files: perl.c pp.c pp_sys.c toke.c util.c

Subject: Clean up and document API for hashes
Date: Tue, 25 Feb 1997 13:24:02 -0500
From: Gurusamy Sarathy <gsar@engin.umich.edu>
Files: hv.c hv.h pod/perldelta.pod pod/perlguts.pod
Msg-ID: <199702251824.NAA14859@aatma.engin.umich.edu>

    (applied based on p5p patch as commit a61fe43df197fcc70e6f310c06ee17d52b606c45)

Subject: pp_undef was not always freeing memory
Date: Thu, 27 Feb 1997 01:53:51 -0500 (EST)
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: pp.c
Msg-ID: <199702270653.BAA13949@monk.mps.ohio-state.edu>

    (applied based on p5p patch as commit 1da885048b65b5be1bd3077c6fc45f92c567e1b5)

Subject: Don't examine rx->exec_tainted if pregexec() fails
From: Chip Salzenberg <chip@perl.com>
Files: pp_hot.c

 TESTS

Subject: New test op/taint.t
Date: Tue, 25 Feb 1997 11:36:53 -0800 (PST)
From: Tom Phoenix <rootbeer@teleport.com>
Files: MANIFEST t/op/taint.t

    private-msgid: <Pine.GSO.3.95q.970225101328.18288M-100000@kelly.teleport.com

Subject: Patch to t/op/rand.t
Date: Tue, 25 Feb 1997 18:19:34 -0800 (PST)
From: Tom Phoenix <rootbeer@teleport.com>
Files: t/op/rand.t

    private-msgid: <Pine.GSO.3.95q.970225181321.13796Q-100000@kelly.teleport.com

 UTILITIES

Subject: Add --lax option to pod2man; use it in perldoc
From: Nat <gnat@frii.com>
Files: pod/pod2man.PL utils/perldoc.PL

Subject: Eliminate dead code in pod2man
From: Chip Salzenberg <chip@perl.com>
Files: pod/pod2man.PL

27 years agoSanity check linking with $libs
Andy Dougherty [Tue, 25 Feb 1997 19:13:45 +0000 (14:13 -0500)]
Sanity check linking with $libs

On Tue, 25 Feb 1997, Chip Salzenberg wrote:

> According to Andy Dougherty:
> > Chip, do you want to put something like this in, or shall I?  (I'd
> > need to borrow the current sfio unit)
>
> I'm leery of changing a working Configure, especially because _90's
> Configure has patches that haven't been propagated back to the units.

Staying leery is probably healthy.

Looking at _90's Configure, I see we'd have to change some
dependencies to move the sfio sanity checking earlier in the script.
I'd probably be inclined to do this by merging stuff into a single unit
(right after checking whether sfio is available, check whether it's
useful) and giving the whole thing a ?Y:TOP directive.

p5p-msgid: <Pine.SOL.3.95q.970225133824.28903K-100000@fractal.lafayette.edu>
private-msgid: <Pine.SOL.3.95q.970225221634.2486A-100000@fractal.lafayette.edu>

27 years agoUpdates to perlfunc.pod
Tom Phoenix (with help from M.J.T. Guy [Mon, 24 Feb 1997 16:36:06 +0000 (04:36 +1200)]
Updates to perlfunc.pod
Signed-off-by: Tom Phoenix (with help from M.J.T. Guy <tomphoenix@unknown>
27 years agopp_undef was not always freeing memory
Ilya Zakharevich [Thu, 27 Feb 1997 06:53:51 +0000 (01:53 -0500)]
pp_undef was not always freeing memory

As I found, pp_undef in
$a = 'aaa';
$a = 0;
undef $a;
was not reclaiming the storage, as it does in
        $a = 'aaa';
        undef $a;

The patch is below fixes this shortcoming (tests ok, perl rebuilds
itself OK),

Enjoy,

p5p-msgid: <199702270707.CAA13978@monk.mps.ohio-state.edu>
private-msgid: <199702270653.BAA13949@monk.mps.ohio-state.edu>

27 years agoWarn about intrusive sfio behavior
Andy Dougherty [Fri, 28 Feb 1997 16:35:49 +0000 (11:35 -0500)]
Warn about intrusive sfio behavior

private-msgid: <Pine.SOL.3.95q.970228112136.24038G-100000@fractal.lafayette.

27 years agoClean up and document API for hashes
Gurusamy Sarathy [Tue, 25 Feb 1997 18:24:02 +0000 (13:24 -0500)]
Clean up and document API for hashes

On Tue, 25 Feb 1997 09:01:57 EST, Chip Salzenberg wrote:
>I've had 99.44% good reports about _90.  I'd love to go public
>immediately, but Sarathy's perl_call fix is just too valuable to pass
>up, and there's a taint test in the works.  So please don't tell
>anyone that _90 is the first public beta; that will likely be _91.

FWIW, the Beta Zvoid is alive and kicking on Linux 2.0.28.

You might want to put this in too.  (The patches to hv.h are mostly
to an unused macro that will be useful for extension writers).

p5p-msgid: <199702251824.NAA14859@aatma.engin.umich.edu>

27 years agoAllow explicit '-lperl' in link arguments
Doug MacEachern [Thu, 27 Feb 1997 16:25:04 +0000 (11:25 -0500)]
Allow explicit '-lperl' in link arguments

private-msgid: <199702271625.LAA25402@postman.osf.org>

27 years ago_90 op/tie.t gets "Segmentation fault" with Sarathy's perl_call patch
Gurusamy Sarathy [Tue, 25 Feb 1997 22:00:49 +0000 (17:00 -0500)]
_90 op/tie.t gets "Segmentation fault" with Sarathy's perl_call patch

On Tue, 25 Feb 1997 14:41:30 EST, Chip Salzenberg wrote:
>According to Gurusamy Sarathy:
>> On Tue, 25 Feb 1997 10:07:02 PST, David Dyck wrote:
>> >summary:
>> >  op/tie.t gets "Segmentation fault" with  Sarathy's perl_call patch
>>
>> Try this patch.
>
>That's got it.  Now pureperl doesn't complain about op/tie.t.
>

Good.  This here patch sets up the flag in a few more places (to avoid
redundant sigjmp()s which was why t/op/tie.t triggered execution
of docatch()).  This patch fixes that flag leak.

p5p-msgid: <199702252200.RAA16853@aatma.engin.umich.edu>

27 years ago[dummy merge]
Chip Salzenberg [Tue, 25 Feb 1997 01:12:02 +0000 (13:12 +1200)]
[dummy merge]

This merge exists so that the p5p version of the patch and the applied
version are both in the history

27 years ago_90 op/tie.t gets "Segmentation fault" with Sarathy's perl_call patch
Gurusamy Sarathy [Tue, 25 Feb 1997 19:25:39 +0000 (14:25 -0500)]
_90 op/tie.t gets "Segmentation fault" with Sarathy's perl_call patch

On Tue, 25 Feb 1997 10:07:02 PST, David Dyck wrote:
>
>summary:
>  op/tie.t gets "Segmentation fault" with  Sarathy's perl_call patch
>

Oops.  Thanks for that report (Evil how the segfault doesn't show
up under `make test`. Evil.)

Try this patch. (I was being risque with the runlevel pretense...)

p5p-msgid: <199702251925.OAA15498@aatma.engin.umich.edu>

27 years agoDon't warn on use of CCFLAGS
Chip Salzenberg [Tue, 25 Feb 1997 01:12:02 +0000 (13:12 +1200)]
Don't warn on use of CCFLAGS

(this is the same change as commit 69d91587b2e24056be123e153ca28f4e82fa8dff, but as applied)

27 years agoDon't warn on use of CCFLAGS
Andreas Koenig [Tue, 25 Feb 1997 10:38:43 +0000 (11:38 +0100)]
Don't warn on use of CCFLAGS

Subject: CCFLAGS (Was: This message did not make it ...)

>>>>> Raphael Manfredi writes:

  >  Making ODBM_File (dynamic)
  >  Processing hints file hints/hpux.pl
  >  'CCFLAGS' is not a known MakeMaker parameter name.

  > error. Upon closer inspection, it looks like the hint file is
  > doing:

  >  $self->{CCFLAGS} = $Config{ccflags} . ' -DDBM_BUG_DUPLICATE_FREE' ;

  > Why the 'CCFLAGS' key is unknown is beyond me. It would be nice
  > to find the roots of this problem before we go to beta.

Finally, here it is. Unfortunately, I cannot remember how it was lost
from the list.

p5p-msgid: <199702251038.LAA13123@anna.in-berlin.de>

27 years agoFix perl_call_*() when !G_EVAL
Gurusamy Sarathy [Tue, 25 Feb 1997 07:25:56 +0000 (02:25 -0500)]
Fix perl_call_*() when !G_EVAL

On Mon, 24 Feb 1997 15:19:17 EST, Gurusamy Sarathy wrote:
>On Mon, 24 Feb 1997 12:53:57 GMT, Tim Bunce wrote:
>>> From: Tom Christiansen <tchrist@jhereg.perl.com>
>>> >Dprof "works".
>>> Then how come it's not in the core? :-(
>>I'd certainly like it to be there for 5.004.
>
>I'd agree, except there's this bug in perl_call_*() that makes
>it fail to run this fully:
>
>   % perl -d:DProf -e 'sub T { eval { die "burp" } } T(); print "zip\n"'
>   %

Ok, here's a patch for the perl_call_*() problems with error traps,
meant for 5.004 (hope I didn't miss the boat!).

This is a subset of the functionality contained in Michael Schroeder's
stack-of-stacks patch.  The patch itself if simple: code that calls
runops() without explicitly setting up a jmp_buf sets a flag that
indicates doeval() is responsible for catching any longjmp()s
locally.  The three places that call doeval() then call setjmp()
based on this flag.

This patch is binary compatible and minimal (as opposed to the
stack-of-stacks patch which has other issues involved, making it
more complicated).  There's a testsuite with 9 tests (3_28 fails all
but one).

p5p-msgid: <199702250725.CAA09192@aatma.engin.umich.edu>

27 years ago[dummy merge]
Chip Salzenberg [Tue, 25 Feb 1997 01:12:02 +0000 (13:12 +1200)]
[dummy merge]

This merge exists so that the p5p version of the patch and the applied
version are both in the history

27 years agoFollow up on elimination of $` $& $' in libraries
Chip Salzenberg [Tue, 25 Feb 1997 01:12:02 +0000 (13:12 +1200)]
Follow up on elimination of $` $& $' in libraries

(this is the same change as commit efa45da9e2f446e4e79f819d2fa3a099e54425fa, but as applied)

27 years agoFollow up on elimination of $` $& $' in libraries
M.J.T. Guy [Fri, 28 Feb 1997 13:59:42 +0000 (13:59 +0000)]
Follow up on elimination of $` $& $' in libraries

Ilya Zakharevich <ilya@math.ohio-state.edu> wrote
> $' $` $& eliminated from all the known *.pm.

Close, but you overlooked the 'use English' in diagnostics.pm.
See additional patch below.

p5p-msgid: <E0w0Sqc-00046E-00@ursa.cus.cam.ac.uk>

27 years agoMake IPC::Open3 work without fork()
Ilya Zakharevich [Tue, 25 Feb 1997 19:37:07 +0000 (14:37 -0500)]
Make IPC::Open3 work without fork()

private-msgid: <199702251937.OAA10718@monk.mps.ohio-state.edu>

27 years agoDon't require() in a signal handler
Chip Salzenberg [Tue, 25 Feb 1997 01:58:14 +0000 (13:58 +1200)]
Don't require() in a signal handler

27 years ago[dummy merge]
Chip Salzenberg [Tue, 25 Feb 1997 01:12:02 +0000 (13:12 +1200)]
[dummy merge]

This merge exists so that the p5p version of the patch and the applied
version are both in the history

27 years agoUpdate OS/2 Configure diff
Chip Salzenberg [Tue, 25 Feb 1997 01:12:02 +0000 (13:12 +1200)]
Update OS/2 Configure diff

(this is the same change as commit ef05b0d7eb79c109aa2e843e7ab5d5abad251c55, but as applied)

27 years agoUpdate OS/2 Configure diff
Ilya Zakharevich [Tue, 25 Feb 1997 19:06:23 +0000 (14:06 -0500)]
Update OS/2 Configure diff

Chip Salzenberg writes:
>
> (I missed the $randbits case in _90; it is fixed in _91.)

The broken $randbits test (no $ldflags) is just ignored under OS/2
anyway, so this should help.

p5p-msgid: <199702251906.OAA10608@monk.mps.ohio-state.edu>

27 years agoSilence bogus typo warning on $DB::postponed
Gurusamy Sarathy [Thu, 27 Feb 1997 18:02:30 +0000 (13:02 -0500)]
Silence bogus typo warning on $DB::postponed

private-msgid: <199702271802.NAA12505@aatma.engin.umich.edu>

27 years agoFix SEGV when debugging with foreach() lvalue patch
Ilya Zakharevich [Thu, 27 Feb 1997 19:24:36 +0000 (14:24 -0500)]
Fix SEGV when debugging with foreach() lvalue patch

p5p-msgid: <199702271924.OAA14557@monk.mps.ohio-state.edu>

27 years ago[differences between cumulative patch application and perl-5.003_90]
Chip Salzenberg [Fri, 21 Feb 1997 14:41:53 +0000 (02:41 +1200)]
[differences between cumulative patch application and perl-5.003_90]

27 years ago[inseparable changes from match from perl5.003_28 to perl-5.003_90]
Perl 5 Porters [Fri, 21 Feb 1997 14:41:53 +0000 (02:41 +1200)]
[inseparable changes from match from perl5.003_28 to perl-5.003_90]

 BUILD PROCESS

Subject: Re: ccdlflags don't quite work
Date: Mon, 24 Feb 1997 16:07:07 -0500 (EST)
From: Andy Dougherty <doughera@fractal.phys.lafayette.edu>
Files: Configure
Msg-ID: <Pine.SOL.3.95q.970224160630.5700E-100000@fractal.lafayette.e

    (applied based on p5p patch as commit bc75e68b7b49a5ac07dc839aec7ca05eba51175f)

Subject: Use $ccflags, $ldflags, $libs when determining $randbits
From: Chip Salzenberg <chip@perl.com>
Files: Configure

Subject: 'installperl -v' doesn't do enough
Date: Mon, 24 Feb 1997 08:42:59 -0500
From: Spider Boardman <spider@Orb.Nashua.NH.US>
Files: installperl
Msg-ID: <199702241342.IAA25945@Orb.Nashua.NH.US>

    (applied based on p5p patch as commit 7287d43f85a3731984c0e0ef1f20e1515b4140f2)

Subject: installperl breaks running system (for a while)
Date: Mon, 24 Feb 1997 09:12:11 -0500
From: Spider Boardman <spider@Orb.Nashua.NH.US>
Files: installperl
Msg-ID: <199702241412.JAA11829@Orb.Nashua.NH.US>

    (applied based on p5p patch as commit ec2a3d87ba26a9586ad981051a949403027c147a)

 CORE PORTABILITY

Subject: Digital UNIX and 3_28
Date: Sun, 23 Feb 1997 16:27:19 +0200 (EET)
From: Jarkko Hietaniemi <jhi@iki.fi>
Files: Configure MANIFEST ext/NDBM_File/hints/dec_osf.pl ext/ODBM_File/hints/dec_osf.pl hints/dec_osf.sh
Msg-ID: <199702231427.QAA13807@alpha.hut.fi>

    (applied based on p5p patch as commit 58c0852f6968d0b4520a03e22a56226185ab78c6)

Subject: AmigaOS patches to 5.003_28
Date: Sat, 22 Feb 1997 18:08:02 +0100
From: "Norbert Pueschel" <pueschel@imsdd.meb.uni-bonn.de>
Files: README.amiga hints/amigaos.sh t/io/fs.t t/lib/anydbm.t t/lib/db-btree.t t/lib/db-hash.t t/lib/db-recno.t t/lib/gdbm.t t/lib/ndbm.t t/lib/odbm.t t/lib/sdbm.t t/op/magic.t t/op/stat.t
Msg-ID: <77724759@Armageddon.meb.uni-bonn.de>

    (applied based on p5p patch as commit b8e6d11c134e93a7795379ceb62b7f950607c667)

 TESTS

Subject: Don't use <*> where readdir() will do
From: Chip Salzenberg <chip@perl.com>
Files: t/op/stat.t

Subject: Allow for $^X to be 'miniperl'
Date: Sun, 23 Feb 1997 16:22:45 +0100
From: Dominic Dunlop <domo@slipper.ip.lu>
Files: t/op/magic.t

    private-msgid: <v03020903af360f31aced@[194.51.248.65]>

27 years agoinstallperl breaks running system (for a while)
Spider Boardman [Mon, 24 Feb 1997 14:12:11 +0000 (09:12 -0500)]
installperl breaks running system (for a while)

"installperl" causes scripts with "/usr/bin/perl" or equivalent to
fail during the process of installation, which can take a while with
slow disks and/or slow CPUs.  It should install the libraries and such
*before* linking "$installbin/perl" to the version-dependent executable.

p5p-msgid: <199702241412.JAA11829@Orb.Nashua.NH.US>

27 years ago[dummy merge]
Chip Salzenberg [Fri, 21 Feb 1997 14:41:53 +0000 (02:41 +1200)]
[dummy merge]

This merge exists so that the p5p version of the patch and the applied
version are both in the history

27 years ago'installperl -v' doesn't do enough
Spider Boardman [Mon, 24 Feb 1997 13:42:59 +0000 (08:42 -0500)]
'installperl -v' doesn't do enough

"installperl -v" doesn't install enough for even "perl5.00328 -V" to
work.  It should install the version-dependent portions of the "lib"
directory as well as the perl executable.

p5p-msgid: <199702241342.IAA25945@Orb.Nashua.NH.US>

27 years agoDocument one-argument limitation with #! line
Chip Salzenberg [Fri, 21 Feb 1997 14:41:53 +0000 (02:41 +1200)]
Document one-argument limitation with #! line

(this is the same change as commit 3c5180aa4845c30d7891bc90e491cc6f9b6ecd85, but as applied)

27 years agoDocument one-argument limitation with #! line
Tom Phoenix [Mon, 24 Feb 1997 02:41:02 +0000 (18:41 -0800)]
Document one-argument limitation with #! line

On Sun, 23 Feb 1997, Tom Christiansen wrote:

> If it is important to you that your bug be fixed, do monitor the
> Changes file in any development releases since the time you submitted
> the bug, and encourage the maintainers with kind words (but never any
> flames!).  Feel free to resend your bug report if the next released
> version of perl comes out and your bug is still present.

That's good. And how about adding something like this?

p5p-msgid: <Pine.GSO.3.95q.970223151026.18178P-100000@kelly.teleport.com>
private-msgid: <Pine.GSO.3.95q.970223182745.15989A-100000@kelly.teleport.com

27 years agoRe: ccdlflags don't quite work
Andy Dougherty [Mon, 24 Feb 1997 21:07:07 +0000 (16:07 -0500)]
Re: ccdlflags don't quite work

On Mon, 24 Feb 1997, Matthias Urlichs wrote:

> Two problems:
>
> # How will the perl executable find the installed shared $libperl?
> # Add $xxx to ccdlflags.
> ... and it does this every time I reconfigure Perl without first saying
> "rm config.sh", so the list grows longer and longer and... Ugh. Probably
> the best solution is to add a Configure variable which does nothing but
> remember if we have added $xxx yet.

I think this will fix the problem.

p5p-msgid: <Pine.SOL.3.95q.970224160630.5700E-100000@fractal.lafayette.edu>
private-msgid: <Pine.SOL.3.95q.970224160630.5700E-100000@fractal.lafayette.e

27 years agoFix typo in 'Tolkien quotation typo' fix
Jarkko Hietaniemi [Fri, 21 Feb 1997 15:38:06 +0000 (03:38 +1200)]
Fix typo in 'Tolkien quotation typo' fix

27 years agoAmigaOS patches to 5.003_28
Norbert Pueschel [Sat, 22 Feb 1997 17:08:02 +0000 (18:08 +0100)]
AmigaOS patches to 5.003_28

Here are some patches for AmigaOS, mainly to switch of unwanted tests.

With these patches, perl 5.003_28 passes all tests under AmigaOS.

You should check if there are other hint files that need the new
dont_use_nlink variable added.

p5p-msgid: <77724759@Armageddon.meb.uni-bonn.de>

27 years agopumpkin-1.9.pod
Andy Dougherty [Mon, 24 Feb 1997 21:06:02 +0000 (16:06 -0500)]
pumpkin-1.9.pod

private-msgid: <Pine.SOL.3.95q.970224155702.5700D-100000@fractal.lafayette.e

27 years agoDigital UNIX and 3_28
Jarkko Hietaniemi [Sun, 23 Feb 1997 14:27:19 +0000 (16:27 +0200)]
Digital UNIX and 3_28

Please try this on 3_28 in as many Digital UNIX / DEC OSF/1 boxes you
can and report the results back to me. I have already 100% positive
make test results from

Digital UNIX 4.0 cc static/shared, gcc static/shared
Digital UNIX 4.0B cc static/shared, gcc static/shared
DEC OSF/1 3.2C cc static/shared, gcc static/shared

so you can test some other os / compiler combinations first.

p5p-msgid: <199702231427.QAA13807@alpha.hut.fi>

27 years ago[dummy merge]
Chip Salzenberg [Fri, 21 Feb 1997 14:41:53 +0000 (02:41 +1200)]
[dummy merge]

This merge exists so that the p5p version of the patch and the applied
version are both in the history

27 years agoAdd documentation and '-h' option to perlbug
Chip Salzenberg [Fri, 21 Feb 1997 14:41:53 +0000 (02:41 +1200)]
Add documentation and '-h' option to perlbug

This was the summary description that Chip wrote for the two changes:

4fc194b2052979cb8d28873cf62c0ce84cd214f5
3f4fc4634a226330cf98eab6871f26f45d52e425

27 years agoperlbug.pod for 5.004
Gurusamy Sarathy [Mon, 24 Feb 1997 20:09:51 +0000 (15:09 -0500)]
perlbug.pod for 5.004

On Mon, 24 Feb 1997 16:10:10 GMT, "M.J.T. Guy" wrote:
>You wrote to perl5-porters:
>> I hope I got Kenneth right as the AUTHOR of perlbug.
>
>Doesn't he deserve the E<lt> treatment, like the rest of you?
>
>>+ Kenneth Albanowski <kjahds@kjahds.com>, subsequently I<doc>tored by

Sure.  That, and a few pod2man noises fixed below.  Goes on top
of the previous patch.

p5p-msgid: <199702242009.PAA02849@aatma.engin.umich.edu>

27 years agoRe: Hash key created by subroutine call? (fwd)
Gurusamy Sarathy [Mon, 24 Feb 1997 22:29:30 +0000 (17:29 -0500)]
Re: Hash key created by subroutine call? (fwd)

On Mon, 24 Feb 1997 06:20:14 MST, Tom Christiansen wrote:
>Considering how frequently this gets asked, how about
>putting it in perlsub?
>
>------- start of forwarded message -------
>From: Rick Smith <ricks@sd.znet.com>
>Newsgroups: comp.lang.perl.misc
>Subject: Hash key created by subroutine call?
>Date: Sat, 22 Feb 1997 21:47:52 -0800

I seem to be in a documental state today.

p5p-msgid: <199702242229.RAA04395@aatma.engin.umich.edu>

27 years agoperlbug.pod for 5.004
Gurusamy Sarathy [Sun, 23 Feb 1997 11:00:00 +0000 (00:00 +1300)]
perlbug.pod for 5.004

On Sun, 23 Feb 1997 15:06:24 MST, Tom Christiansen wrote:
>Maybe this should talk more about the perlbug program, or be podfully
>included in that script?
>
>--tom

I've made some additions, along with a merge into utils/perlbug.PL.
(I didn't see Tom(rootbeer)'s suggestion before making the patch,
so that's not in.)

I hope I got Kenneth right as the AUTHOR of perlbug.

Do we need a `ln -s utils/perlbug pod/perlbug.pod` too?

p5p-msgid: <199702240854.DAA27128@aatma.engin.umich.edu>

27 years agoPost-28 INSTALL updates
Andy Dougherty [Mon, 24 Feb 1997 22:09:09 +0000 (17:09 -0500)]
Post-28 INSTALL updates

Here are some more updates to the INSTALL file.  Specifically, I revised
the malloc section a bit to give more of an overall perspective (and,
admittedly, less detail).

p5p-msgid: <Pine.SOL.3.95q.970224170713.5700H-100000@fractal.lafayette.edu>
private-msgid: <Pine.SOL.3.95q.970224170713.5700H-100000@fractal.lafayette.e

27 years agoMore thoroughly test rand() and srand()
Tom Phoenix [Tue, 31 Dec 1996 17:59:20 +0000 (05:59 +1200)]
More thoroughly test rand() and srand()

27 years agoC<use vars> didn't work until 5.002
Chip Salzenberg [Tue, 19 Nov 1996 00:07:00 +0000 (12:07 +1200)]
C<use vars> didn't work until 5.002

27 years agoTrivial patch to make ExtUtils::Install more -w clean
Tim Bunce [Mon, 24 Feb 1997 16:05:17 +0000 (16:05 +0000)]
Trivial patch to make ExtUtils::Install more -w clean

p5p-msgid: <9702241605.AA17436@toad.ig.co.uk>

27 years agoPod::Text fixes
Roderick Schertler [Sat, 22 Feb 1997 18:03:08 +0000 (13:03 -0500)]
Pod::Text fixes

Here are some bug fixes for Pod::Text.

Is it a bug that text given to =for and =begin doesn't get indented?  It
seems like a bug to me, but I imagine one could argue the other way so I
wanted to test opinion before fixing it.

p5p-msgid: <350.856634588@eeyore.ibcinc.com>

27 years agoFix fd leak in IO::Pipe
Roderick Schertler [Sun, 23 Feb 1997 19:29:57 +0000 (14:29 -0500)]
Fix fd leak in IO::Pipe

Subject: Re: Another goof with IO::Pipe?

On Fri, 21 Feb 1997 18:51:10 -0500 (EST), Ilya Zakharevich <ilya@math.ohio-state.edu> said:
>
> IO::Pipe basically does:
>  $fh = $rw ? $me->reader() : $me->writer(); # close the other end
>         bless $io, "IO::Handle";
>         $io->fdopen($fh, $mode);
>
> In the corresponding place IPC::Open3 adds:
>
>  $fh->close;
>
> Which one is correct?

The ->close is necessary, the IO::Pipe code is leaving two copies of the
pipe in the child.  It dup()s the pipe onto STDIN/STDOUT, but doesn't
then close the pipe.

Here's an fd scan of a child whose parent did ->reader:

    FD  Type            Mode  UID       Inode    File System  FS Type  rdev
    0   char special    0620  roderick  686      13054,42068  dg/ux    23, 8
    1   FIFO special    0000  roderick  23         900,    0  dg/ux       --
    2   char special    0620  roderick  686      13054,42068  dg/ux    23, 8
    4   FIFO special    0000  roderick  23         900,    0  dg/ux       --

fd 4 is original write side of the pipe (the read side was fd 3) which
should have been closed by this point.

p5p-msgid: <pzn2sv722y.fsf@eeyore.ibcinc.com>

27 years agoDon't clobber $1 et al in debugger's DB::sub()
Ilya Zakharevich [Fri, 21 Feb 1997 14:53:46 +0000 (02:53 +1200)]
Don't clobber $1 et al in debugger's DB::sub()

27 years agoDon't assume that sizeof(int) >= sizeof(void*)
Chip Salzenberg [Fri, 21 Feb 1997 14:59:31 +0000 (02:59 +1200)]
Don't assume that sizeof(int) >= sizeof(void*)

27 years agoUpdate VMS version
Chip Salzenberg [Thu, 20 Feb 1997 16:52:53 +0000 (04:52 +1200)]
Update VMS version

27 years agoHints for DC/OSx
Stephen Zander [Mon, 24 Feb 1997 21:24:54 +0000 (13:24 -0800)]
Hints for DC/OSx

Subject: Re: Perl 5.003_28 is available

Chip Salzenberg wrote:
> This is our sixth (yes, _sixth_) beta candidate.  If I could give one
> piece of advice to any pumpkin holder, it would be:

Better late than never...

With the patch to hints/dcosx below 5.003_28 passes avarything
except test 9 of op/pack.  This seems to be a problem with
0xFFFFFFFF being treated as unsigned while
unpack("I",pack("I",0xFFFFFFFF) is signed.

./perl -e 'print 0xFFFFFFFF, "\n", unpack("I",pack("I",0xFFFFFFFF)),"\n";'
4294967295
-1

I'm assuming that gcvt is doing something strange.  Any suggestions/experiences?

p5p-msgid: <199702242124.NAA03796@wsuse5.mckesson.com>

27 years agoUltrix hints
Spider Boardman [Sat, 22 Feb 1997 09:51:48 +0000 (04:51 -0500)]
Ultrix hints

Subject: Re: Perl 5.003_28 is available

The hint file for ULTRIX is out of date with respect to the size
of the Perl_yylex() function.  I found this after enabling -O2,
which works in recent versions of the compiler.  Also, despite
the inability of the standard dist unit to detect the fact, the
ULTRIX dirent struct does too have a d_namelen field.

Patch for that attached.

Also, of course, as expected, ULTRIX still fails t/io/fs.t test
#25, since [f]truncate can't be used to extend a file on that OS.
All other tests are fine.

p5p-msgid: <199702220951.EAA08156@Orb.Nashua.NH.US>

27 years agoAutomatically call srand() before rand() if user didn't
Chip Salzenberg [Wed, 19 Feb 1997 19:04:03 +0000 (07:04 +1200)]
Automatically call srand() before rand() if user didn't

27 years ago[differences between cumulative patch application and perl5.003_28]
Chip Salzenberg [Fri, 21 Feb 1997 16:41:00 +0000 (04:41 +1200)]
[differences between cumulative patch application and perl5.003_28]

27 years ago[inseparable changes from patch from perl5.003_27 to perl5.003_28]
Perl 5 Porters [Fri, 21 Feb 1997 16:41:00 +0000 (04:41 +1200)]
[inseparable changes from patch from perl5.003_27 to perl5.003_28]

 CORE LANGUAGE CHANGES

Subject: Don't let C<sub foo;> undefine &foo
From: Chip Salzenberg <chip@perl.com>
Files: op.c

Subject: Make code, doc agree on $ENV{PATH} and `cmd`
From: Chip Salzenberg <chip@perl.com>
Files: pod/perlsec.pod pp_sys.c

Subject: Optimize keys() and values() in void context
From: Chip Salzenberg <chip@perl.com>
Files: doop.c op.c

 CORE PORTABILITY

Subject: VMS patches post _27
Date: Thu, 20 Feb 1997 01:58:46 -0500 (EST)
From: Charles Bailey <bailey@HMIVAX.HUMGEN.UPENN.EDU>
Files: MANIFEST dosish.h hv.c lib/ExtUtils/MM_VMS.pm lib/ExtUtils/xsubpp perl.c perlsdio.h pod/perldelta.pod pod/perlvar.pod t/op/closure.t unixish.h vms/Makefile vms/descrip.mms vms/ext/filespec.t vms/genconfig.pl vms/vms.c vms/vmsish.h

    private-msgid: <01IFMEMPN1IU0057E2@hmivax.humgen.upenn.edu>

Subject: Re: OS/2 patch for _27
Date: Thu, 20 Feb 1997 19:24:16 -0500 (EST)
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: INSTALL README.os2 lib/Test/Harness.pm os2/Changes os2/OS2/PrfDB/t/os2_prfdb.t os2/os2.c os2/os2ish.h os2/perl2cmd.pl perl.c pod/perldelta.pod t/TEST t/harness t/op/magic.t
Msg-ID: <199702210024.TAA03174@monk.mps.ohio-state.edu>

    (applied based on p5p patch as commit 833d3f255ed68b969f062cec63d33f853ed9237c)

 DOCUMENTATION

Subject: INSTALL updates since _26
Date: Tue, 18 Feb 1997 16:00:08 -0500 (EST)
From: Andy Dougherty <doughera@fractal.phys.lafayette.edu>
Files: INSTALL
Msg-ID: <Pine.SOL.3.95q.970218155815.2014F-100000@fractal.lafayette.e

    (applied based on p5p patch as commit a8247d96fd6167a3b920e63aedee5592cd6e29a7)

Subject: Document "$$0" change
From: Chip Salzenberg <chip@perl.com>
Files: pod/perldelta.pod

Subject: Don't recommend impossible //o for C<$x =~ $y>
From: Chip Salzenberg <chip@perl.com>
Files: pod/perlop.pod

Subject: Correct doc that claimed that <FH> was never false
From: Chip Salzenberg <chip@perl.com>
Files: pod/perldelta.pod pod/perlop.pod

Subject: Document C<$?> vs. $SIG{CHLD}
From: Ulrich Pfeifer <pfeifer@charly.informatik.uni-dortmund.de>
Files: pod/perlvar.pod

Subject: Add pumpkin.pod
From: Chip Salzenberg <chip@perl.com>
Files: MANIFEST Porting/pumpkin.pod

Subject: Don't say "associat*ve arr*y"
From: Chip Salzenberg <chip@perl.com>
Files: MANIFEST gv.h hv.c lib/Env.pm lib/overload.pm opcode.pl pod/perl.pod pod/perldelta.pod pod/perldiag.pod pod/perlfunc.pod pod/perlguts.pod pod/perlmod.pod pod/perltie.pod pod/perltoc.pod pod/perltrap.pod x2p/a2p.pod

 OTHER CORE CHANGES

Subject: Fix a typo
From: Chip Salzenberg <chip@perl.com>
Files: pp_sys.c

Subject: Fix perl_call_sv(..., G_NOARGS)
From: Chip Salzenberg <chip@perl.com>
Files: perl.c

Subject: Fix SIGSEGV when cloning sub with complex expression
From: Chip Salzenberg <chip@perl.com>
Files: op.c

27 years agoINSTALL updates since _26
Andy Dougherty [Tue, 18 Feb 1997 21:00:08 +0000 (16:00 -0500)]
INSTALL updates since _26

Here are my proposed updates to the INSTALL file.  Comments are welcome.
(I'd particularly welcome an AIX bos.* blurb from someone who understand
those things.)

It's not all that much, really; just some sections have been moved around
into what I hope is a more logical order.

p5p-msgid: <Pine.SOL.3.95q.970218155815.2014F-100000@fractal.lafayette.edu>
private-msgid: <Pine.SOL.3.95q.970218155815.2014F-100000@fractal.lafayette.e

27 years agoAdd test for grep() and wantarray
Hugo van der Sanden [Tue, 18 Feb 1997 11:05:59 +0000 (11:05 +0000)]
Add test for grep() and wantarray

Subject: Re: subscript when wantarray

In <199702172207.RAA20755@rio.atlantic.net>, Chip Salzenberg writes:
:First, here's a patch that makes the first case print "2"
[...]
:Second, the reason that it prints "2" instead of "3" is that
:C<(undef)[0]> evaluates to the empty list in an array context,
:but to C<undef> in a scalar context.  This is on purpose.
Yep, I expected that.

Patch worked fine here; below is an extra test using the same code as my
original example.

p5p-msgid: <199702181105.LAA17895@tyree.iii.co.uk>

27 years agoRe: OS/2 patch for _27
Ilya Zakharevich [Fri, 21 Feb 1997 00:24:16 +0000 (19:24 -0500)]
Re: OS/2 patch for _27

The bulk of this patch is README.os2 which has a lot of formatting
stripped (it is at the end). The rest (340 lines of -u) is quite simple.

INSTALL
t/harness documented.
lib/Test/Harness.pm
Can test files with -T on hash-bang line.
Gets more info from tests which coredump.
Will not try to load Devel::CoreStack on each test.
os2/Changes
Updated.
os2/OS2/PrfDB/t/os2_prfdb.t
File closed.
os2/os2.c
Poor-man's setup of environment if DLL run from non-conforming EXE.
os2/os2ish.h
Poor-man's setup of environment if DLL run from non-conforming EXE.
ALTERNATE_SHEBANG defined.
os2/perl2cmd.pl
Updated for ALTERNATE_SHEBANG.
perl.c
If ALTERNATE_SHEBANG defined, just ignore -S on the starting
line. (It might be already used by OS to start perl.)
pod/perldelta.pod
Malloc's PERL_DEBUG_MSTATS, EMERGENCY_SBR, PACK_MALLOC,
TWO_POT_OPTIMIZE documented.
t/harness
Does not require -I../lib any more.
t/op/magic.t
Now works under OS/2 with both tests.
t/TEST
Adds an advice to use
perl harness
to get better granularity.
README.os2
A lot of markup stripped (in the hope that smart converters
will add it back), minor updates.

p5p-msgid: <199702210024.TAA03174@monk.mps.ohio-state.edu>

27 years agoTweak tests to notice $dont_use_nlink
Chip Salzenberg [Mon, 17 Feb 1997 22:29:53 +0000 (10:29 +1200)]
Tweak tests to notice $dont_use_nlink

27 years agoDon't run locale test if -DNO_LOCALE
Chip Salzenberg [Mon, 17 Feb 1997 22:22:49 +0000 (10:22 +1200)]
Don't run locale test if -DNO_LOCALE

27 years agoInclude 'study' in regexp.t
Chip Salzenberg [Wed, 19 Feb 1997 22:43:28 +0000 (10:43 +1200)]
Include 'study' in regexp.t

27 years agoprototype error in File::stat
Graham Barr [Tue, 18 Feb 1997 07:48:40 +0000 (07:48 +0000)]
prototype error in File::stat

The prototype for lstat in File::stat is wrong. It stats that is expects
a GLOB. It should be expecting a scalar, the same as File::stat::stat.

p5p-msgid: <199702180748.HAA14151@ultra-boy>

27 years agoRemove redundant clearerr() from IO::Seekable
Chip Salzenberg [Tue, 18 Feb 1997 00:25:45 +0000 (12:25 +1200)]
Remove redundant clearerr() from IO::Seekable

27 years ago[dummy merge]
Chip Salzenberg [Fri, 21 Feb 1997 16:41:00 +0000 (04:41 +1200)]
[dummy merge]

This merge exists so that the p5p version of the patch and the applied
version are both in the history

27 years agoAvoid $` $& $' in libraries
Chip Salzenberg [Fri, 21 Feb 1997 16:41:00 +0000 (04:41 +1200)]
Avoid $` $& $' in libraries

(this is the same change as commit 2724d5068a405436d3f2cd00aeb8f7b460b24fec, but as applied)

27 years agoAvoid $` $& $' in libraries
Ilya Zakharevich [Fri, 21 Feb 1997 02:07:30 +0000 (21:07 -0500)]
Avoid $` $& $' in libraries

Subject: Pessimal $` $& $' in libraries

Corrected:

lib/diagnostics.pm
lib/Getopt/Long.pm
lib/Pod/Text.pm
os2/OS2/REXX/REXX.pm

$' $` $& eliminated from all the known *.pm.

Btw, by this I proudly introduce usage of $+ in standard perl modules ;-).

p5p-msgid: <199702210207.VAA03560@monk.mps.ohio-state.edu>

27 years agoDebugger patch
Ilya Zakharevich [Fri, 21 Feb 1997 07:37:59 +0000 (02:37 -0500)]
Debugger patch

private-msgid: <199702210737.CAA03951@monk.mps.ohio-state.edu>

27 years agoFix the Tolkein quote
Chip Salzenberg [Tue, 18 Feb 1997 01:29:35 +0000 (13:29 +1200)]
Fix the Tolkein quote

27 years agoMinor update to malloc.c
Ilya Zakharevich [Fri, 21 Feb 1997 02:44:13 +0000 (21:44 -0500)]
Minor update to malloc.c

This change of defines may satisfy Nick in his future life, when he
needs an effecient malloc() with DEBUGGING perl again:

p5p-msgid: <199702210244.VAA03676@monk.mps.ohio-state.edu>

27 years agoUndo signal patch -- it broke die() in signal
Chip Salzenberg [Tue, 18 Feb 1997 00:20:07 +0000 (12:20 +1200)]
Undo signal patch -- it broke die() in signal

27 years agoNo version of AIX has working setre[ug]id()
Keith Neufeld [Tue, 18 Feb 1997 04:49:16 +0000 (16:49 +1200)]
No version of AIX has working setre[ug]id()

27 years agoNew hints for Digital UNIX
Jarkko Hietaniemi [Tue, 18 Feb 1997 02:25:14 +0000 (14:25 +1200)]
New hints for Digital UNIX

27 years agoWhen overloading, don't throw away nomethod's value
Ilya Zakharevich [Tue, 18 Feb 1997 21:22:53 +0000 (09:22 +1200)]
When overloading, don't throw away nomethod's value

27 years agoTurn off 'expression tainted' flag at end of runops()
Chip Salzenberg [Tue, 18 Feb 1997 21:08:02 +0000 (09:08 +1200)]
Turn off 'expression tainted' flag at end of runops()

27 years agoDon't taint $x in C<$x = ($tainted =~ /(\w+)/)>
Chip Salzenberg [Wed, 19 Feb 1997 21:58:01 +0000 (09:58 +1200)]
Don't taint $x in C<$x = ($tainted =~ /(\w+)/)>

27 years ago[shell changes from patch from perl5.003_27 to perl5.003_28]
Chip Salzenberg [Fri, 21 Feb 1997 16:41:00 +0000 (04:41 +1200)]
[shell changes from patch from perl5.003_27 to perl5.003_28]

Change from running these commands:

 # VMS no longer uses 'make' to build Perl
 rm -f vms/Makefile

 # ready to patch
 exit 0

27 years ago[differences between cumulative patch application and perl5.003_27]
Chip Salzenberg [Tue, 18 Feb 1997 01:22:00 +0000 (13:22 +1200)]
[differences between cumulative patch application and perl5.003_27]

27 years ago[inseparable changes from patch from perl5.003_26 to perl5.003_27]
Perl 5 Porters [Tue, 18 Feb 1997 01:22:00 +0000 (13:22 +1200)]
[inseparable changes from patch from perl5.003_26 to perl5.003_27]

 BUILD PROCESS

Subject: Fix eval "" in Configure
Date: Fri, 14 Feb 1997 13:09:53 -0500
From: John L. Allen <allen@gateway.grumman.com>
Files: Configure

    Subject: Re: Configure problem on IRIX - me dumb

    p5p-msgid: <9702141809.AA17001@gateway.grumman.com>

Subject: Don't link with -lsfio if sfio is not requested
From: Chip Salzenberg <chip@perl.com>
Files: Configure

Subject: perl5.003_26 Configure change "win" for AIX 4
Date: Fri, 14 Feb 1997 13:59:02 -0600 (CST)
From: Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu>
Files: Configure

    p5p-msgid: <Pine.OSF.3.95.970214135751.32654A-100000@dogbert.cc.ndsu.NoDak.edu>
    private-msgid: <Pine.OSF.3.95.970214135751.32654A-100000@dogbert.cc.ndsu.NoD

 CORE LANGUAGE CHANGES

Subject: Better looks_like_number() function [sv.c]
Date: Fri, 14 Feb 1997 18:08:52 +0100
From: Gisle Aas <aas@bergen.sn.no>
Files: sv.c
Msg-ID: <199702141708.SAA17546@bergen.sn.no>

    (applied based on p5p patch as commit 8dbaa58ee2aba7cc22d84199a674c58bbf108b46)

Subject: Remove redundant functions UNIVERSAL::{class,is_instance}
Date: 14 Feb 1997 15:52:21 +0000
From: Gisle Aas <aas@bergen.sn.no>
Files: pod/perldelta.pod pod/perlobj.pod t/op/universal.t universal.c
Msg-ID: <hwwsbpeq2.fsf@bergen.sn.no>

    (applied based on p5p patch as commit 77bb9b23081b62119e8fbe9f5655b8802e4537ae)

Subject: Allow C<setpgrp $$>
Date: 16 Feb 1997 23:19:12 -0500
From: Roderick Schertler <roderick@gate.net>
Files: pp_sys.c
Msg-ID: <pzraigyshr.fsf@eeyore.ibcinc.com>

    (applied based on p5p patch as commit 3d2573a84a1aa655d5da58c57b3fc20e04d40f9f)

Subject: Fix syntax error on C<&$1>
From: Chip Salzenberg <chip@perl.com>
Files: toke.c

Subject: Fix grep() with refs in array context
From: Chip Salzenberg <chip@perl.com>
Files: pp.c

 CORE PORTABILITY

Subject: Eliminate $^S; add C<use vmsish qw(status exit time)>
Date: Mon, 17 Feb 1997 02:45:26 -0500 (EST)
From: Charles Bailey <bailey@HMIVAX.HUMGEN.UPENN.EDU>
Files: MANIFEST gv.c lib/English.pm lib/ExtUtils/MM_VMS.pm lib/ExtUtils/Mksymlists.pm lib/ExtUtils/xsubpp mg.c op.c perl.c perl.h pod/perldelta.pod pod/perlmod.pod pod/perlvar.pod pp_ctl.c pp_sys.c utils/perldoc.PL vms/Makefile vms/config.vms vms/descrip.mms vms/ext/Stdio/Stdio.pm vms/ext/Stdio/Stdio.xs vms/ext/XSSymSet.pm vms/ext/vmsish.pm vms/vms.c vms/vmsish.h win32/makedef.pl

    private-msgid: <01IFI9CFKL0S004R2V@hmivax.humgen.upenn.edu>

 LIBRARY AND EXTENSIONS

Subject: Remove Fatal.pm
From: Chip Salzenberg <chip@perl.com>
Files: MANIFEST lib/Fatal.pm pod/perldelta.pod pod/perlmod.pod pod/roffitall t/lib/fatal.t

Subject: Refresh MakeMaker to 5.40
From: Andy Dougherty <doughera@lafcol.lafayette.edu>
Files: lib/ExtUtils/Liblist.pm lib/ExtUtils/MM_Unix.pm lib/ExtUtils/MakeMaker.pm lib/ExtUtils/Mksymlists.pm

 OTHER CORE CHANGES

Subject: Fix core dump when embedding
From: Chip Salzenberg <chip@perl.com>
Files: perl.c

Subject: Re: Fragile signals
Date: Thu, 13 Feb 1997 01:44:39 -0500 (EST)
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: mg.c
Msg-ID: <199702130644.BAA07572@monk.mps.ohio-state.edu>

    (applied based on p5p patch as commit 09df8c7df7dfc9853902f1fdd8a6d95f53be66fc)

Subject: Make format strings correspond exactly to parameters
Date: 13 Feb 1997 17:24:31 -0500
From: Roderick Schertler <roderick@gate.net>
Files: doio.c ext/DB_File/DB_File.xs ext/Opcode/Opcode.xs gv.c op.c perl.c pp_ctl.c pp_sys.c regcomp.c toke.c
Msg-ID: <pz7mkc1h0g.fsf@eeyore.ibcinc.com>

    (applied based on p5p patch as commit bf81aadd817bdea29720b072eef945df2da8463b)

Subject: Don't try to attach 'o' magic to read-only values
From: Chip Salzenberg <chip@perl.com>
Files: sv.c

Subject: Fix carriage-return message
From: Chip Salzenberg <chip@perl.com>
Files: toke.c

Subject: In <=>, test for equality first
From: Chip Salzenberg <chip@perl.com>
Files: pp.c

Subject: Don't mark sv_{true,false} PADTMP
From: Chip Salzenberg <chip@perl.com>
Files: op.c

27 years agoMake format strings correspond exactly to parameters
Roderick Schertler [Thu, 13 Feb 1997 22:24:31 +0000 (17:24 -0500)]
Make format strings correspond exactly to parameters

Subject: Re: Hereis weirdness in 5.003_26

On Thu, 13 Feb 1997 15:13:18 -0500 (EST), Chip Salzenberg <chip@atlantic.net> said:
> According to Ilya Zakharevich:
>>
>> Why was not it catched by gcc prototypes?
>
> I don't use gcc.  Does anyone who uses gcc compile with -Wproto ?

-Wformat, you mean.  I wasn't previously, but I will in the future.  It
turned up a few bugs (some universal, some which would trigger only
where I32 != int or the like).  I think -Wformat should be added to
CFLAGS automatically when appropriate.

p5p-msgid: <pz7mkc1h0g.fsf@eeyore.ibcinc.com>

27 years agoFix formatting in perldiag
Chip Salzenberg [Mon, 17 Feb 1997 05:58:57 +0000 (17:58 +1200)]
Fix formatting in perldiag

27 years agoRe: Fragile signals
Ilya Zakharevich [Thu, 13 Feb 1997 06:44:39 +0000 (01:44 -0500)]
Re: Fragile signals

Ilya Zakharevich writes:
>
> Some messages of mine get stuck in the chain:
>
>         Nico => Debian => floor
>
> I repost them now:
>
> As you may remember, the following program
>
> #!/usr/bin/perl
> sub foo { ++$a }
> $SIG{ALRM} = sub { $run = 0 ; return };
> while (1) {
>   $a = 0;
>   $run = 1;
>   alarm 1;
>   foo while $run;
>   print $i++, ": $a\n";
> }
> die "Not reached";
> __END__
>
> will not call malloc/realloc/free in sighandler, but nevertheless it
> would not run more than approx. 25 iterations. This shows that 4% of
> subroutine-calling frame is not signal-safe.
>
> The following patch makes it run approx. twice as long (on _17) (thus

Well, the following patch makes it run 4 times as long. The only
failure mode I have seen was silent walking off runops().

Here is the statistics (number of iterations before failure):

   Before:
                23, 24, 14, 22, 19, 13, 30, 11
        (4 core dumps, 2 attempts to free unref scalar, 1 panic:
                leave-scope, 1 silent)
   After:
                42, 936, 63, 131, 280.
        (all silent)

It looks like handling of internal stacks needs a lot of
improvement. What this patch does is only a symptomatic treatment, but
it may go into 004 nevertheless (especially if one can see why op
becomes 0 at some moment). I tried saving/restoring op in the handler,
but this does not help...

p5p-msgid: <199702130644.BAA07572@monk.mps.ohio-state.edu>

27 years agoUpdate To-Do list
Tim Bunce [Mon, 10 Feb 1997 19:00:59 +0000 (19:00 +0000)]
Update To-Do list

Subject: Re: Perl Futures (was: Tentative 5.005 and 5.006 schedules)

p5p-msgid: <9702101900.AA25293@toad.ig.co.uk>

27 years agoAllow C<setpgrp $$>
Roderick Schertler [Mon, 17 Feb 1997 04:19:12 +0000 (23:19 -0500)]
Allow C<setpgrp $$>

Subject: Re: Perl question, re: POSIX setpgrp

On Fri, 14 Feb 1997 16:31:53 GMT, Chris Vo <chrisv@on.bell.ca> said:
>
>>> POSIX setpgrp can't take an argument at ./check_ntp line 21.
>
> where line 21 reads as:
>       setpgrp (0, $$);

Replace that with just

        setpgrp;

This does the same thing and will work on all systems.

I think there's a bug here.  Perl is expecting a pgrp arg of 0 to mean
$$, but it doesn't allow an explicit $$.  Even this seems a little odd,
as on neither system I've got available at the moment is it mentioned
that a 0 pgrp means getpid() (they both mention that a 0 pid means
getpid(), though).

p5p-msgid: <pzraigyshr.fsf@eeyore.ibcinc.com>

27 years agopod2man: missing '-' in name section shouldn't be fatal
Ulrich Pfeifer [Mon, 10 Feb 1997 17:38:45 +0000 (18:38 +0100)]
pod2man: missing '-' in name section shouldn't be fatal

Subject: pod2man: missing '-' in name section is fatal

p5p-msgid: <yfmzpxcimsa.fsf@ls6.informatik.uni-dortmund.de>

27 years agoRemove redundant functions UNIVERSAL::{class,is_instance}
Gisle Aas [Fri, 14 Feb 1997 15:52:21 +0000 (15:52 +0000)]
Remove redundant functions UNIVERSAL::{class,is_instance}

Nick Ing-Simmons <nik@tiuk.ti.com> writes:

> Loose them!

p5p-msgid: <hwwsbpeq2.fsf@bergen.sn.no>

27 years agoRemove non-portable locale tests
Chip Salzenberg [Mon, 17 Feb 1997 05:59:16 +0000 (17:59 +1200)]
Remove non-portable locale tests

27 years agoBetter looks_like_number() function [sv.c]
Gisle Aas [Fri, 14 Feb 1997 17:08:52 +0000 (18:08 +0100)]
Better looks_like_number() function [sv.c]

The old looks_like_number() function was too forgiving.  It classified
strings like these as numbers:

   "+"
   "+e+"
   "e3"
   "."

With this patch perl -lwe 'print int(".")' prints:

   Argument "." isn't numeric in int at -e line 1.
   0

as it should and perl still passes all its tests.

p5p-msgid: <199702141708.SAA17546@bergen.sn.no>

27 years agoRefresh Test::Harness to 1.15
Andreas Koenig [Mon, 17 Feb 1997 05:59:16 +0000 (17:59 +1200)]
Refresh Test::Harness to 1.15

27 years agoRefresh CPAN.pm to 1.21
Andreas Koenig [Mon, 17 Feb 1997 05:59:13 +0000 (17:59 +1200)]
Refresh CPAN.pm to 1.21