Re: Inexplicably slow regex

Can't confirm your observations:

use Benchmark qw( cmpthese );

$_ = '
          df asdf sadf s f
          df asdf sadf s f
          df asdf sadf s f
          df asdf sadf s f
          df asdf sadf s f
          df asdf sadf s f
          df asdf sadf s f
          df asdf sadf s f
';
our $KEY="df";

cmpthese(-3, {
   caret => '/^ \s* $KEY /mx',
   cr => '/\n \s* $KEY /mx',
});
[download]

Result

Rate caret cr
cr 1118569/s -- -1%
caret 1133366/s 1% --

s$$([},&%#}/&/]+}%&{})*;#$&&s&&$^X.($'^"%]=\&(|?*{%
+.+=%;.#_}\&"^"-+%*).}%:##%}={~=~:.")&e&&s""`$''`"e

Comment on Re: Inexplicably slow regex Select or Download Code

Replies are listed 'Best First'.
Re^2: Inexplicably slow regex by Skeeve (Parson) on Sep 12, 2006 at 20:13 UTC
Even with 0.5MB: Still can't confirm! `use Benchmark qw( cmpthese ); our $KEY="df"; { local $/; open my $hosts,'/etc/hosts'; $_= <$hosts>; close $hosts; } $_= $_ x (0.510241024 / length $_); s/$KEY/XX/g; $_.= "\n df "; print length $_,"\n"; cmpthese(-3, { caret => '/^ \s* $KEY /mx', cr => '/\n \s* $KEY /mx', });` [download] Result `524114` `Rate cr caret` `cr 1108181/s -- -1%` `caret 1118184/s 1% --` `s$$([},&%#}/&/]+}%&{});#$&&s&&$^X.($'^"%]=\&(\|?{%` `+`.+=%;.#_}\&"^"-+%*).}%:##%}={~=~:.")&e&&s""`$''`"e	[reply] [d/l] [select]
Re^3: Inexplicably slow regex by Anonymous Monk on Sep 13, 2006 at 00:22 UTC
It looks like it needs many lines to have a partial match in order to slow things down. Try executing this line before adding the matching line: `s/^/ /mg;` [download]	[reply] [d/l]
Re^4: Inexplicably slow regex by Skeeve (Parson) on Sep 13, 2006 at 06:48 UTC
I've put the line just before `$_.= "\n df ";` and still no big difference: `542450` `Rate cr caret` `cr 1105787/s -- -2%` `caret 1131435/s 2% --` `s$$([},&%#}/&/]+}%&{});#$&&s&&$^X.($'^"%]=\&(\|?{%` `+`.+=%;.#_}\&"^"-+%*).}%:##%}={~=~:.")&e&&s""`$''`"e	[reply] [d/l] [select]
Re^5: Inexplicably slow regex by Anonymous Monk on Sep 13, 2006 at 16:45 UTC
Re^2: Inexplicably slow regex by Anonymous Monk on Sep 12, 2006 at 19:47 UTC
Perhaps the difference is that the text I'm searching (read from a file) is 0.5MB. Also, I only expect one match in the entire file so your example may not show a difference since a match is found immediately.	[reply]