Regular expression match (?:XX|X)* dumps core

[p5pRT]

Migrated from rt.perl.org#938 (status was 'resolved')

Searchable as RT938$

[p5pRT]

From garethr@cre.canon.co.uk

The following program causes a segmentation violation​:

  $_ = 'X' x 50000;
  /(?​:XX|X)*/;

Perl Info

Site configuration information for perl 5.00502:

Configured by abw at Thu Nov 19 10:50:08 GMT 1998.

Summary of my perl5 (5.0 patchlevel 5 subversion 2) configuration:
  Platform:
    osname=solaris, osvers=2.6, archname=sun4-solaris
    uname='sunos bandanna 5.6 generic_105181-03 sun4u sparc sunw,ultra-1 '
    hint=previous, useposix=true, d_sigaction=define
    usethreads=undef useperlio=undef d_sfio=undef
  Compiler:
    cc='gcc', optimize='-O', gccversion=2.8.0
    cppflags='-I/opt/gnu/include -I/user/perl/build/include'
    ccflags ='-I/opt/gnu/include -I/user/perl/build/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=n, prototype=define
  Linker and Libraries:
    ld='gcc', ldflags ='-L/opt/gnu/lib -L/user/perl/build/lib'
    libpth=/opt/gnu/lib /lib /usr/lib /usr/ccs/lib /user/perl/build/lib
    libs=-lsocket -lnsl -lgdbm -ldl -lm -lc -lcrypt -ldb
    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/opt/gnu/lib -L/user/perl/build/lib'

Locally applied patches:
    


@INC for perl 5.00502:
    /apps/perl5/lib/5.00502/sun4-solaris
    /apps/perl5/lib/5.00502
    /apps/perl5/lib/site_perl/5.005/sun4-solaris
    /apps/perl5/lib/site_perl/5.005
    .


Environment for perl 5.00502:
    HOME=/user/garethr
    LANG=C
    LD_LIBRARY_PATH=/usr/dt/lib:/usr/openwin/lib:/apps/X11R5/lib:/usr/lib
    LOGDIR (unset)
    PATH=/usr/dt/bin:/usr/openwin/bin:/bin:/usr/bin:/usr/ucb:/bin:/usr/bin:/cre/bin:/apps/gcc-2.8.0/bin:/apps/frame5/bin:/apps/emacs/bin:/apps/tex/bin:/apps/perl5/bin:/user/web/bin/scripts:/user/web/bin/SunOS-5.6:/apps/gnu/bin:/apps/mysql/bin:/usr/ccs/bin:/apps/X11R5/bin:/usr/openwin/bin
    PERL_BADLANG (unset)
    SHELL=/cre/bin/bash

[p5pRT]

From [Unknown Contact. See original ticket]

This regexp bug can _still_ cause a segfault in bleadperl as of today.

--cut--
#!/usr/local/bin/perl

$_ = 'X' x 50000;
/(?​:XX|X)*/;
--cut--

[p5pRT]

From @vanstyn

In <20000918115335.A27376@​athenaeum.globnix.org>, Michael Stevens writes​:
:This regexp bug can _still_ cause a segfault in bleadperl as of today.
:
:--cut--
:#!/usr/local/bin/perl
:
:$_ = 'X' x 50000;
:/(?​:XX|X)*/;

Yes, this is not likely to be fixed any time soon​: the recursion exceeds
your stack size. Increasing the stack size for perl may allow you to
match longer strings before it breaks, but a fix involves flattening
the recursion out of the regexp matcher, which is a) difficult and b)
not of the highest priority. But patches welcome, of course. :)

Hugo

[p5pRT]

From [Unknown Contact. See original ticket]

On Mon, Sep 18, 2000 at 01​:33​:17PM +0100, Hugo wrote​:

Yes, this is not likely to be fixed any time soon​: the recursion exceeds
your stack size. Increasing the stack size for perl may allow you to
match longer strings before it breaks, but a fix involves flattening
the recursion out of the regexp matcher, which is a) difficult and b)
not of the highest priority. But patches welcome, of course. :)

It would be nice if it could at least say "this is too big, I give up".

Code wise, the regexp stuff scares me.

Michael.

[p5pRT]

From @vanstyn

In <20000918141050.A29064@​athenaeum.globnix.org>, Michael Stevens writes​:
:On Mon, Sep 18, 2000 at 01​:33​:17PM +0100, Hugo wrote​:
:> Yes, this is not likely to be fixed any time soon​: the recursion exceeds
:> your stack size. Increasing the stack size for perl may allow you to
:> match longer strings before it breaks, but a fix involves flattening
:> the recursion out of the regexp matcher, which is a) difficult and b)
:> not of the highest priority. But patches welcome, of course. :)
:
:It would be nice if it could at least say "this is too big, I give up".
:
:Code wise, the regexp stuff scares me.

It should do​: that is what the REG_INFTY stuff is there for. However
it is difficult (impossible?) for Configure to determine the right
value to set REG_INFTY to - I believe it is usually too high under
Linux, for example - and I have seen suggestions that the REG_INFTY
check may not be triggered when it should in all cases.

HTH,

Hugo

[p5pRT]

From The RT System itself

still core dumps in bleadperl as of 18/sep/2000, perl 5.7.0 DEVEL7093