From siracusa@gmail.com

This is a bug report for perl from siracusa@​gmail.com,
generated with the help of perlbug 1.36 running under perl 5.10.0.

Much of this report is copied and pasted from the following URL​:

http​://stackoverflow.com/questions/322173

Consider the following code​:

  BEGIN
  {
  package MyEx;

  sub new { bless {}, shift }

  package main;

  eval { die MyEx->new };

  if($@​)
  {
  die "Really die" unless($@​->isa('MyEx'));
  }
  }

  print "OK\n";

Run it and it prints "OK" as expected. Run it in the debugger, however,
and it dies with this error​:

  Can't call method "isa" without a package or object reference ...

The upshot is that $@​ is not an object when the code runs under the
debugger. Instead, it's an unblessed scalar containing this string​:

  " at somefile.pl line 9
  eval {...} called at somefile.pl line 9
  main​::BEGIN() called at somefile.pl line 16
  eval {...} called at somefile.pl line 16
  "

(Internal newlines and spacing preserved. That's the literal text of the
string, including the "..."s.)

Setting dieLevel to 0, either through the use of a ~/.perldb file containing​:

  parse_options('dieLevel=0')

or by setting the PERLDB_OPT environment variablt to "dieLevel=0", did not
solve the problem.

local()izing $SIG{__DIE__} inside the eval block is a workaround, but
should not be necessary and is not done in much real code. This makes it
impossible to use the debugger during the development of such code (e.g.,
several of my larger CPAN modules).

Here's an explanation of the bug from brian d foy​:

This is a problem with perl5db.pl creating __DIE__ handlers. If I localize
$SIG{__DIE__} in your eval, things work as you expect.

  eval
  {
  local $SIG{__DIE__};
  die MyEx->new;
  };

If you don't do that, you're getting the handler from DB​::dbdie, which uses
Carp​::longmess. That shouldn't happen if dieLevel is 0, but by default it
is 1, and it gets set to 1 if it is not defined. This was a patch to
perl5db.pl back in 2001, and previously the default had been 0.

You're supposed to turn this off with​:

  PERLDB_OPT="dieLevel=0" perl5.10.0 -d program

But there is still a code reference in $SIG{__DIE__} after that, and it's a
reference to dbdie. I think this is a bug in handling the global variable
$prevdie in perl5db.pl's dieLevel. At the end of that subroutine, there is​:

  # perl5db.pl dieLevel, around line 7777
  elsif ($prevdie) {
  $SIG{__DIE__} = $prevdie;
  print $OUT "Default die handler restored.\n";
  }

But notice that after restoring $SIG{__DIE__}, it keeps the previous value
in $prevdie, meaning whatever is in there leaks to another call. When I run
that command line, there are two calls to dieLevel before it handles
PERLDB_OPT, so $prevdie is probably dirty.

So, that's as far as I got before I didn't want to think about perl5db.pl
anymore.

It seems like another workaround could be to change the default dieLevel
back to 0 instead of 1.

BTW, the "perldoc perldebug" documentation seems to imply that the default
value of dieLevel is still 0​:

  "By default, the debugger leaves your exceptions and warnings alone,
  because altering them can break correctly running programs."

so apparently that patch in 2001 that brian refers to only updated the
code, but not the docs.

-John

Flags​:
  category=utilities
  severity=high

Site configuration information for perl 5.10.0​:

Configured by john at Sat Sep 20 20​:27​:53 EDT 2008.

Summary of my perl5 (revision 5 version 10 subversion 0) configuration​:
  Platform​:
  osname=darwin, osvers=9.5.0, archname=darwin-2level
  uname='darwin link.local 9.5.0 darwin kernel version 9.5.0​: wed
sep 3 11​:29​:43 pdt 2008; root​:xnu-1228.7.58~1release_i386 i386 '
  config_args=''
  hint=recommended, useposix=true, d_sigaction=define
  useithreads=undef, usemultiplicity=undef
  useperlio=define, d_sfio=undef, uselargefiles=define, usesocks=undef
  use64bitint=undef, use64bitall=undef, uselongdouble=undef
  usemymalloc=n, bincompat5005=undef
  Compiler​:
  cc='cc', ccflags ='-fno-common -DPERL_DARWIN -no-cpp-precomp
-fno-strict-aliasing -pipe -I/usr/local/include',
  optimize='-O3',
  cppflags='-no-cpp-precomp -fno-common -DPERL_DARWIN
-no-cpp-precomp -fno-strict-aliasing -pipe -I/usr/local/include'
  ccversion='', gccversion='4.0.1 (Apple Inc. build 5488)', gccosandvers=''
  intsize=4, longsize=4, ptrsize=4, doublesize=8, byteorder=1234
  d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=16
  ivtype='long', ivsize=4, nvtype='double', nvsize=8, Off_t='off_t',
lseeksize=8
  alignbytes=8, prototype=define
  Linker and Libraries​:
  ld='env MACOSX_DEPLOYMENT_TARGET=10.3 cc', ldflags =' -L/usr/local/lib'
  libpth=/usr/local/lib /usr/lib
  libs=-ldbm -ldl -lm -lutil -lc
  perllibs=-ldl -lm -lutil -lc
  libc=/usr/lib/libc.dylib, so=dylib, useshrplib=false, libperl=libperl.a
  gnulibc_version=''
  Dynamic Linking​:
  dlsrc=dl_dlopen.xs, dlext=bundle, d_dlsymun=undef, ccdlflags=' '
  cccdlflags=' ', lddlflags=' -bundle -undefined dynamic_lookup
-L/usr/local/lib'

Locally applied patches​:

@​INC for perl 5.10.0​:
  /usr/local/lib/perl5/5.10.0/darwin-2level
  /usr/local/lib/perl5/5.10.0
  /usr/local/lib/perl5/site_perl/5.10.0/darwin-2level
  /usr/local/lib/perl5/site_perl/5.10.0
  .

Environment for perl 5.10.0​:
  DYLD_LIBRARY_PATH (unset)
  HOME=/Users/john
  LANG=en_US
  LANGUAGE (unset)
  LC_ALL=C
  LD_LIBRARY_PATH (unset)
  LOGDIR (unset)
  PATH=/usr/local/bin​:/usr/bin​:/bin​:/usr/local/sbin​:/usr/sbin​:/sbin​:/usr/X11R6/bin​:/usr/local/mysql/bin​:/usr/local/pgsql/bin​:/Developer/Tools​:/Users/john/bin​:/usr/local/teTeX/bin/powerpc-apple-darwin6.8​:.
  PERLDB_OPT=dieLevel=0
  PERL_BADLANG (unset)
  PERL_READLINE_NOWARN=
  SHELL=/bin/tcsh

From @rgs

2008/11/28 via RT John Siracusa <perlbug-followup@​perl.org>​:

It seems like another workaround could be to change the default dieLevel
back to 0 instead of 1.

BTW, the "perldoc perldebug" documentation seems to imply that the default
value of dieLevel is still 0​:

"By default, the debugger leaves your exceptions and warnings alone,
because altering them can break correctly running programs."

so apparently that patch in 2001 that brian refers to only updated the
code, but not the docs.

Here's a pointer to the patch :
http​://public.activestate.com/cgi-bin/perlbrowse/p/11120
and the original mail from Ilya, which is not so helpful​:
http​://markmail.org/message/c6dwk4areelar37p?q=debugger+date​:20010703&page=1

I don't know what defaults are preferable for dielevel and warnlevel
in the debugger. I don't use it enough.

The RT System itself - Status changed from 'new' to 'open'

From @kevindawson

brian d foy++

looks like no one has checked this recently either!

well all is now well in perl5db.pl version 1.42

that is what I used to test with.

I suspect this is another item fixed by shlom-fish wake :)

PERLDB_OPTS="dieLevel=0" #works

or we could just reset the default to 0 as suggested by brian as that
works as well.

imho we should close this and discuss else whare wether 0 should be the
new default value

From [Unknown Contact. See original ticket]

brian d foy++

looks like no one has checked this recently either!

well all is now well in perl5db.pl version 1.42

that is what I used to test with.

I suspect this is another item fixed by shlom-fish wake :)

PERLDB_OPTS="dieLevel=0" #works

or we could just reset the default to 0 as suggested by brian as that
works as well.

imho we should close this and discuss else whare wether 0 should be the
new default value