Take into account that signals behaviour changed in Perl 5.7.3. From then on, signals are deferred and aren't delivered to the program until perl decides it's safe to do so. If you want to get signals immediatly, you'll have to set the environment variable PERL_SIGNALS to the value "unsafe". You can read more about this in perlipc, "Deferred Signals (Safe Signals)".
Update: There's a slow regex in perlre. Try this code with and without the environment variable PERL_SIGNALS:
$SIG{'ALRM'} = sub { die "alrm received"; };
alarm 2;
'aaaaaaaaaaaa' =~ /((a{0,5}){0,5}){0,5}[c]/;
| [reply]
[d/l] |
The alarm does go of for the example you gave me, however, a regex given by leronim below :
my $re = ('a*' x80).'(b|c)';
my $str = "a" x 80;
$str =~ /$re/;
Never sets off the alarm. Any idea why that is the case.
Thanks. | [reply]
[d/l] |
$ perl
$SIG{'ALRM'} = sub { die "alrm received"; };
alarm 2;
my $re = ('a*' x80).'(b|c)';
my $str = "a" x 80;
$str =~ /$re/;
__END__
... runs forever ...
$ PERL_SIGNALS=unsafe perl
$SIG{'ALRM'} = sub { die "alrm received"; };
alarm 2;
my $re = ('a*' x80).'(b|c)';
my $str = "a" x 80;
$str =~ /$re/;
__END__
alrm received at - line 1.
| [reply]
[d/l] |
Can anyone give me a particular regex against a string that would take a long time to complete...
From Progamming Perl:
$_ = "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaab";
/a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*a*[b]/;
"... If you remove the "b" from the string the pattern will probably run for many, many years before failing. Many, many millennia. Acutally, billions and billions of years.*"
later
"* Actually, it's more in the order of septillions and septillions. We don't know how long it would take. We didn't care to wait around watching it not fail. In any event, your computer is likely to crash before the heat death of the universe, and this regular expression takes longer than either of those."
Any help? :-)
Apologies to O'Reilly for any infringments.
| [reply]
[d/l] |
From doc:perlipc for v5.8.5:
As Perl interpreter only looks at the signal flags when it about to execute a new opcode if a signal arrives during a long running opcode (e.g. a regular expression operation on a very large string)the signal will not be seen until operation completes.
Also, think about what you are asking the OS to do: first, you tell the OS to interupt you in x seconds then, you hog all the CPU cycles so nothing else gets done. On a preemptive multitasker this will work but the performance can still suffer.
s//----->\t/;$~="JAPH";s//\r<$~~/;{s|~$~-|-~$~|||s
|-$~~|$~~-|||s,<$~~,<~$~,,s,~$~>,$~~>,,
$|=1,select$,,$,,$,,1e-1;print;redo}
| [reply] |
particular never-ending regex:
my $re = ('a*' x80).'(b|c)';
my $str = "a" x 80;
$str =~ /$re/;
s;;Just-me-not-h-Ni-m-P-Ni-lm-I-ar-O-Ni;;tr?IerONim-?HAcker ?d;print
| [reply]
[d/l]
[select] |