Permalink
Commits on Jun 11, 1997
  1. [differences between cumulative patch application and perl5.004_01]

    [editor's note: The changes between this and 5.004 were processed from
    the m1t2 release, which was a bad idea as it was the _01 release which
    had the final corrected attributions.  The differences between the
    various m*t* releases do that; I considered it most valuable just to
    look at the _NN releases.  Many patches have been separated out and/or
    applied from the p5p archives nonetheless.]
    Tim Bunce committed Jun 11, 1997
  2. [inseparable changes from changes to perl-5.004_01-mt2]

     BUILD PROCESS
    
    Subject: [PATCH] INSTALL-1.18
    From: Andy Dougherty <doughera@fractal.phys.lafayette.edu>
    Files: INSTALL
    Msg-ID: Pine.SOL.3.95q.970529142739.662D-100000@fractal.lafayette.edu
    
        (applied based on p5p patch as commit 0dcb58f)
    
    Subject: improved gnuwin32 Configure support
    From: Chris Faylor <cgf@bbc.com>
    Files: Configure
    
        private-msgid: 199706070318.XAA09214@hardy.bbc.com
    
    Subject: installhtml problems finding splitpod
    From: Larry W. Virden <lvirden@cas.org>
    Files: installhtml INSTALL
    
    Subject: oddity in Configure
    From: Mike Stok <mike@stok.co.uk>
    Files: Configure
    
    Subject: Compiling perl5.004 on NEWS-OS 4.x
    From: Makoto MATSUSHITA (=?ISO-2022-JP?B?GyRCJF4kRCQ3JD8kXiQzJEgbKEI=?=) <matusita@ics.es.osaka-u.ac.jp>
    Files: Configure hints/newsos4.sh
    Msg-ID: 19970521132814F.matusita@ics.es.osaka-u.ac.jp
    
        (applied based on p5p patch as commit 4e81aff)
    
     CORE LANGUAGE
    
    Subject: first true value returned by scalar C<...> is wrong
    From: Hans Mulder <hansm@euronet.nl>
    Files: pp_ctl.c t/op/flip.t
    
    Subject: Regex Bug in 5.003_26 thru 003_99a
    From: Andreas Karrer <karrer@ife.ee.ethz.ch>
    Files: regcomp.h regcomp.c regexec.c
    Msg-ID: 199705161915.PAA18721@rio.atlantic.net
    
        (applied based on p5p patch as commit c8756f3)
    
    Subject: -w interacts badly with -Dt
    From: Spider Boardman <spider@Orb.Nashua.NH.US>
    Files: sv.c
    
    Subject: No DESTROY on untie. Tie memory leak fixed.
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: pp_hot.c
    Msg-ID: 199705172156.RAA20561@aatma.engin.umich.edu
    
        (applied based on p5p patch as commit 5117ca9)
    
    Subject: magic_clear_all_env proto should match svt_clear
    From: Nick Ing-Simmons <nik@tiuk.ti.com>
    Files: proto.h mg.c
    
    Subject: ENV leaks on win32 (was Re: Comments on ENV patch sought)
    From: Hans Mulder <hansm@euronet.nl>
    Files: embed.h perl.h proto.h global.sym mg.c t/op/magic.t
    Msg-ID: 199705292240.AAA01135@mail.euronet.nl
    
        (applied based on p5p patch as commit 66b1d55)
    
    Subject: Patch to show @INC when require dies
    From: Jim Avera <avera@hal.com>
    Files: pp_ctl.c
    Msg-ID: 9705230121.AA27872@membrane.hal.com
    
        (applied based on p5p patch as commit 2683423)
    
    Subject: bug with m// nested inside s///e
    From: Hans Mulder <hansm@euro.net>
    Files: op.c t/op/subst.t
    
     DOCUMENTATION
    
    Subject: perlembed Win32 update
    From: Doug MacEachern <dougm@opengroup.org>
    Files: pod/perlembed.pod
    
    Subject: perldiag.pod patch - "(W) substr outside string" is "(S)evere" if used as lvalue.
    From: John Hughes <john@AtlanTech.COM>
    Files: pod/perldiag.pod
    
    Subject: local(%ENV) looses magic - document behaviour
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: pod/perlsub.pod
    
    Subject: perlguts caveats
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: pod/perlguts.pod
    Msg-ID: 199705180052.UAA22066@aatma.engin.umich.edu,
    	199705180202.WAA22826@aatma.engin.umich.edu,
    Msg-IDs: 1997May17.235722.2033087@hmivax.humgen.upenn.edu,
    	 199705301341.JAA05204@aatma.engin.umich.edu
    
        (applied based on p5p patch as commits:
          04343c6
          6e2bfd7
          9edb2b4)
    
    Subject: pod2man produces broken pages
    From: Davin Milun <milun@cs.Buffalo.EDU>
    Files: pod/pod2man.PL
    Msg-ID: 199705310447.AAA15721@obelix.cs.Buffalo.EDU
    Msg-ID: 1997May25.192350.2055977@hmivax.humgen.upenn.edu
    
        (applied based on p5p patch as commits:
          3ad4915
          ca023ce)
    
    Subject: Perl 5 pod2man fix
    From: Franklin Chen <chen@adi.com>
    Files: pod/pod2man.PL
    
        Subject: perlguts man page corrupted
        
        perlguts.1 has garbled headings, e.g., the following is a result of
        "man perlguts":
        
        ===
             What is an
        
             Perl uses a special typedef IV which is a simple integer
             type that is guaranteed to be large enough to hold a pointer
             (as well as an integer).
        ===
        
        p5p-msgid: 199705210013.UAA09599@menhaden.adi.com
    
    Subject: reference form chomp to chop in perlfunc
    From: Hans Mulder <hansm@euronet.nl>
    Files: pod/perlfunc.pod
    
    Subject: pod2man gags if "=pod" is before "=head1 NAME"
    From: Warren Hyde <whyde@pezz.sps.mot.com>
    Files: pod/pod2man.PL
    Msg-ID: 9705212115.AA21730@pezz.sps.mot.com
    
        (applied based on p5p patch as commit a8aaa22)
    
    Subject: perlfunc.pod unclear about return value range of rand
    From: "Tuomas J. Lukka" <tjl@lukka.student.harvard.edu>
    Files: pod/perlfunc.pod
    
        Subject: Re: Array feature request 
        
        Tom Phoenix writes:
        >On Fri, 16 May 1997, Tuomas J. Lukka wrote:
        >
        >> To detract further, the perlfunc.pod says
        >> 
        >> 	"returns a random fractional number between 0 and EXPR".
        >> 
        >> Does that mean inclusive or exclusive i.e. do I run a risk of one
        >> in a billion receiving an undef one past the array limit and growing
        >> the array? If yes or if not, should that be mentioned?
        >
        >           0  <=  rand(EXPR)  <  EXPR
        >
        >Hope this helps!
        
        Yes, this behaviour is expected but it is definitely not documented
        in the pods. 
        
        p5p-msgid: m0wSMiC-000C9xC@lukka.student.harvard.edu
    
    Subject: Error in perllol manpage
    From: Chris Wick <cwick@lmc.com>
    Files: pod/perllol.pod
    
    Subject: 5.004 removed deprecated %OVERLOAD support silently
    From: Jonathan Biggar <jon@sems.com>
    Files: pod/perldelta.pod
    
        This is a bug report for perl from jon@sems.com,
        generated with the help of perlbug 1.17 running under perl 5.004.
        
        p5p-msgid: 199705232319.QAA28388@clamp.netlabs.com
    
    Subject: Documentation bugs
    From: Stephen Potter <spp@psa.pencom.com>
    Files: pod/perldata.pod pod/perldiag.pod pod/perlfaq8.pod pod/perlfaq9.pod pod/perlop.pod pod/perlsub.pod pod/perltoot.pod
    
    Subject: make html --> unusable xref links
    From: "Darren/Torin/Who Ever..." <torin@daft.com>
    Files: INSTALL pod/perldiag.pod installhtml
    
        Jim Avera writes:
         > By default (without editing the Makefile) the resulting .html files
         > are useless -- they have broken cross-reference links.
         > 
         > If you set HTMLROOT=. in pod/Makefile, then cross-references
         > say .//pod/... which does not work either.  
         > Setting HTMLROOT to an absolute path doesn't work either (the
         > spurrious /pod/ is always there).
        
         [seconded]
        
        p5p-msgid: 199705162008.XAA06906@alpha.hut.fi
    
    Subject: pod/*.html -- all hyperlinks are invalid
    From: "Darren/Torin/Who Ever..." <torin@daft.com>
    Files: INSTALL pod/perldiag.pod installhtml
    
        The INSTALL file says:
        
         (cd pod && make html && mv *.html <www home dir>)
        
        But the resulting *.html files have hyperlinks that look like this;
        
         <A HREF="/pod/perlfaq.html">perlfaq</A>
        
        The /pod/ part shouldn't be there.  As a result, all of the hyperlinks are
        invalid (unless, of course, you install the *.html files in /pod/).
        
        [Please do not change anything below this line]
        
        p5p-msgid: 199705171830.OAA15652@erawan.cognex.com
    
    Subject: Re: make html - any takers? 
    From: "Darren/Torin/Who Ever..." <torin@daft.com>
    Files: INSTALL pod/perldiag.pod installhtml
    
        On Sat, 07 Jun 1997 02:27:49 -0000, Tim Bunce wrote:
        >My reading of the following is that the best short term fix is for
        >INSTALL to be changed so it no longer suggests 
        >
        >	(cd pod && make html && mv *.html <www home dir>)
        >
        >Comments welcome. Proven portable and tested patches preferred :-)
        
        Win32 uses something similar to the suggested method with
        installhtml, and it seems to work reasonably well.  I'll supply
        a patch if I wander anywhere close to a UNIX box in the
        next 24 hours, and someone else hasn't already done it. :-)
        
        >When I follow the doc in installhtml, I get htmlwhich mostly works.
        >
        >Here's a sample:
        >
        >perl -w /ldata2/gnu/perl5.004/installhtml \
        >	--podpath=lib:ext:pod:vms:win32:plan9:os2:h2pl:x2p:Porting:eg \
        >	--podroot=/ldata2/gnu/perl5.004 \
        >	--htmldir=/projects/sprs_lwv/HTML/perl \
        >	--htmlroot=/projects/sprs_lwv/HTML/perl \
        >	--recurse \
        >	--splithead=pod/perlipc.pod \
        >	--splititem=pod/perlfunc.pod \
        >	--libpods=perlfunc:perlguts:perlvar:perlrun:perlop \
        >	--verbose
        
         - Sarathy.
           gsar@engin.umich.edu
        
        p5p-msgid: 199706081749.NAA04552@aatma.engin.umich.edu
    
    Subject: Re: make html --> unusable xref links
    From: "Darren/Torin/Who Ever..." <torin@daft.com>
    Files: INSTALL pod/perldiag.pod installhtml
    
        According to Jarkko Hietaniemi  <jhi@iki.fi>:
        :Jim Avera writes:
        : > If you set HTMLROOT=. in pod/Makefile, then cross-references
        : > say .//pod/... which does not work either.  
        : > Setting HTMLROOT to an absolute path doesn't work either (the
        : > spurrious /pod/ is always there).
        : > 
        : > What is the intent of the current setup (p54rc1)?
        
        When I follow the doc in installhtml, I get htmlwhich mostly works.
        
        Here's a sample:
        
        perl -w /ldata2/gnu/perl5.004/installhtml \
        	--podpath=lib:ext:pod:vms:win32:plan9:os2:h2pl:x2p:Porting:eg \
        	--podroot=/ldata2/gnu/perl5.004 \
        	--htmldir=/projects/sprs_lwv/HTML/perl \
        	--htmlroot=/projects/sprs_lwv/HTML/perl \
        	--recurse \
        	--splithead=pod/perlipc.pod \
        	--splititem=pod/perlfunc.pod \
        	--libpods=perlfunc:perlguts:perlvar:perlrun:perlop \
        	--verbose
        
        p5p-msgid: 1997May16.191039.2033079@hmivax.humgen.upenn.edu
    
    Subject: Re: make html --> unusable xref links
    From: "Darren/Torin/Who Ever..." <torin@daft.com>
    Files: INSTALL pod/perldiag.pod installhtml
    
        p5p-msgid: 87hgg2y1h4.fsf@perv.daft.com
    
    Subject: make html --> unusable xref links
    From: "Darren/Torin/Who Ever..." <torin@daft.com>
    Files: INSTALL pod/perldiag.pod installhtml
    
        I don't understand how "make html" is intended to be used.
        
        By default (without editing the Makefile) the resulting .html files
        are useless -- they have broken cross-reference links.
        
        If you set HTMLROOT=. in pod/Makefile, then cross-references
        say .//pod/... which does not work either.  
        Setting HTMLROOT to an absolute path doesn't work either (the
        spurrious /pod/ is always there).
        
        What is the intent of the current setup (p54rc1)?
        
        -Jim
        
        p5p-msgid: 9705161931.AA01075@membrane.hal.com
    
    Subject: 5.004 POD stuff
    From: "Darren/Torin/Who Ever..." <torin@daft.com>
    Files: INSTALL pod/perldiag.pod installhtml
    Msg-ID: 9705191839.AA28702@lemming.engeast
    
        (applied based on p5p patch as commit 1f8d200)
    
     LIBRARY AND EXTENSIONS
    
    Subject: sdbm can fail if a config.h exists in system directories
    From: Tim Bunce <Tim.Bunce@ig.co.uk>
    Files: ext/SDBM_File/sdbm/Makefile.PL
    
    Subject: LWP and SIG __DIE__ traps not playing well together!
    From: Gisle Aas <aas@bergen.sn.no>
    Files: lib/AutoLoader.pm
    
    Subject: Memory Consumption of autosplit_lib_modules/sv_gets (workaround)
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: lib/AutoSplit.pm
    
    Subject: Comments of this Sys::Syslog patch
    From: Jarkko Hietaniemi <jhi@iki.fi>
    Files: lib/Sys/Syslog.pm
    
        Subject: clpm: 5.004 Sys::Syslog
        
        Russ Allbery  <rra@stanford.edu> writes:
        >Richard Dows <marius@randomc.com> writes:
        >
        >> I wanted to use Sys::Syslog, and so I tested it with a small script.  I
        >> get an error like so:
        >
        >> Undefined subroutine &Sys::Syslog::hostname called at
        >> /usr/local/lib/perl5/Sys/Syslog.pm line 92.  BEGIN failed--compilation
        >> aborted at stest line 3.
        >
        >This problem is fixed in (the just released) 5.004.
        
        However, 5.004 changed the behavior of Syslog.pm in an icky way.
        
        In 5.003, Syslog called Sys::Hostname to get the name of the
        host, and happily used that as the logging host.
        
        Now, Syslog calls Sys::Hostname to get the name of the host,
        strips off all the domain info, and uses that as the host:
        
        sub connect {
            unless ($host) {
        	require Sys::Hostname;
        	my($host_uniq) = Sys::Hostname::hostname();
        	($host) = $host_uniq =~ /([\w\-]+)/;
            }
        
           ....etc.....
        
        Was this intentional? It screws up systems which use an FQDN as their
        hostname and don't list the stripped-down version in their /etc/hosts
        file, because inet_aton fails on the stripped-down name.
        
        Was /([\w\-\.]+)/ what was meant?
        
        -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
                 Alan Schwartz           | Disclaimer: I represent no one
                                         |
        alansz@cogsci.berkeley.edu       | "Life is what happens to you while
        UC Berkeley                      |  you're busy making other plans"
        Cognitive Psychology             |          - J. Lennon
        -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
        
        p5p-msgid: 199705231621.TAA16790@alpha.hut.fi
    
    Subject: Patch to CPAN.pm (perl5.004) for ncftp
    From: "Richard L. Maus <rmaus@monmouth.com>
    Files: lib/CPAN.pm
    Msg-ID: 337FBAC8.167EB0E7@monmouth.com
    
        (applied based on p5p patch as commit 1f8d200)
        
    
    Subject: [PATCH] Harness.pm bug w/perl5.004 & VMS
    From: Dan Sugalski <sugalsd@lbcc.cc.or.us>
    Files: lib/Test/Harness.pm
    Msg-ID: 3.0.1.32.19970530102300.008a2730@stargate.lbcc.cc.or.us
    
        (applied based on p5p patch as commit 2d8ca5d)
    
    Subject: more Fcntl constants [PATCH]
    From: Jarkko Hietaniemi <jhi@iki.fi>
    Files: ext/Fcntl/Fcntl.pm ext/Fcntl/Fcntl.xs
    
     PORTABILITY
    
    Subject: win32: additional default libraries
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: lib/ExtUtils/MM_Win32.pm
    Msg-ID: 199705291332.JAA21560@aatma.engin.umich.edu
    
        (full description at 14ba8c9)
        (base was difficult to find)
    
    Subject: clean up perlocal.pod output on VMS
    From: Peter Prymmer <pvhp@forte.com>
    Files: lib/ExtUtils/MM_VMS.pm
    
    Subject: Re: Term::ReadKey on Win32: set console
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: lib/Term/ReadLine.pm
    
    Subject: Pod::Text nit for Win32
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: lib/Pod/Text.pm
    
    Subject: MakeMaker stumbles on Win32 UNC paths
    From: Warren Jones <wjones@TC.FLUKE.COM>
    Files: lib/ExtUtils/MM_Win32.pm
    
    Subject: build problem on SGI R10000 PowerChallenge (IRIX 6.2) lseek proto
    From: Jarkko Hietaniemi <jhi@iki.fi>
    Files: doio.c
    
    Subject: Perl 5.004 + Linux 2.0.30 & semctl()
    From: Andy Dougherty <doughera@fractal.phys.lafayette.edu>
    Files: doio.c
    
    Subject: lib/io_udp.t fails on VMS
    From: Jonathan Hudson <Jonathan.Hudson@jrhudson.demon.co.uk>
    Files: pp_sys.c
    Msg-ID: XFMail.970522181042.Jonathan.Hudson@jrhudson.demon.co.uk
    
        (applied based on p5p patch as commit 490ab35)
    
    Subject: (NEXT|OPEN)STEP hints
    From: Gerd Knops <gerti@BITart.com>
    Files: hints/next_3.sh hints/next_4.sh
    
    Subject: win32: user defined shell
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: pod/perlrun.pod win32/win32.c
    Msg-ID: 199705291339.JAA21682@aatma.engin.umich.edu
    
        (applied based on p5p patch as commit 174c211)
    
    Subject: win32: Configure cf_email
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: win32/Makefile win32/config.bc win32/config.vc win32/config_sh.PL win32/makefile.mk
    Msg-ID: 199705301335.JAA05079@aatma.engin.umich.edu
    
        (applied based on p5p patch as commit d484a82)
    
    Subject: README.win32 nits
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: README.win32
    
    Subject: SVR4 hints for DDE SMES Supermax Enterprise Server
    From: Jarkko Hietaniemi <jhi@iki.fi>
    Files: hints/svr4.sh
    
    Subject: porting.help
    From: Tim Bunce <Tim.Bunce@ig.co.uk>
    Files: Porting/pumpkin.pod Porting/preprel
    
    Subject: "Major 5.004 Win32 update (Borland win32 support, and other patches)",
    	 "($a,undef,$b) = qw(a b c) and ties delaying DESTROY fixes"
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: MANIFEST pod/perlguts.pod win32/include/sys/socket.h EXTERN.h opcode.h perl.h regcomp.h ext/Fcntl/Fcntl.pm ext/SDBM_File/Makefile.PL lib/ExtUtils/Install.pm lib/ExtUtils/MM_Unix.pm lib/ExtUtils/MM_Win32.pm lib/ExtUtils/MakeMaker.pm lib/ExtUtils/Mksymlists.pm lib/File/DosGlob.pm t/op/mkdir.t t/op/stat.t win32/win32.h win32/win32io.h win32/win32iop.h README.win32 doio.c gv.c mg.c op.c perlio.c pp.c pp_ctl.c pp_hot.c pp_sys.c util.c win32/Makefile win32/config.bc win32/config.vc win32/config_H.bc win32/config_H.vc win32/makedef.pl win32/makefile.mk win32/makeperldef.pl win32/perlglob.c win32/perllib.c win32/win32.c win32/win32io.c win32/win32sck.c
    
    Subject: Re: Maintenance release (remove PERL_DUMMY_SIZE)
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: opcode.h perl.h regcomp.h win32/win32.h gv.c
    
    Subject: ENV leaks on win32 (was Re: Comments on ENV patch sought)
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: win32/win32.h win32/win32io.h win32/win32iop.h global.sym mg.c perl.c t/op/magic.t util.c win32/makedef.pl win32/win32.c win32/win32io.c
    
    Subject: win32: ExtUtils::Liblist support
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: lib/ExtUtils/Liblist.pm win32/Makefile win32/config.bc win32/makefile.mk
    
    Subject: Re: borland C++Perl embedding failures re __declspec()
    From: Gurusamy Sarathy <gsar@engin.umich.edu>
    Files: win32/win32.c
    
     TESTS
    
    Subject: Tests depend on locale
    From: "Jan D." <jan.djarv@mbox200.swipnet.se>
    Files: t/lib/safe2.t t/op/mkdir.t
    Msg-ID: 199705191230.PAA21070@alpha.hut.fi,
    	199705191127.NAA08148@ostrich.gaia.swipnet.se
    
        (applied based on p5p patch as commit 3458556)
    
    Subject: More simple regexp tests and test docs
    From: Hans Mulder <hansm@euronet.nl>
    Files: t/op/re_tests t/op/regexp.t
    
    Subject: Re: Using undef to ignore values returned from split
    From: Hugo van der Sanden <hv@crypt.compulink.co.uk>
    Files: t/op/split.t
    
     UTILITIES
    
    Subject: bad test of -A flag in h2xs
    From: "Jeffrey S. Haemer" <jsh@woodcock.boulder.qms.com>
    Files: utils/h2xs.PL
    
    Subject: h2xs missing from utils/Makefile
    From: Hans Mulder <hansm@euronet.nl>
    Files: utils/Makefile
    
    Subject: final newline missing in MANIFEST generated by h2xs
    From: Hans Mulder <hansm@euronet.nl>
    Files: utils/h2xs.PL
    Perl 5 Porters committed with Tim Bunce Jun 11, 1997
  3. Tests depend on locale

    This is (sort of) documented in INSTALL.  But you are right,
    the tests should force the LC_ALL to be 'C'.  Note: LANG is
    not enough, and neither is POSIX as readily available as C.
    Here are the patches:
    
    p5p-msgid: 199705191230.PAA21070@alpha.hut.fi
    Signed-off-by: Jarkko Hietaniemi <jhi@iki.fi>
    Jan D committed with Tim Bunce May 19, 1997
  4. PATCH: bug in perlbug w.r.t. environment variables

    Here is patch for the problem, this adds sorting all of the environment
    (I see little point in sorting just some).
    
    p5p-msgid: 199705191857.VAA09154@alpha.hut.fi
    Signed-off-by: Jarkko Hietaniemi <jhi@iki.fi>
    Jan D committed with Tim Bunce Jun 11, 1997
  5. win32: Configure cf_email

    Subject: [PATCH] win32: cf_email (was Re: Win32 embedding w/ 5.004)
    
    On Fri, 30 May 1997 11:01:57 +0800, "Gary Ng (Systek)" wrote:
    >Nick and Sarathy contribute much more that me. BTW, anyone 
    >notice that perl -V on win32 still show 'garyng' no matter who build it
    >:-)
    >Can that be removed ?
    
    Actually perl -V doesn't show that for me, but perlbug does.
    Here's a patch.
    
    p5p-msgid: 199705301335.JAA05079@aatma.engin.umich.edu
    Gurusamy Sarathy committed with Tim Bunce May 27, 1997
  6. h2xs missing from utils/Makefile

    Hans Mulder committed with Tim Bunce Jun 11, 1997
  7. win32: user defined shell

    This patch replaces the patch in win32.3 in the gnat collection.
    
    The reasons for reworking that patch should be evident from
    reading the patch below.
    
    p5p-msgid: 199705291339.JAA21682@aatma.engin.umich.edu
    Gurusamy Sarathy committed with Tim Bunce May 24, 1997
  8. Re: Using undef to ignore values returned from split

    Hugo van der Sanden committed with Tim Bunce Jun 11, 1997
  9. lib/io_udp.t fails on VMS

    perl 5.004 built without error or warning on VMS AXP/DECC with DECCRTL
    (UCX) sockets (no sockshr library). However it fails the
    lib/io_udp.t test for the following reasons:
    
    1. The 'fromlen' parameter in pp_sysread *must* be sizeof(struct
       sockaddr) or the DECCRTL fails with an invalid buffer size error.
    
    2. The DECCRTL/UCX getpeerhost() function returns defined and a blank
       'sockaddr' for udp hosts. A similar fix to that in vms/sockadapt.h
       (for sockshr) is required for DECCRTL in pp_sys.c
    
    The following diff (unix, sorry VMS folks) patches pp_sys.c so that
    the udp test is successful using UCX.
    
    p5p-msgid: XFMail.970522181042.Jonathan.Hudson@jrhudson.demon.co.uk
    Jonathan Hudson committed with Tim Bunce May 22, 1997
  10. More simple regexp tests and test docs

    Hans Mulder committed with Tim Bunce Jun 11, 1997
  11. win32: additional default libraries

    I found the default libs list is still incomplete (incomplete
    enough for Tk to fail to build).
    
    This patch adds the additional default libraries to MM_Win32.pm.
    Note LDLOADLIBS can be overridden from Makefile.PL (although
    Tk/win32 seems to want to set it directly in its Makefile,
    rather than via an attribute list, and the direct setting gets
    overridden by the default because the default comes later in
    the makefile).
    
    The patch is against the jumbo patch that nobody seems to like.
    It is *my* new baseline, sorry.
    
     - Sarathy.
       gsar@engin.umich.edu
    
    p5p-msgid: 199705291332.JAA21560@aatma.engin.umich.edu
    Gurusamy Sarathy committed with Tim Bunce May 27, 1997
  12. [dummy merge]

    This merge exists so that the p5p version of the patch and the applied
    version are both in the history
    Tim Bunce committed Jun 11, 1997
  13. Compiling perl5.004 on NEWS-OS 4.x

    Here (attached below) is a patch for compiling perl5 on NEWS-OS 4.x. It
    contains a new hints file for NEWS-OS 4.x and some modification to
    Configure.  I've tested on perl5.003, perl5.003_99 and perl5.004.
    
    Background: Current release of NEWS-OS is version 6.1.x and it is
    SystemV variant. However, NEWS-OS 4.x is 4.3BSD-variant and commonly
    used in Japan. So, I named 'newsos' for newer release, and 'newsos4'
    for older release. Unfortunately, NEWS-OS 4.x has no uname(1).
    
    p5p-msgid: 19970521132814F.matusita@ics.es.osaka-u.ac.jp
    Makoto MATSUSHITA (=?ISO-2022-JP?B?GyRCJF4kRCQ3JD8kXiQzJEgbKEI=?=) committed with Tim Bunce May 21, 1997
  14. op/groups test fails on Linux (groups in /bin)

    (this is the same change as commit 7b7ef14, but as applied)
    Tim Bunce committed Jun 11, 1997
  15. op/groups test fails on Linux (groups in /bin)

    The test op/groups checks for groups in /usr/bin and /usr/ucb.
    This fails on some Linux distributions that has groups in /bin.
    
    A better test would be to just run groups (take whatever is in the
    path) and see if it outputs a list that includes the current
    group.  Possibly you could prepend /bin:/usr/bin:/usr/ucb to the
    current path.  Something like this.
    
    p5p-msgid: 199705191120.NAA08130@ostrich.gaia.swipnet.se
    Jan D committed with Tim Bunce Jun 11, 1997
  16. INSTALL-1.18

    This patch fixes two problems in the INSTALL file in 5.004.  This fix is
    appropriate for 5.004_01.
    
    First, the INSTALL file erroneously claims you can set all the install*
    variables from the Configure command line.  You can't.  That's awaiting my
    -Dinstallprefix patch, which is so far down on the ToDo list that it may
    never get done.
    
    Second, I misspelled 'override'.  I had two 'v's and one 'r'. :-)
    
    p5p-msgid: Pine.SOL.3.95q.970529142739.662D-100000@fractal.lafayette.edu
    Andy Dougherty committed with Tim Bunce May 29, 1997
  17. SVR4 hints for DDE SMES Supermax Enterprise Server

    jhi committed with Tim Bunce Jun 11, 1997
  18. 5.004 POD stuff

    Caught a small typo in perldiag.pod; patch below.
    
    p5p-msgid: 9705191839.AA28702@lemming.engeast
    Signed-off-by: "Paul D. Smith" <psmith@BayNetworks.COM>
    Signed-off-by: Gurusamy Sarathy <gsar@engin.umich.edu>
    Signed-off-by: Jarkko Hietaniemi <jhi@iki.fi>
    Signed-off-by: Michael R Cook <mcook@cognex.com>
    Signed-off-by: Jim Avera <avera@hal.com>
    Signed-off-by: Larry W. Virden <lvirden@cas.org>
    Darren/Torin/Who Ever committed with Tim Bunce May 8, 1997
  19. misc perl5.004 doc fixes

    here are a few doc changes that fix a few typos, etc.
    
    p5p-msgid: 199705160419.AAA16317@cas.org
    Larry W. Virden committed with Tim Bunce Jun 11, 1997
  20. pod2man gags if "=pod" is before "=head1 NAME"

    This is strictly a pod2man issue:
    when using =pod and =cut to intermingle docs into a perl program,
    pod2man does not allow the "=pod" directive to appear before the
    "=head1 NAME" line, and assumes that your pod manpage is improper.
    
    This bug was a problem in 5.003 as well, and here's the patch...
    
    p5p-msgid: 9705212115.AA21730@pezz.sps.mot.com
    Warren Hyde committed with Tim Bunce May 20, 1997
  21. (NEXT|OPEN)STEP hints

    Gerd Knops committed with Tim Bunce Jun 11, 1997
  22. pod2man produces broken pages

    Georg Moritz <gm@mgm-net.de> wrote:
    
    >A single quote in a pod file beginning a line, translated to \*(L'
    >by B<pod2man> lets the formatter choke, i.e. the line is (in the best
    >case) treated as a comment.
    
    I think it would be better to leave the pods alone, and fix pod2man:
    
    p5p-msgid: 1997May25.192350.2055977@hmivax.humgen.upenn.edu
    Signed-off-by: Hans Mulder <hansm@icgned.nl>
    Davin Milun committed with Tim Bunce May 25, 1997
  23. pathname bug in xsubpp on win32

    If xsubpp is run with the argument "c:\i.xs", it will
    generate bad #line directives, like:
    
     #line 1 "c:\i.xs"
    
    which will cause VC++ to barf over a bad "\i"
    escape sequence.
    
    p5p-msgid: 199705230126.SAA23401@clamp.netlabs.com
    Jonathan Biggar committed with Tim Bunce Jun 11, 1997
  24. pod2man produces broken pages

    Georg Moritz <gm@mgm-net.de> wrote:
    
    >A single quote in a pod file beginning a line, translated to \*(L'
    >by B<pod2man> lets the formatter choke, i.e. the line is (in the best
    >case) treated as a comment.
    
    I think it would be better to leave the pods alone, and fix pod2man:
    
    p5p-msgid: 1997May25.192350.2055977@hmivax.humgen.upenn.edu
    private-msgid: 199705310447.AAA15721@obelix.cs.Buffalo.EDU
    Signed-off-by: Hans Mulder <hansm@icgned.nl>
    Davin Milun committed with Tim Bunce May 25, 1997
  25. Pod::Text nit for Win32

    Gurusamy Sarathy committed with Tim Bunce Jun 11, 1997
  26. perlguts caveats

    I've planted more warning signs per Ilya request.
    
    p5p-msgid: 199705301341.JAA05204@aatma.engin.umich.edu
    Signed-off-by: Ilya Zakharevich <ilya@math.ohio-state.edu>
    Gurusamy Sarathy committed with Tim Bunce May 27, 1997
  27. Re: Term::ReadKey on Win32: set console

    Gurusamy Sarathy committed with Tim Bunce Jun 11, 1997
  28. perlguts caveats

    Subject: Re: perlguts additions 
    
    On Sat, 17 May 1997 21:35:50 EDT, Spider Boardman wrote:
    >Your first referenct to `mg_get', here:
    >
    >+a tied array, the caller will usually need to call C<mg_get(val)> to
    >
    >should be to `mg_set'.
    >
    
    Oops, thanks.
    
    p5p-msgid: 199705180202.WAA22826@aatma.engin.umich.edu
    Signed-off-by: Ilya Zakharevich <ilya@math.ohio-state.edu>
    Gurusamy Sarathy committed with Tim Bunce May 17, 1997
  29. clean up perlocal.pod output on VMS

    Peter Prymmer committed with Tim Bunce Jun 11, 1997
  30. perlguts caveats

    Subject: perlguts additions
    
    Here are some additions that describe tied hash and array access.
    
    p5p-msgid: 199705180052.UAA22066@aatma.engin.umich.edu
    Signed-off-by: Ilya Zakharevich <ilya@math.ohio-state.edu>
    Gurusamy Sarathy committed with Tim Bunce Apr 30, 1997
  31. [dummy merge]

    This merge exists so that the p5p version of the patch and the applied
    version are both in the history
    Tim Bunce committed Jun 11, 1997
  32. Patch to show @INC when require dies

    When 'require' dies because it can't find the file in @INC, it would
    be helpful if perl printed out the contents of @INC, to help debu
    the problem (especially when perl is invoked in devious ways, such
    as via a C application in which perl is embedded with funny secret
    -I args passed to perl_parse!).
    
    I would like to contribute a patch to 5.004 to do just that.
    
    p5p-msgid: 9705230121.AA27872@membrane.hal.com
    Jim Avera committed with Tim Bunce May 22, 1997
  33. perl5.004 on AIX: Patches

    (this is the same change as commit a9581ec, but as applied)
    Tim Bunce committed Jun 11, 1997
  34. perl5.004 on AIX: Patches

    I took a further look at the problems which I described in my previous
    post, and managed to make a patch for the -Duseperlio problem. [this]
    makes both -Duseperlio and -Dshrplib work [...]
    
    Should have spotted this earlier, but hey...
    
    p5p-msgid: Pine.A32.3.93.970519163700.25188A-100000@junior.uwc.ac.za
    Peter van Heusden committed with Tim Bunce Jun 11, 1997
  35. ENV leaks on win32 (was Re: Comments on ENV patch sought)

    Subject: [PATCH] for NETaa13787: %ENV=(); doesn't clear the environment
    
    Perl maintains two representations of the environment:
    
    (A) a hash named %ENV, used by the perl script
    (B) a char** named environ, which is passed to child processes
    
    Obviously, the intent is to keep tho two in sync.
    This fails in two situations:
    
    (1) A list assignment to %ENV clears (A) but not (B);
    (2) Assigning to $0 has the side effect of deleting the key
    	NoNeSuCh form (B) but not from (A).
    
    $ perl -e '%ENV=(); print "home\n" if exists $ENV{HOME}; exec "echo \$HOME";'
    /Users/hansm
    $ perl -e '$ENV{NoNeSuCh} = "foo"; $0 = "bar"; exec "echo \$NoNeSuCh";'
    
    $ perl -e '$ENV{NoNeSuCh} = "foo"; exec "echo \$NoNeSuCh";'
    foo
    $ 
    
    I've complained about rpoblem (1) before; and Larry assigned
    it bug ID NETaa13787 when he entered it into DDTS.
    
    The patch below attempts to remedy both problems, at least on
    Unix platforms.  I don't know how to handle the environment
    on VMS and WIN32; my code simply calls DIE('unimplemented"),
    which is honest but won't make users on those plaforms happy.
    
    p5p-msgid: 199705292240.AAA01135@mail.euronet.nl
    Signed-off-by: Peter Prymmer <pvhp@forte.com>
    Hans Mulder committed with Tim Bunce May 29, 1997