From [email protected]

Folks,

looks like the Perl bug system was having some problems the other day. I
am re-sending my bug report, with a few additional notes​:

1. Removing the Autoloader did *not* fix the problem.

2. We are now at a complete loss.

Any suggestions, assistance, additional debugging requests, etc. would be
greatly appreciated.

David.

** WARNING ** This message was composed by a temporarily crippled programmer
using a speech recognition system. Please don't be alarmed if some extremely
strange word usage shows up in the message. Many thanks....

---------- Forwarded message ----------
Date​: Wed, 9 Jun 1999 18​:55​:35 -0400
From​: Mail Delivery Subsystem <MAILER-DAEMON@​perl.org>
To​: davidb@​chelsea.net
Subject​: Returned mail​: unknown mailer error 13

The original message was received at Wed, 9 Jun 1999 18​:55​:30 -0400
from IDENT​:root@​perl.com [199.45.135.9]

  ----- The following addresses had permanent fatal errors -----
"|/home/richard/bugs/scripts/tm_mailproc.pl"
  (expanded from​: <perlbugtron@​perl.org>)

  ----- Transcript of session follows -----
RFgen​::logg(Initial​: [12849] (/home/richard/bugs/logs/pb_19990609) /home/richard/bugs/scripts/tm_mailproc.pl set) can't open today's log (/home/richard/bugs/logs/pb_19990609)! Permission denied at /home/richard/site_perl/RFgen.pm line 193, <STDIN> chunk 150.
554 "|/home/richard/bugs/scripts/tm_mailproc.pl"... unknown mailer error 13

from perl5.

From [email protected]

Delivery Status:
Reporting-MTA: dns; defender.perl.org
Received-From-MTA: DNS; perl.com
Arrival-Date: Wed, 9 Jun 1999 18:55:30 -0400

Final-Recipient: RFC822; [email protected]
X-Actual-Recipient: RFC822; |/home/richard/bugs/scripts/[email protected]
Action: failed
Status: 5.0.0
Last-Attempt-Date: Wed, 9 Jun 1999 18:55:35 -0400

from perl5.

From [email protected]

Message RFC822:
Return-Path: [email protected]
Received: from jhereg.perl.com (IDENT:[email protected] [199.45.135.9])
by defender.perl.org (8.9.3/8.9.3/Debian/GNU) with ESMTP id SAA12846
for [email protected]; Wed, 9 Jun 1999 18:55:30 -0400
Received: from killian.chelsea.net (killian.chelsea.net [207.25.36.14])
by jhereg.perl.com (8.9.0/8.9.0) with ESMTP id GAA09814
for [email protected]; Wed, 9 Jun 1999 06:41:27 -0600
Received: from spaten.chelsea.net (spaten.chelsea.net [207.25.36.16])
by killian.chelsea.net (8.9.1/8.9.1) with ESMTP id IAA03310;
Wed, 9 Jun 1999 08:41:37 -0400 (EDT)
From: David Birnbaum [email protected]
Received: (from davidb@localhost)
by spaten.chelsea.net (8.9.1/8.9.1) id IAA17614;
Wed, 9 Jun 1999 08:41:36 -0400 (EDT)
Date: Wed, 9 Jun 1999 08:41:36 -0400 (EDT)
Message-Id: [email protected]
To: [email protected]
Cc: [email protected]
Subject: Random core dump
MIME-Version: 1.0

This is a bug report for perl from [email protected],
generated with the help of perlbug 1.26 running under perl 5.00502.

[Please enter your report here]

We have suddenly become perplexed with a core dump that occurs
approximately once every three or four times we run a particular Perl
program. Unfortunately, the library being included is rather long,
and I can't include a small sample program to reproduce the error.
However, here is information I can provide:

1. We are using the Autoloader module. The problem seems to go away
   when we turn it off, but we want to run a few more nights to
   make certain.

2. We cannot reproduce the problem when perl is running under a
   "truss", even after dozens of attempts.

3. The problem showed up when we made this change to a signal
   handling routine in the library:

*** 33,42 ****

@isa = qw( AutoLoader );

! $SIG{"CHLD"} = sub {
$Admutil::last_pid = waitpid( -1, WNOHANG );
$Admutil::caught_pid{$Admutil::last_pid} = $?;
};

Some documentation

--- 33,44 ----

@isa = qw( AutoLoader );

! sub REAPER {
! $SIG{"CHLD"} = &REAPER;
$Admutil::last_pid = waitpid( -1, WNOHANG );
$Admutil::caught_pid{$Admutil::last_pid} = $?;
};

• $SIG{"CHLD"} = &REAPER;

  Some documentation

  which was to try and debug a problem with "unknown subroutine"
  appearing another part of the code - occasionally (suspicious,
  me thinks).

4. I am no gdb expert, but:

(gdb) where
#0 0x6b0a0 in Perl_newSV ()
#1 0x7ee3c in save_scalar_at ()
#2 0x7f028 in Perl_save_scalar ()
#3 0x60204 in Perl_pp_gvsv ()
#4 0x9f0d0 in Perl_runops_standard ()
#5 0x27ddc in perl_run ()
#6 0x2594c in main ()

is what it has to say about where it is dying.

5. I do not think this library is doing anything particularly esoteric
   in terms of its programming or functionality. I've gone through the
   code by inspection (I originally wrote it) to look for anything
   strange, but nothing jumps out at me.

I'm sorry I cannot include anything more useful. Hopefully this will
jog somebody's memory about a potential cause or can give us a few
new ideas to try.

Thanks,

David.

[Please do not change anything below this line]

Site configuration information for perl 5.00502:

Configured by mprewitt at Sat Dec 5 09:54:18 EST 1998.

Summary of my perl5 (5.0 patchlevel 5 subversion 2) configuration:
Platform:
osname=solaris, osvers=2.6, archname=sun4-solaris
uname='sunos corona 5.6 generic_105181-09 sun4m sparc sunw,sparcstation-20 '
hint=previous, useposix=true, d_sigaction=define
usethreads=undef useperlio=undef d_sfio=undef
Compiler:
cc='gcc', optimize='-O', gccversion=2.8.1
cppflags='-I/opt/bind/8.1.2/include -I/usr/local/include'
ccflags ='-I/opt/bind/8.1.2/include -I/usr/local/include'
stdchar='unsigned char', d_stdstdio=define, usevfork=false
intsize=4, longsize=4, ptrsize=4, doublesize=8
d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=16
alignbytes=8, usemymalloc=y, prototype=define
Linker and Libraries:
ld='gcc', ldflags ='-L/opt/bind/8.1.2/bind/lib -L/usr/local/lib'
libpth=/usr/local/lib /lib /usr/lib /usr/ccs/lib
libs=-lsocket -lnsl -ldl -lm -lbind -lc -lcrypt
libc=/lib/libc.so, so=so, useshrplib=false, libperl=libperl.a
Dynamic Linking:
dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=undef, ccdlflags=' '
cccdlflags='-fPIC', lddlflags='-G -L/usr/local/lib -L/opt/bind/8.1.2/bind/lib'

Locally applied patches:

@inc for perl 5.00502:
/opt/perl/5.005_02/lib/perl/sun4-solaris
/opt/perl/5.005_02/lib/perl
/opt/siteperl/5.005_02/siteperl/sun4-solaris
/opt/siteperl/5.005_02/siteperl
.

Environment for perl 5.00502:
HOME=/home/davidb
LANG (unset)
LD_LIBRARY_PATH (unset)
LOGDIR (unset)
PATH=/home/davidb/bin:/usr/local/bin:/usr/sbin:/usr/bin:/usr/ccs/bin:/usr/ucb:.
PERL_BADLANG (unset)
SHELL=/bin/ksh

from perl5.

From [Unknown Contact. See original ticket]

Folks,

We are proceeding under the assumption that this bug is actually a core
dump from a malloc() call in a signal handler (as per various and sundry
comments in the Perl book). Is there any way to verify this? In any
case, we are modifying our code to see if that is indeed the case. Wish
we knew why it suddenly started....

Thanks (if anyone is indeed out there).

David.

** WARNING ** This message was composed by a temporarily crippled programmer
using a speech recognition system. Please don't be alarmed if some extremely
strange word usage shows up in the message. Many thanks....

---------- Forwarded message ----------
Date​: Mon, 14 Jun 1999 09​:35​:52 -0400
From​: Richard Foley <richard@​perl.org>
To​: David Birnbaum <davidb@​chelsea.net>
Subject​: [ID 19990614.001 Perl core dump]

Your e-mail has been received by the Perl Bug Squashing Team.

Ticket ID '[19990614.001]' has been assigned. Please include this ticket ID in the
subject line of any followup messages related to this issue.

This is an automatic confirmation message.
--
Perl Bug Squashing Team
perlbug@​perl.org
 

from perl5.