Andreas Koenig [Fri, 10 Jan 1997 12:43:39 +0000 (13:43 +0100)]
Re: MakeMaker and 'make uninstall'
>>>>> Hugo van der Sanden writes:
> I tried uninstalling a package today, the first time I've done so. The
> package was MailFolder0.03, and as far as I can see the uninstall:
> target in the Makefile was provided by MakeMaker. This does:
[...]
> and FULLEXT is defined as 'Mail'. The result was that most of the
> MailTools package was also removed.
Thanks for the report and _Sorry_.
Folks, I think, I have to deprecate the uninstall target. It was never
documented, it was never really clever, it just caused me a lot of
grief. I'd rather design a new solution from scratch.
If nobody objects, I'd recommend the following patch to
lib/ExtUtils/MM_Unix.pm. It makes use of the "nonono" switch of the
uninstall routine and will not uninstall anything.
Chip, can you sneak that in, please? Thanks!
p5p-msgid: <
199701101243.NAA26400@anna.in-berlin.de>
Ilya Zakharevich [Wed, 8 Jan 1997 08:25:47 +0000 (03:25 -0500)]
Re: xsubpp and Tk ==> segfault
Ilya Zakharevich writes:
>
> I recall that the reason why void code may be marked as XSRETURN(1) is
> that XS TUTORIAL recommended to mark XSUB as void if you manually put
It is not perlxstut, it is perlxs.
> in stack manipulation! I do not know why this horrible practice is
> blessed, but we need a workaround which would allow this mess, and
> would not make an XSUB
[Add: ] ... segfault if some Perl code was called in between.
As I found out, the only difference between an XSUB with a CODE:
section which is declared `void' and one which is declared "SV *" is
that the first one does not contain a line:
SV * RETVAL;
Is there any trouble if such a line is present?
If nobody will strongly object to it, I will remove an advice to
mark a CODE:-XSUBs as `void' from the POD(s).
I can also add a warning to xsubpp if a `void' XSUB sets ST(*). What
about it? (Warning will be issued only once per .xs .)
> Solution: Make the check for CODE section stricter: check for actuall
> occurence of "ST(" (with embedded spaces) before making the decision
> to XSRETURN(1).
The following patch is checked with Perl 3_17 and Tk (it checks for ST
at the LHS as an assignment):
Enjoy,
p5p-msgid: <
199701080825.
DAA15813@monk.mps.ohio-state.edu>
Paul Marquess [Wed, 8 Jan 1997 16:55:02 +0000 (16:55 +0000)]
low priority patches
this set of patches isn't super critical, but it does tidy a few things
up in anticipation of lexical warnings.
p5p-msgid: <
9701081655.AA27349@claudius.bfsec.bt.co.uk>
Chip Salzenberg [Tue, 14 Jan 1997 04:01:03 +0000 (16:01 +1200)]
Avoid "uninitialized" warnings from POSIX::constant()
Gurusamy Sarathy [Tue, 14 Jan 1997 18:15:25 +0000 (13:15 -0500)]
Re: Uninitialized value in Carp.pm ?
private-msgid: <
199701141815.NAA07960@aatma.engin.umich.edu>
Chip Salzenberg [Wed, 15 Jan 1997 19:24:00 +0000 (07:24 +1200)]
[dummy merge]
This merge exists so that the p5p version of the patch and the applied
version are both in the history
Chip Salzenberg [Wed, 15 Jan 1997 19:24:00 +0000 (07:24 +1200)]
PATCH: AutoSplit
(this is the same change as commit
bb8fceff88bc3fe9e820d0761f1b0451a870ac65, but as applied)
Graham Barr [Mon, 11 Mar 1996 11:01:58 +0000 (06:01 -0500)]
PATCH: AutoSplit
Below is a patch which moves where AutoSplit splits a file. Currently AutoSplit
split when it sees a sub. This inhibits individual routines from having their
own use or require statements.
This patch causes AutoSplit to split at the last line starting with
a } before a sub.
The reason behind this is that I have a package which is a collection
of utility routines each requiring separate packages, as I am trying
to reduce what is loaded by autoloading the routines, it seems only
sensible to only 'use' packages when the routine which needs them it
loaded
p5p-msgid: <
9603111010.AA29935@tiuk.ti.com>
Roderick Schertler [Thu, 9 Jan 1997 04:45:58 +0000 (23:45 -0500)]
Re: FileCache::cacheout clobbers $_
This bug report was from last July, but the bug is still there.
FileCache.pm and cacheout.pl clobber $_.
p5p-msgid: <pz3ewb3189.fsf@eeyore.ibcinc.com>
Paul Marquess [Tue, 14 Jan 1997 12:47:40 +0000 (12:47 +0000)]
Refresh DB_File to 1.10
this patch works around an incompatability that was introduced in
Berkeley DB 1.86
p5p-msgid: <
9701141247.AA21242@claudius.bfsec.bt.co.uk>
Johan Vromans [Fri, 10 Jan 1997 18:22:36 +0000 (06:22 +1200)]
Refresh Getopt::Long to 2.6
Graham Barr [Mon, 13 Jan 1997 02:07:59 +0000 (14:07 +1200)]
Fix for fd leak in IO::File::new_tmpfile
Signed-off-by: Graham Barr <gbarr@ti.com>
Paul Marquess [Sun, 12 Jan 1997 15:09:33 +0000 (15:09 +0000)]
patch for db-recno.t
Chip,
here is a patch for db-recno.t which makes it print a diagnostic
message when any of test 51,53 or 55 fail.
p5p-msgid: <
9701121509.AA11147@claudius.bfsec.bt.co.uk>
Chip Salzenberg [Wed, 15 Jan 1997 19:24:00 +0000 (07:24 +1200)]
[dummy merge]
This merge exists so that the p5p version of the patch and the applied
version are both in the history
Chip Salzenberg [Wed, 15 Jan 1997 19:24:00 +0000 (07:24 +1200)]
Advice on TEST failure
(this is the same change as commit
1ce51ea5e928b0e6fe3d33c847a736aaee083a49, but as applied)
Dominic Dunlop [Fri, 10 Jan 1997 09:19:07 +0000 (10:19 +0100)]
Advice on TEST failure
At 18:00 +0100 1997-01-09, Paul Marquess wrote:
>What are peoples views on this one. Should I simply write a diagnostic
>message to STDERR when test 51 fails? Some thing like this...
>
> base/cond.........ok
> lib/db-btree......ok
> lib/db-hash.......ok
> lib/db-recno...... not ok 51
> There is a known problem with older versions of Berkeley DB. You
> can ignore the errors if you're never going to use the broken
> functionality (recno databases with a modified bval), otherwise
> you'll have to upgrade your DB library or OS. If you feel like
> upgrading, the most recent version is 1.85. Check out
> http://www.bostic.com/db for more details.
Looks good to me. As of 5.003_20, hints/broken-db.msg is a shell scriptlet
that contains a similar (but not quite as useful) message for use by
Configure. It Would Be Nice if the same text (the same file?) were used in
both situations. But no big deal.
And, shortly afterwards, Tom Phoenix wrote:
>You're right. Who needs fine print? :-) What if 'make test' wrapped up
>with a message like this one?
>
> Failed 1 test, 96.55% okay.
> u=1.13 s=3.62 cu=60.49 cs=73.18 files=140 tests=3165
>
> ### Since not all tests were successful, you may want to run
> ### some tests individually to see the diagnostic messages,
> ### if any. See the INSTALL document's section on "make test".
That's a good idea, too, whether or not lib/db-recno blats something onto
the screen even when TEST is run without -v. Here's a patch which
implements it. (Well, with slightly different ordering.) I've also taken
the liberty of removing the RCS information from TEST, as it's meaningless
by now, of making the exit behaviour the same for one and for multiple
failures, and of trying to make the difference between "files" and "tests"
clearer. (This always confused me: I thought maybe "files" was the number
of files that got touched in the course of the tests.)
p5p-msgid: <v03010d01aefbaefcf3bc@[194.51.248.78]>
Chip Salzenberg [Mon, 13 Jan 1997 03:35:33 +0000 (15:35 +1200)]
Make installperl quieter; only shared libraries need 0555
Chip Salzenberg [Sat, 11 Jan 1997 04:26:37 +0000 (16:26 +1200)]
Fix infinite loop for undef function in @SIG{__WARN__,__DIE__}
Chip Salzenberg [Mon, 13 Jan 1997 01:17:21 +0000 (13:17 +1200)]
Fail regex that starts with '{'
Chip Salzenberg [Wed, 15 Jan 1997 19:24:00 +0000 (07:24 +1200)]
[shell changes from patch from perl5.003_20 to perl5.003_21]
Change from running these commands:
# this file is obsolete
rm -f hints/aux.sh
# new (and old) tests and scripts must be executable
touch t/op/cmp.t t/op/universal.t pod/rofftoc
chmod +x t/op/cmp.t t/op/universal.t pod/rofftoc pod/roffitall
# ready to patch
exit 0
Chip Salzenberg [Tue, 7 Jan 1997 23:52:00 +0000 (11:52 +1200)]
[differences between cumulative patch application and perl5.003_20]
Perl 5 Porters [Tue, 7 Jan 1997 23:52:00 +0000 (11:52 +1200)]
[inseparable changes from patch from perl5.003_19 to perl5.003_20]
BUILD PROCESS
Subject: Make Configure default to the first domain in /etc/resolv.conf
From: Chip Salzenberg <chip@atlantic.net>
Files: Configure
Subject: Start all helper scripts with $startsh
From: Chip Salzenberg <chip@atlantic.net>
Files: Configure
Subject: Support libperl.so under FreeBSD
Date: Sun, 5 Jan 1997 22:41:49 +0100
From: Ollivier Robert <roberto@keltia.freenix.fr>
Files: Configure Makefile.SH
Msg-ID: <Mutt.
19970105224149.roberto@keltia.freenix.fr>
(applied based on p5p patch as commit
b126116e5ae3d57fa007f8a42fd506805b35163b)
CORE LANGUAGE CHANGES
Subject: Rescind named closures
From: Chip Salzenberg <chip@atlantic.net>
Files: Makefile.SH op.c perly.c perly.c.diff perly.y pp_hot.c
Subject: Fix: empty @_ when calling empty-proto subs without parens
Date: Sat, 04 Jan 1997 10:29:04 +0000
From: Graham Barr <bodg@tiuk.ti.com>
Files: perly.c perly.y
(applied based on p5p patch as commit
3112f5de73952f91aa4e8005d9852dfddbcf0402)
CORE PORTABILITY
Subject: Configure/perl5/Compartmented Mode Workstation (fwd)
Date: Mon, 06 Jan 1997 13:15:38 -0500 (EST)
From: Andy Dougherty <doughera@fractal.phys.lafayette.edu>
Files: Configure hints/dec_osf.sh
private-msgid: <Pine.SOL.3.95.
970106131505.1662C-100000@fractal.lafayette.ed
Subject: Remove obsolete file "dl_os2.xs".
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: MANIFEST
DOCUMENTATION
Subject: tiny doc patches
Date: Sat, 04 Jan 1997 11:12:13 -0500
From: Roderick Schertler <roderick@gate.net>
Files: pod/perlapio.pod pod/perlnews.pod pod/perltoc.pod
Msg-ID: <23338.
852394333@eeyore.ibcinc.com>
(applied based on p5p patch as commit
b681178584626ba3718f1279845fd452317134c1)
Subject: doc patch for defined on perlfunc.pod
Date: 04 Jan 1997 21:28:30 -0500
From: Roderick Schertler <roderick@gate.net>
Files: pod/perlfunc.pod
Msg-ID: <pz91686ek1.fsf@eeyore.ibcinc.com>
(applied based on p5p patch as commit
38e3adfd2e3d40b46e465482945c4f3de4bb50ef)
Subject: doc patch: perldsc
Date: 04 Jan 1997 21:25:58 -0500
From: Roderick Schertler <roderick@gate.net>
Files: pod/perldsc.pod pod/perltoc.pod
Msg-ID: <pzafqo6eo9.fsf@eeyore.ibcinc.com>
(applied based on p5p patch as commit
4d42f92e5bf79556508016b7af91233b12e526eb)
Subject: scalar caller doc fix
Date: Mon, 06 Jan 1997 22:34:20 -0500
From: Roderick Schertler <roderick@gate.net>
Files: pod/perlfunc.pod
Msg-ID: <18245.
852608060@eeyore.ibcinc.com>
(applied based on p5p patch as commit
218104faecb0ec19e0f4f89e084959e757a5230f)
Subject: Misc perlfunc updates
From: Tom Christiansen <tchrist@mox.perl.com>
Files: pod/perlfunc.pod pod/perltoc.pod
LIBRARY AND EXTENSIONS
Subject: sigaction() problems
Date: Mon, 06 Jan 1997 15:42:04 -0500
From: Roderick Schertler <roderick@gate.net>
Files: ext/POSIX/POSIX.pm ext/POSIX/POSIX.pod
Msg-ID: <12808.
852583324@eeyore.ibcinc.com>
(applied based on p5p patch as commit
84e96f2bcc509ba2fb5d2c9608a30cc3cfdea41a)
Subject: Fix importation of FileHandle methods; fix POSIX docs
From: Chip Salzenberg <chip@atlantic.net>
Files: ext/POSIX/POSIX.pm ext/POSIX/POSIX.pod lib/FileHandle.pm
Subject: Patch: make hints files warn about db-recno failures
Date: Sun, 5 Jan 1997 12:34:25 +0100
From: Dominic Dunlop <domo@slipper.ip.lu>
Files: MANIFEST hints/aux.sh hints/broken-db.msg hints/freebsd.sh
(applied based on p5p patch as commit
692df45da95e2b7d14c4560347ef4555bb40b621)
OTHER CORE CHANGES
Subject: Fix C< sub foo (&@); sub bar (&); foo {}, bar {}, bar {} >
From: Chip Salzenberg <chip@atlantic.net>
Files: perly.c perly.c.diff perly.y
Subject: plug for safe/opcode leaks
Date: Tue, 07 Jan 1997 17:20:46 -0500
From: Doug MacEachern <dougm@osf.org>
Files: op.c
Msg-ID: <
199701072220.RAA02117@postman.osf.org>
(applied based on p5p patch as commit
5cbfc2849d37f748a8facbcbf1c889c575943488)
Subject: Fix Dynaloader failures with DProf
Date: Mon, 06 Jan 1997 12:18:46 -0500
From: Gurusamy Sarathy <gsar@engin.umich.edu>
Files: pp_hot.c
private-msgid: <
199701061718.MAA26909@aatma.engin.umich.edu>
TESTS
Subject: New test: comp/proto.t
Date: Mon, 06 Jan 1997 09:13:03 +0000
From: Graham Barr <bodg@tiuk.ti.com>
Files: MANIFEST t/comp/proto.t
(applied based on p5p patch as commit
8c1635e65dc1b3900503d444e985e3f0e5601454)
Graham Barr [Sun, 5 Jan 1997 20:13:03 +0000 (20:13 +0000)]
New test: comp/proto.t
Chip Salzenberg wrote:
>
> According to Graham Barr:
> > BTW: I am working on a test (comp/proto.t) for this, and
> > prototypes in general.
>
> Thanks for telling me -- saved me some work.
OK, here is what I have done for comp/proto.t. More could probably
be added, but where do you stop.
I did find something weird though which caused the script to fail
just by changing the order of definitions ???
Chip Salzenberg [Tue, 7 Jan 1997 23:52:00 +0000 (11:52 +1200)]
[dummy merge]
This merge exists so that the p5p version of the patch and the applied
version are both in the history
Dominic Dunlop [Sat, 4 Jan 1997 22:34:25 +0000 (23:34 +0100)]
Patch: make hints files warn about db-recno failures
Herewith some minor Configure and hints patches against 5.003_19:
1. Amend Configure so that MachTen's antediluvian test, which holds that
-1 is not an integer, does not bleat during non-blocking IO tests. (Tom
Christiansen posted a while back about several instances of this sort of
thing; this patch addresses the only one I can see. Are there
still others?)
2. Make hints files for A/UX, FreeBSD, and MachTen give notification
of the db-recno test failures caused by these systems' old db
libraries. This patch due mainly to Roderick Schertler, who implemented
a suggestion I made. I hope it's OK to put a shared stub in hints/. It
has the .shx extension so that it doesn't show up in Configure's list of
systems for which there are hints. It may be that the hints for some
version of IRIX need the same tweak, but I couldn't work out which one
from Martijn Koster's recent posting, and so left well enough alone.
p5p-msgid: <v03010d00aef53ac4d18a@[194.51.248.68]>
Chip Salzenberg [Tue, 7 Jan 1997 23:52:00 +0000 (11:52 +1200)]
perlpod.pod possible patches
(this is the same change as commit
dd21fd8df4b58122c680bb0b8e543f75fb0fcd93, but as applied)
Larry W. Virden [Tue, 7 Jan 1997 12:56:30 +0000 (07:56 -0500)]
perlpod.pod possible patches
I am discouraged because I have pretty much determined that a current
incompatibility between Mail::Mailer, perl 5.003_19, and perlbug has
resulted in a number of patchs and bug reports that I have sent over the
past week appear to have disappeared into a sink hole. Anyways, I will
attempt to do this by hand. The following is a patch to perlpod.pod to
fix some things that are not formatting properly or which appear to be
incomplete thoughts.
p5p-msgid: <
9701070756.AA1185@cas.org>
Signed-off-by: x2487 <lvirden@cas.org>
Graham Barr [Fri, 3 Jan 1997 21:29:04 +0000 (21:29 +0000)]
Fix: empty @_ when calling empty-proto subs without parens
Graham Barr wrote:
<snip>
> OK, so I thought that prototypes in Socket.pm (which is probably
> a good idea) could help here, but... guess what the following
> script outputs
>
> sub fred ()
> {
> warn join(" ",@_);
> }
>
> fred;
> @_ = qw(a b c);
> &fred;
> fred;
> __END__
>
> It outputs ...
>
> Warning: something's wrong at yyy line 3.
> a b c at yyy line 3.
> a b c at yyy line 3.
>
> OK I can belive the second one as it has the & style call (which
> the pod states overrides prototypes) but I was very amazed at
> the last call.
>
> Is this a bug ?? I think so as the pod suggests that a sub defined
> as
>
> sub mytime ();
>
> and a statement such as
>
> mytime +2;
>
> will by the same as
>
> mytime() + 2
>
> but this is not the case, it seems it would be
>
> &mytime(@_) + 2
>
OK, here is a patch which fixes this. It requires a modification
to perly.y and perly.c (I cannot re-generate perly.c here
but it is a simple fix)
it chamges the output of the above script to
Warning: something's wrong at yyy line 3.
Warning: something's wrong at yyy line 3.
a b c at yyy line 3.
Chip Salzenberg [Tue, 7 Jan 1997 23:52:00 +0000 (11:52 +1200)]
[dummy merge]
This merge exists so that the p5p version of the patch and the applied
version are both in the history
Roderick Schertler [Tue, 7 Jan 1997 03:34:20 +0000 (22:34 -0500)]
scalar caller doc fix
The second bit is due to Tom Phoenix, I stuck it in since I was in there
already.
p5p-msgid: <18245.
852608060@eeyore.ibcinc.com>
Chip Salzenberg [Tue, 7 Jan 1997 23:52:00 +0000 (11:52 +1200)]
Re: constant function inlining
(this is the same change as commit
f0893e72a8aa916a77953ff2f1847494d1527fb8, but as applied)
Roderick Schertler [Tue, 7 Jan 1997 20:27:50 +0000 (15:27 -0500)]
Re: constant function inlining
On Tue, 7 Jan 1997 18:21:28 +0000, Tim Bunce <Tim.Bunce@ig.co.uk> said:
>
> A few words about redefining [constant] functions would be worthwhile.
Right, patch attached.
p5p-msgid: <pzk9pp1b95.fsf@eeyore.ibcinc.com>
Roderick Schertler [Sun, 5 Jan 1997 02:25:58 +0000 (21:25 -0500)]
doc patch: perldsc
Mostly just fixed a couple of bracing and sorting problems, plus a little
bit of cosmetics. Dunno how this bitrot seeps in.
p5p-msgid: <1623.
846039492@jinete.perl.com>
private-msgid: <pzafqo6eo9.fsf@eeyore.ibcinc.com>
Fabien TASSIN [Sun, 5 Jan 1997 23:47:01 +0000 (00:47 +0100)]
pod2html.PL patch (for 5.003-19)
I've just found 2 bugs:
- incorrect order for tags in the header.
- bad links if the filename contains one (or more) "_".
Here is the patch :
nexus:~perl/5.003-19/src/pod $ diff -c pod2html.PL.orig pod2html.PL
p5p-msgid: <
199701052347.
AAA21297@solar5>
Roderick Schertler [Sun, 5 Jan 1997 02:28:30 +0000 (21:28 -0500)]
doc patch for defined on perlfunc.pod
I keep getting this question, so I thought I'd doc it. I still wonder
whether it's a bug. How does this sound, Larry?
p5p-msgid: <1509.
846038569@jinete.perl.com>
private-msgid: <pz91686ek1.fsf@eeyore.ibcinc.com>
Robin Barker [Tue, 7 Jan 1997 17:19:59 +0000 (17:19 +0000)]
File::Basename::dirname bugs
private-msgid: <12393.
9701071719@tempest.cise.npl.co.uk>
Roderick Schertler [Sat, 4 Jan 1997 16:12:13 +0000 (11:12 -0500)]
tiny doc patches
p5p-msgid: <23338.
852394333@eeyore.ibcinc.com>
Roderick Schertler [Mon, 6 Jan 1997 15:39:52 +0000 (10:39 -0500)]
More magic variable tests
private-msgid: <7043.
852565192@eeyore.ibcinc.com>
Roderick Schertler [Mon, 6 Jan 1997 20:42:04 +0000 (15:42 -0500)]
sigaction() problems
Working out an example of non-restarting syscalls I found three
problems.
- sigaction warns if there are no flags in the SigAction structure. I
think the SigAction constructor should treat the SigSet and flags
args as optional. Minimalist patch appended.
$ ./perl -MPOSIX=:signal_h -we '
sigaction 2, new POSIX::SigAction sub { }'
Use of uninitialized value at -e line 1.
$ ./perl -MPOSIX=:signal_h -we '
sigaction 2, new POSIX::SigAction sub { }, undef, 0'
$ _
- POSIX::constant warns on an arg-less macro.
$ ./perl -MPOSIX=:signal_h -lwe 'print SIGALRM'
Use of uninitialized value at /usr/local/lib/perl5/POSIX.pm line 197.
14
$ ./perl -MPOSIX=:signal_h -lwe 'print SIGALRM(0)'
14
$ _
- sigaction doesn't actually work.
$ ./perl -MPOSIX=:signal_h -we '
sigaction SIGALRM(0),
new POSIX::SigAction sub { die "alarm\n" }, undef, 0;
kill "ALRM", $$'
SIGALRM handler "CODE(0x223970)" not defined.
$ _
p5p-msgid: <12808.
852583324@eeyore.ibcinc.com>
Chip Salzenberg [Mon, 6 Jan 1997 22:00:11 +0000 (10:00 +1200)]
Finish OP= warnings: none on ^=
Ollivier Robert [Sun, 5 Jan 1997 21:41:49 +0000 (22:41 +0100)]
Support libperl.so under FreeBSD
Subject: Re: 5.003_18 + shared libperl + FreeBSD 3.0-CURRENT
According to Andy Dougherty:
> There's a section in Configure to add such things. Edit Configure and
> look for a section like
After many tests, here are the two patches that enable use of either shared
or static libperl. Please everyone verify that I didn't broke anything
(especially with the second patch in Makefile.SH)...
Differences between NetBSD & FreeBSD surprise me because we're supposed to
have almost the same ld/ld.so...
p5p-msgid: <Mutt.
19970105224149.roberto@keltia.freenix.fr>
Chip Salzenberg [Sat, 4 Jan 1997 22:51:38 +0000 (10:51 +1200)]
Fix $^X on systems that set it to Perl's basename
Doug MacEachern [Tue, 7 Jan 1997 22:20:46 +0000 (17:20 -0500)]
plug for safe/opcode leaks
CHECKOP was bailing out at compile time here without destroying
the newly created/unscoped op.
p5p-msgid: <
199701072220.RAA02117@postman.osf.org>
Chip Salzenberg [Tue, 7 Jan 1997 23:52:00 +0000 (11:52 +1200)]
[shell changes from patch from perl5.003_19 to perl5.003_20]
Change from running these commands:
# this file is obsolete
rm -f ext/DynaLoader/dl_os2.xs
# this file was renamed
if test -f t/pragma/warn-global
then
mv t/pragma/warn-global t/pragma/warn-1global
fi
# new (and nearly new) tests must be executable
touch t/comp/proto.t
chmod +x t/comp/proto.t t/comp/use.t t/harness
# ready to patch
exit 0
Chip Salzenberg [Sat, 4 Jan 1997 05:44:00 +0000 (17:44 +1200)]
[differences between cumulative patch application and perl5.003_19]
Perl 5 Porters [Sat, 4 Jan 1997 05:44:00 +0000 (17:44 +1200)]
[inseparable changes from patch from perl5.003_18 to perl5.003_19]
CORE LANGUAGE CHANGES
Subject: Make method cache invisible to user code
From: Chip Salzenberg <chip@atlantic.net>
Files: dump.c gv.c gv.h hv.c op.c perl.c pp_hot.c pp_sys.c sv.c toke.c
Subject: Never parse "{m,s,y,tr,q{,q,w,x}}:{,:}" as package or label
From: Chip Salzenberg <chip@atlantic.net>
Files: toke.c
CORE PORTABILITY
Subject: Fix $^X under HP-UX
From: Chip Salzenberg <chip@atlantic.net>
Files: hints/hpux.sh toke.c
Subject: New hints/hpux.sh
Date: Tue, 31 Dec 1996 15:09:32 -0800
From: Jeff Okamoto <okamoto@hpcc123.corp.hp.com>
Files: hints/hpux.sh
private-msgid: <
199612312309.
AA283393772@hpcc123.corp.hp.com>
DOCUMENTATION
Subject: Perlguts, version 28
Date: Fri, 3 Jan 1997 13:10:46 -0800
From: Jeff Okamoto <okamoto@hpcc123.corp.hp.com>
Files: pod/perlguts.pod
private-msgid: <
199701032110.
AA102535846@hpcc123.corp.hp.com>
Subject: Miscellaneous pod patches
From: Ralf S. Engelschall <rse@engelschall.com>
Files: pod/Makefile pod/perldebug.pod pod/perlfunc.pod pod/perlguts.pod
Subject: expanded flock() docs
Date: Fri, 03 Jan 1997 19:31:11 -0500
From: Roderick Schertler <roderick@gate.net>
Files: pod/perlfunc.pod
Msg-ID: <4481.
852337871@eeyore.ibcinc.com>
(applied based on p5p patch as commit
1fd81fbbe87d964ad1f7dbdce41e36f3781dcf82)
Subject: Use Text::Wrap in buildtoc; run buildtoc
From: Ulrich Pfeifer <pfeifer@charly.informatik.uni-dortmund.de>
Files: pod/buildtoc pod/perltoc.pod
Subject: Remove obsolete perlovl.pod
From: Chip Salzenberg <chip@atlantic.net>
Files: MANIFEST plan9/mkfile pod/perlovl.pod vms/Makefile vms/descrip.mms
OTHER CORE CHANGES
Subject: Fix segv when calling named closures
From: Chip Salzenberg <chip@atlantic.net>
Files: pp_hot.c
Subject: Finish rationalizing "undef value" warnings
From: Chip Salzenberg <chip@atlantic.net>
Files: doop.c pp.c sv.c t/op/assignwarn.t
Subject: Arrange for all "_<file" entries to be in %main::
From: Chip Salzenberg <chip@atlantic.net>
Files: gv.c lib/perl5db.pl
Subject: Introduce CVf_NODEBUG flag
Date: Wed, 01 Jan 1997 15:42:05 -0500
From: Gurusamy Sarathy <gsar@engin.umich.edu>
Files: cv.h pp_hot.c
Msg-ID: <
199701012042.PAA25994@aatma.engin.umich.edu>
(applied based on p5p patch as commit
a3d90dd510fe5a67ed9b80e603493d285c30aa97)
Subject: Reword 'may be "0"' warning per Larry; fix its line number
From: Chip Salzenberg <chip@atlantic.net>
Files: op.c pod/perldiag.pod
Subject: 5.003_18: perl_{con,des}truct fixes
Date: Fri, 03 Jan 1997 15:42:04 -0500
From: Doug MacEachern <dougm@osf.org>
Files: perl.c perl.h pod/perlembed.pod pod/perltoc.pod t/op/sysio.t
Msg-ID: <
199701032042.PAA06766@postman.osf.org>
(applied based on p5p patch as commit
316c7b3d7b47e3143f94c7f8621e854c519d1e87)
Subject: Fix lost value from READLINE after TIEHANDLE
From: Gurusamy Sarathy <gsar@engin.umich.edu>
Files: pp_hot.c sv.h
TESTS
Subject: Create t/pragma directory; populate with new and old
From: Paul Marquess <pmarquess@bfsec.bt.co.uk>
Files: MANIFEST Makefile.SH t/TEST t/comp/use.t t/lib/locale.t t/op/overload.t t/op/use.t t/pragma/locale.t t/pragma/overload.t t/pragma/strict-refs t/pragma/strict-subs t/pragma/strict-vars t/pragma/strict.t t/pragma/subs.t t/pragma/warn-global t/pragma/warning.t
Subject: New tests: comp/colon.t and op/assignwarn.t
From: Robin Barker <rmb@cise.npl.co.uk>
Files: MANIFEST t/comp/colon.t t/op/assignwarn.t
Roderick Schertler [Sat, 4 Jan 1997 00:31:11 +0000 (19:31 -0500)]
expanded flock() docs
p5p-msgid: <4481.
852337871@eeyore.ibcinc.com>
Tim Bunce [Fri, 3 Jan 1997 17:48:46 +0000 (17:48 +0000)]
Re: perldelta, take 3
> From: Tom Christiansen <tchrist@mox.perl.com>
Some quick thoughts:
> =head1 NAME
>
> perldelta - what's new for perl5.004 (version 0.3)
To avoid any confusion: s/version/perldelta version/
> =head2 Internal Change: Safe Module Absorbed into Opcode
>
> A new Opcode module subsumes 5.003's Safe module. The Safe
> interface is still available, so existing scripts should still
> work, but users are encouraged to read the new Opcode documentation.
> (Note too, that after 5.004 is installed, you will no longer be
> able to run Safe under the 5.003 binary.)
Opcode does not supercede Safe. Safe is now implemented on top of Opcode.
Opcode is purely functional. Safe implements OO 'compartments'.
I'd suggest you say something like:
=head2 New Opcode Module and Revised Safe Module
A new Opcode module supports the creation, manipulation and application
of opcode masks. The revised Safe module has a new API and is implemented
using the new Opcode module. Users are encouraged to read the new Opcode
and Safe documentation. (Note too, that after 5.004 is installed, you will
no longer be able to run Safe under the 5.003 binary.)
> =item use Module VERSION LIST
>
> If the VERSION argument is present between Module and LIST, then the
> C<use> will fail if the $VERSION variable in package Module is
> less than VERSION.
>
> Note that there is not a comma after the version!
+ This is similar to the existing version checking mechanism in the
+ Exporter module but it's faster and can be used with modules that
+ don't use the Exporter. It is the recommended method for new code.
> =head2 New Built-in Methods
>
> The C<UNIVERSAL> package automatically contains the following methods that
> are inherited by all other classes:
>
> =item isa ( CLASS )
> =item can ( METHOD )
> =item VERSION ( [ VERSION ] )
>
> use A 1.2 qw(some imported subs);
>
> A->VERSION( 1.2 );
> $ref->is_instance(); # True
I don't see is_instance described anywhere. I think it needs a few words.
> =head1 Pragmata
>
> Three new pragmatic modules exist:
>
> =item use ops
>
> Restrict unsafe operations when compiling
! Disable named/unsafe opcodes when compiling perl code
> =head1 Efficiency Enhancements
>
> All hash keys with the same string are only allocated once, so
> even if you have 100 copies of the same hash, the immutable keys
> never have to be re-allocated.
>
> Functions that do nothing but return a fixed value are now inlined.
! Functions with an empty prototype that do nothing but return a
! fixed value are now inlined. E.g., sub PI () { 3.14159 }
p5p-msgid: <
9701031748.AA15335@toad.ig.co.uk>
Doug MacEachern [Fri, 3 Jan 1997 20:42:04 +0000 (15:42 -0500)]
5.003_18: perl_{con,des}truct fixes
Here's my current patch to clean/reset global variables. I trust more
work needs to be done in this area, there are comments that might help
if something pops up down the line.
I did not test MULTIPLICITY extensively, but a simple test case seems to
work, I've added that to perlembed.pod.
A *huge* _thank you_ to Chip for your help in this area, 5.004 will make
many once frusterated embedders very happy!
p5p-msgid: <
199701032042.PAA06766@postman.osf.org>
Jarkko Hietaniemi [Thu, 2 Jan 1997 05:27:47 +0000 (17:27 +1200)]
Make libs clean under '-w'
Gurusamy Sarathy [Wed, 1 Jan 1997 20:42:05 +0000 (15:42 -0500)]
Introduce CVf_NODEBUG flag
Subject: Re: debugger and XSUBs
On Wed, 01 Jan 1997 13:45:32 EST, Chip Salzenberg wrote:
>According to Gurusamy Sarathy:
>> P.S: Porters: Come to think of if, I can't seem to find a
>> good enough reason for why even XSUBs (like Alias::attr())
>> must be called by DB::sub() and not directly by perl (when
>> perldb). Anyone else can?
>
>Nope.
Actually, there is a case for it, since you may want to profile
XSUBs.
>> I ask because the fix will be much simpler in perl (just skip the
>> OPpENTERSUB_DB indirection in pp_entersub() for XSUBs).
>
>I vote for this change. It's even a performance improvement.
Here's a simple patch that offers the best of both worlds. I have added
a new flag CVf_NODEBUG, that can be used to turn off the DB::sub
indirection for specific CVs. This is most likely to be used
on XSUBs that must depend on the calling scope's structure
(which C<-d> alters), but can be used on the CV of plain subs
as well.
This facility will also be useful in Dprof, where one can conceivably
turn off the profiling of all subs except the target one in the
interest of accurately timing the target sub's performance.
I do the following now in the BOOT: section of Alias.xs to
disable debugging of Alias::attr():
BOOT:
{
GV *gv = gv_fetchpv("Alias::attr", FALSE, SVt_PVCV);
if (gv && GvCV(gv))
CvNODEBUG_on(GvCV(gv));
}
Perlanoids will be happy to note that this patch has no effect
unless the -d switch is used.
p5p-msgid: <
199701012042.PAA25994@aatma.engin.umich.edu>
Chip Salzenberg [Thu, 2 Jan 1997 00:22:16 +0000 (12:22 +1200)]
Free memory of large lexical variables when leaving scope
Chip Salzenberg [Sat, 4 Jan 1997 05:44:00 +0000 (17:44 +1200)]
[shell changes from patch from perl5.003_18 to perl5.003_19]
Change from running these commands:
# these files are obsolete
rm -f pod/perlovl.pod
rm -f t/lib/locale.t
rm -f t/op/overload.t
rm -f t/op/use.t
# new test directory must exists
test -d t/pragma || mkdir t/pragma
# new tests must be executable
t="t/comp/colon.t t/op/assignwarn.t t/pragma/locale.t \
t/pragma/overload.t t/pragma/strict.t t/pragma/subs.t t/pragma/warning.t"
touch $t
chmod +x $t
# now that we have splain.PL, this need not be executable
chmod a-x lib/diagnostics.pm
# ready to patch
exit 0
Chip Salzenberg [Tue, 31 Dec 1996 20:59:00 +0000 (08:59 +1200)]
[differences between cumulative patch application and perl5.003_18]
Perl 5 Porters [Tue, 31 Dec 1996 20:59:00 +0000 (08:59 +1200)]
[inseparable changes from patch from perl5.003_17 to perl5.003_18]
CORE LANGUAGE CHANGES
Subject: Inherited overloading
Date: Sun, 29 Dec 1996 08:12:54 -0500 (EST)
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: gv.c lib/overload.pm perl.h sv.c sv.h t/op/overload.t
Chip Salzenberg writes:
>
> Patch now, tarchive later:
Below is the fixed overloading patch.
Note that in between AMG_names got const on it (a good thing!), but as
a corollary I needed to cast away const-ness to actually use it
(since, say, newSVpv does not have const args).
Enjoy,
p5p-msgid: <
199612291312.IAA02134@monk.mps.ohio-state.edu>
Subject: Closures at file scope must be anonymous
From: Chip Salzenberg <chip@atlantic.net>
Files: op.c
Subject: Warn on '{if,while} ($x = X)' where X is glob, readdir, or <FH>
From: Chip Salzenberg <chip@atlantic.net>
Files: op.c pod/perldiag.pod
DOCUMENTATION
Subject: Re: perldiag.pod entry for "Scalar value @%s{%s} ..."
Date: Tue, 31 Dec 1996 11:50:19 -0500
From: Roderick Schertler <roderick@gate.net>
Files: pod/perldiag.pod
Msg-ID: <2043.
852051019@eeyore.ibcinc.com>
(applied based on p5p patch as commit
c885792efecf3f527b3b5099727cc16b03eee1dc)
OTHER CORE CHANGES
Subject: Get rid of 'Leaked scalars'
From: Chip Salzenberg <chip@atlantic.net>
Files: cop.h gv.c op.c
TESTS
Subject: Expanded locale.t and misc.t
From: Jarkko Hietaniemi <jhi@cc.hut.fi>
Files: t/lib/locale.t t/lib/misc.t
Subject: Expanded my.t
From: Chip Salzenberg <chip@atlantic.net>
Files: t/lib/my.t
Roderick Schertler [Tue, 31 Dec 1996 16:50:19 +0000 (11:50 -0500)]
Re: perldiag.pod entry for "Scalar value @%s{%s} ..."
On Mon, 30 Dec 1996 21:28:30 -0500, Gurusamy Sarathy <gsar@engin.umich.edu> said:
> On Mon, 30 Dec 1996 21:09:12 EST, Roderick Schertler wrote:
>>
>> +=item Scalar value @%s{%s} better written as $%s{%s}
>> +
>> +This is just like the previous entry, but for hashes instead of arrays.
>> +
>
> Won't that look odd under -Mdiagnostics? Better to spell it out
> again.
Right you are. This diff goes on top of the one above.
p5p-msgid: <2043.
852051019@eeyore.ibcinc.com>
Dominic Dunlop [Sat, 28 Dec 1996 09:56:41 +0000 (10:56 +0100)]
Locale-related pod patches, take 2
[Ahem. Had the wrong thing in the scratch-pad, didn't I? Please ignore my
previous full posting of a slightly-tweaked perllocale.pod. This mail
contains what I really meant to send.]
Herewith (quick, before _18 appears) locale-related patches to the
documentation in perl5.003_17/pod. The main effect is to add
locale-related information to pods other than perllocale.pod, although
there are some tiny tweaks to that pod too. Produces no complaints from
pod2man; not checked for layout since 5.003_13.
p5p-msgid: <v03007800aeea9e488b36@[194.51.248.77]>
Dominic Dunlop [Mon, 30 Dec 1996 12:31:06 +0000 (00:31 +1200)]
Updates to perllocale.pod
Ilya Zakharevich [Sat, 28 Dec 1996 03:47:24 +0000 (22:47 -0500)]
File::Copy under OS/2
Chip Salzenberg writes:
>
> Patch now, tarchive later:
>
> file: $CPAN/authors/id/CHIPS/perl5.003_17.pat.gz
Almost clean under OS/2: the only problem is with File::Copy: the
test for syscopy was inverted, and test contained some Un*xisms.
Note that the POD contains some line noise, I marked it with ?????.
Enjoy,
p5p-msgid: <
199612280347.WAA00293@monk.mps.ohio-state.edu>
Tom Phoenix [Mon, 30 Dec 1996 17:24:16 +0000 (09:24 -0800)]
Improving Config.pm
private-msgid: <Pine.GSO.3.95.
961230091244.13467L-100000@solaris.teleport.co
Tom Phoenix [Mon, 30 Dec 1996 01:46:21 +0000 (17:46 -0800)]
More tests
private-msgid: <Pine.GSO.3.95.
961229170736.15213M-100000@solaris.teleport.co
Graham Barr [Mon, 30 Dec 1996 07:00:18 +0000 (07:00 +0000)]
test harness for C<use x.xxxx>
Here is the test harness for version numbers in use statements. It
checks both the C<use x.xxx> and C<use MODULE x.xxxx>
Paul suggested in an earlier message, the introduction
of a pragma directory. I would suggest that this test
should also go in the pragma directory.
p5p-msgid: <
32C76882.
3F3C7999@tiuk.ti.com>
Chip Salzenberg [Fri, 27 Dec 1996 00:55:26 +0000 (12:55 +1200)]
Add missing syms to global.sym; update magic doc
Chip Salzenberg [Mon, 30 Dec 1996 21:52:01 +0000 (09:52 +1200)]
Fix core dump on perl_construct()/perl_destruct() loop
Chip Salzenberg [Mon, 30 Dec 1996 21:34:20 +0000 (09:34 +1200)]
Don't forget $c in C<(($a,$b,$c)=(1,2))=(3,4,5)>
Chip Salzenberg [Mon, 30 Dec 1996 22:57:15 +0000 (10:57 +1200)]
Ultrix setlocale() workaround
Chip Salzenberg [Sun, 29 Dec 1996 21:20:34 +0000 (09:20 +1200)]
Warn on 'undef $x; $x OP 1' where OP is *=, /=, %=, or **=
Chip Salzenberg [Mon, 30 Dec 1996 21:33:01 +0000 (09:33 +1200)]
Hide lexicals from C<use>d or C<require>d module (!)
Chip Salzenberg [Tue, 31 Dec 1996 20:59:00 +0000 (08:59 +1200)]
[shell changes from patch from perl5.003_17 to perl5.003_18]
Change from running these commands:
# new tests must be executable
touch t/op/use.t
chmod +x t/op/use.t
# ready to patch
exit 0
Chip Salzenberg [Fri, 27 Dec 1996 18:22:00 +0000 (06:22 +1200)]
[differences between cumulative patch application and perl5.003_17]
Perl 5 Porters [Fri, 27 Dec 1996 18:22:00 +0000 (06:22 +1200)]
[inseparable changes from patch from perl5.003_16 to perl5.003_17]
Jarkko Hietaniemi [Thu, 26 Dec 1996 13:06:04 +0000 (15:06 +0200)]
tiny perllocale.pod diff for _16
private-msgid: <
199612261306.PAA21161@alpha.hut.fi>
Kenneth Albanowski [Wed, 25 Dec 1996 04:00:10 +0000 (23:00 -0500)]
perlpod.pod patch for _16
This documents the new =for/=begin/=end behavior, and slightly changes the
emphasis on HTML in description of E<>, hopefully for the better.
p5p-msgid: <Pine.LNX.3.93.
961224225906.337B-100000@kjahds.com>
Jeff Okamoto [Wed, 25 Dec 1996 01:44:23 +0000 (17:44 -0800)]
Perlguts, version 27
private-msgid: <
199612250144.
AA059528263@hpcc123.corp.hp.com>
Ilya Zakharevich [Thu, 26 Dec 1996 19:54:34 +0000 (14:54 -0500)]
Newer debugger
Here are the fruits of my spending _days_ trying to understand why
MakeMaker did not work on my extension.
Docs in the second chunk.
Enjoy,
p5p-msgid: <
199612242305.SAA10757@monk.mps.ohio-state.edu>
private-msgid: <
199612261954.OAA12999@monk.mps.ohio-state.edu>
Andreas Koenig [Wed, 25 Dec 1996 22:01:43 +0000 (10:01 +1200)]
Refresh MakeMaker to 5.39
Chip Salzenberg [Thu, 26 Dec 1996 02:07:36 +0000 (14:07 +1200)]
Test C< ()=() >
Ilya Zakharevich [Thu, 26 Dec 1996 00:31:14 +0000 (12:31 +1200)]
New tests: op/method.t and op/locale.t
Signed-off-by: Ilya Zakharevich <ilya@math.ohio-state.edu>
Chip Salzenberg [Wed, 25 Dec 1996 20:55:25 +0000 (08:55 +1200)]
Limit effects of "=pod" to a single file
Chip Salzenberg [Wed, 25 Dec 1996 09:42:02 +0000 (21:42 +1200)]
Fix {,un}tainting of $1 etc. when C<use locale>
Ilya Zakharevich [Wed, 25 Dec 1996 21:05:42 +0000 (16:05 -0500)]
Minor OS/2 fixes
These are pretty minor (though fix one test-case and can make CPAN.pm
behave better).
p5p-msgid: <
199612252105.QAA11890@monk.mps.ohio-state.edu>
Ralf S. Engelschall [Thu, 26 Dec 1996 02:04:03 +0000 (14:04 +1200)]
Suggest "usemymalloc='n'" for FreeBSD 2.*
Winfried Koenig [Wed, 25 Dec 1996 22:45:45 +0000 (00:45 +0200)]
perl5.003_15 and Interactive Unix
with my patches, perl5.003_15 compiled with cc, runs all tests.
Compiled with gcc I get this failed tests:
op/goto...........FAILED on test 8
op/sysio..........FAILED on test 0
lib/autoloader....Goto undefined subroutine at ./lib/autoloader.t line 58.
FAILED on test 3
lib/english.......FAILED on test 0
lib/fatal.........FAILED on test 0
lib/hostname......FAILED on test 0
Failed 6/135 tests, 91.11% okay.
On my Linux System I get similar failures:
op/goto...........FAILED on test 8
op/sysio..........Negative length at ./op/sysio.t line 64.
FAILED on test 1
lib/autoloader....Goto undefined subroutine at ./lib/autoloader.t line 58.
FAILED on test 0
lib/english.......FAILED on test 0
lib/fatal.........Can't use an undefined value as filehandle reference at ./lib/fatal.t line 21.
lib/hostname......FAILED on test 0
Failed 6/135 tests, 94.81% okay.
So I think that this problems are already fixed. My changes to
perl.h are necessary to compile x2p/malloc.c, but there may be
better solutions.
p5p-msgid: <m0vd254-0004oKC@incom.rhein-main.de>
Chip Salzenberg [Thu, 26 Dec 1996 01:07:14 +0000 (13:07 +1200)]
Support named closures
Chip Salzenberg [Fri, 27 Dec 1996 18:22:00 +0000 (06:22 +1200)]
[shell changes from patch from perl5.003_16 to perl5.003_17]
Change from running these commands:
# new tests must be executable
touch t/lib/locale.t t/op/method.t
chmod +x t/lib/locale.t t/op/method.t
# ready to patch
exit 0
Chip Salzenberg [Tue, 24 Dec 1996 23:25:00 +0000 (11:25 +1200)]
[differences between cumulative patch application and perl5.003_16]
Perl 5 Porters [Tue, 24 Dec 1996 23:25:00 +0000 (11:25 +1200)]
[inseparable changes from patch from perl5.003_15 to perl5.003_16]
CORE PORTABILITY
Subject: _13: patches for unicos/unicosmk
Date: Fri, 20 Dec 1996 14:38:50 -0600
From: Dean Roehrich <roehrich@cray.com>
Files: Configure MANIFEST hints/unicos.sh hints/unicosmk.sh
private-msgid: <
199612202038.OAA22805@poplar.cray.com>
LIBRARY AND EXTENSIONS
Subject: Refresh IO to 1.14
From: Graham Barr <gbarr@ti.com>
Files: MANIFEST ext/IO/IO.xs ext/IO/README ext/IO/lib/IO/File.pm ext/IO/lib/IO/Handle.pm ext/IO/lib/IO/Pipe.pm ext/IO/lib/IO/Seekable.pm ext/IO/lib/IO/Select.pm ext/IO/lib/IO/Socket.pm t/lib/io_dup.t t/lib/io_pipe.t t/lib/io_sel.t t/lib/io_sock.t t/lib/io_tell.t t/lib/io_udp.t t/lib/io_xs.t
OTHER CORE CHANGES
Subject: Fix 'foreach(@ARGV) { while (<>) { push @p,$_ } }'
From: Chip Salzenberg <chip@atlantic.net>
Files: cop.h pp_hot.c scope.c
Subject: Eliminate warnings from C< undef $x; $x OP= "foo" >
From: Chip Salzenberg <chip@atlantic.net>
Files: doop.c pp.c pp.h pp_hot.c
Subject: Try again to improve method caching
Date: Mon, 23 Dec 1996 20:13:56 -0500 (EST)
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: gv.c sv.c
Msg-ID: <
199612240113.UAA09487@monk.mps.ohio-state.edu>
(applied based on p5p patch as commit
81c78688fe5c3927ad37ba29de14c86e38120317)
Subject: Be more careful about 'o' magic memory management
From: Chip Salzenberg <chip@atlantic.net>
Files: mg.c sv.c
Subject: Fix bad pointer refs when localized object loses magic
From: Chip Salzenberg <chip@atlantic.net>
Files: scope.c
Ilya Zakharevich [Tue, 24 Dec 1996 01:13:56 +0000 (20:13 -0500)]
Try again to improve method caching
Subject: Re: Autoloading broken?!
Chip Salzenberg writes:
>
> According to Ilya Zakharevich:
> >
>
> Well, I can only guess what your message was going to say... But if
> you build stock _14, you'll find that MakeMaker doesn't work, because
> SelfLoader doesn't work.
>
> I think it has something to do with your patch finding completely
> empty functions (no XSUB and no code) and ignoring -- or even removing
> -- them, under the assumption they're bad cache entries. But that
> approach can make declarations like "sub Foo::bar;" evaporate into
> nothingness, when such declarations are sometimes used to force a call
> to Foo::AUTOLOAD().
In a correct package - FOO. I think it would call some AUTOLOAD
anyway, this is why this case slipped through my testing.
> That's my understanding, anyway.
Thanks, I found this too (and fixed it). I think it should work better
now. So far only other places which I found broken by my previous
patch are "overloading + AUTOLOADing", and "->can + AUTOLOAD".
These 3 cases work now (after correcting a bug in overload.t's AUTOLOAD).
p5p-msgid: <
199612240113.UAA09487@monk.mps.ohio-state.edu>
Jarkko Hietaniemi [Tue, 24 Dec 1996 04:09:31 +0000 (16:09 +1200)]
Update to perllocale.pod
Chip Salzenberg [Tue, 24 Dec 1996 04:07:29 +0000 (16:07 +1200)]
Edit INSTALL to describe new binary compat setup
Robin Barker [Tue, 24 Dec 1996 04:08:19 +0000 (16:08 +1200)]
Fix compilation errors when malloc.c used for x2p
Paul Marquess [Tue, 24 Dec 1996 04:07:24 +0000 (16:07 +1200)]
Use root EXTERN.h when compiling x2p/malloc.c.
Chip Salzenberg [Tue, 24 Dec 1996 04:08:18 +0000 (16:08 +1200)]
Don't recurse into subdirs twice on 'make realclean'
Russell Mosemann [Tue, 24 Dec 1996 04:08:16 +0000 (16:08 +1200)]
Refresh Net::Ping to 2.02
Andreas Koenig [Tue, 24 Dec 1996 04:08:07 +0000 (16:08 +1200)]
Refresh CPAN to 1.09
Chip Salzenberg [Tue, 24 Dec 1996 04:22:19 +0000 (16:22 +1200)]
Fix closures that are not in subroutines
Chip Salzenberg [Tue, 24 Dec 1996 23:25:00 +0000 (11:25 +1200)]
[shell changes from patch from perl5.003_15 to perl5.003_16]
Change from running these commands:
# new test must be executable
touch t/lib/io_sel.t
chmod +x t/lib/io_sel.t
# ready to patch
exit 0
Chip Salzenberg [Mon, 23 Dec 1996 21:07:00 +0000 (09:07 +1200)]
[revert some function caching changes]
[remove] some function caching changes that were supposed to be
improvements, but weren't.
Chip Salzenberg [Mon, 23 Dec 1996 00:58:58 +0000 (12:58 +1200)]
[differences between cumulative patch application and perl5.003_14]
Perl 5 Porters [Mon, 23 Dec 1996 00:58:58 +0000 (12:58 +1200)]
[inseparable changes from patch from perl5.003_13 to perl5.003_14]
CORE LANGUAGE CHANGES
Subject: Eliminate support for {if,unless,while,until} BLOCK BLOCK
From: Chip Salzenberg <chip@atlantic.net>
Files: perly.c perly.c.diff perly.y toke.c
Subject: Taint $x after $x =~ s/pat/xyz/ if pat or xyz is tainted by locale
From: Chip Salzenberg <chip@atlantic.net>
Files: cop.h mg.c pp_ctl.c pp_hot.c
Subject: Complete support for modifying undefined array members in foreach
From: Chip Salzenberg <chip@atlantic.net>
Files: global.sym mg.c perl.h pp.c pp_hot.c proto.h sv.c
DOCUMENTATION
Subject: Update pod/Makefile; s/perli18n/perllocale/
From: Chip Salzenberg <chip@atlantic.net>
Files: ext/POSIX/POSIX.pod lib/I18N/Collate.pm pod/Makefile pod/perl.pod pod/perlmod.pod pod/perlnews.pod pod/roffitall
OTHER CORE CHANGES
Subject: Bug in debugger with import manipulations
Date: Mon, 23 Dec 1996 05:37:48 -0500 (EST)
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: pp_hot.c
Finally I traced why MakeMaker runs wrongly under debugger: debugger
function calling sequence was assuming that
GvCV(CvGV(cv)) == cv
for non-anonymous subs (ne END). MakeMaker managed to break it by
*A::B = \&C::D;
eval 'sub C::D {new one}';
After this CvGV(\&A::B) is *C::D, but &{*C::D} is the "new one".
Patch follows (note that in this case we do not sacrifice having a
subroutine name in debugger output ;-).
Enjoy,
p5p-msgid: <
199612231037.
FAA08617@monk.mps.ohio-state.edu>
Subject: Import and dynamic methods
Date: Mon, 23 Dec 1996 01:45:37 -0500 (EST)
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: gv.c hv.c sv.c
Here is the patch which corrects bad things which happens when you
import subroutines and otherwise manipulate the symbol tree.
I put forward the only chunk which may be controversal, since it may
have a minor performance penalty. It is independent of the others, so
it can be freely deleted.
The manipulations which correctly propagate to method calls:
a) Pruning globs: delete $B::{method}
b) Undefing subroutines: undef &B::method;
c) Importing: *B::method = \&mymethod;
Enjoy,
p5p-msgid: <
199612230645.
BAA08378@monk.mps.ohio-state.edu>
Subject: sv_gets patch
Date: Sun, 22 Dec 1996 03:24:04 -0500 (EST)
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: pp_hot.c
I sent this patch before as a part of
pos $str =
$str = /failing/g
patch. Now I separate it in the case it was tainted by environment:
Synopsis:
$a = <FH>;
does not work as expected if $a is magic (say, tied).
Enjoy,
p5p-msgid: <
199612220824.
DAA07235@monk.mps.ohio-state.edu>
Subject: pos $str patch
Date: Sun, 22 Dec 1996 03:31:21 -0500 (EST)
From: Ilya Zakharevich <ilya@math.ohio-state.edu>
Files: mg.c pp_hot.c t/op/pat.t
This patch was (mostly) posted already.
It fixes the bug:
pos $str
is reset to undef if
$str =~ /failing/g
fails.
Additionally, if fixes the hidded bit (=MGf_MINMATCH) surviving
setting
pos $str = ...
Enjoy,
p5p-msgid: <
199612220831.
DAA07247@monk.mps.ohio-state.edu>
PORTABILITY
Subject: Fix bugs in bincompat3 usage
From: Chip Salzenberg <chip@atlantic.net>
Files: perl.h perl_exp.SH
Subject: VMS patches to 5.003_13
Date: Mon, 23 Dec 1996 01:26:47 -0500 (EST)
From: Charles Bailey <bailey@HMIVAX.HUMGEN.UPENN.EDU>
Files: deb.c ext/POSIX/POSIX.xs gv.c lib/File/Copy.pm mg.c perl.c perl.h proto.h sv.c t/lib/filecopy.t taint.c toke.c util.c vms/Makefile vms/config.vms vms/descrip.mms vms/gen_shrfls.pl vms/genconfig.pl vms/genopt.com vms/perly_c.vms vms/perly_h.vms vms/test.com vms/vms.c vms/vms_yfix.pl
private-msgid: <01IDBYYFYPIS002ASE@hmivax.humgen.upenn.edu>
UTILITIES, LIBRARY, AND EXTENSIONS
Subject: Remove libnet
From: Chip Salzenberg <chip@atlantic.net>
Files: MANIFEST pod/perlmod.pod
Subject: Update IO->VERSION() to 1.1201 for CPAN's sake
From: Chip Salzenberg <chip@atlantic.net>
Files: ext/IO/lib/IO/Handle.pm lib/IO/Handle.pm
Subject: Remodel File::Copy.
From: Chip Salzenberg <chip@atlantic.net>
Files: lib/File/Copy.pm