Re^4: Tokenizing and qr// <=> /g interplay

Hi, I was just refactoring some code and saw a possible opportunity to use this advice. But. Instead of having multiple strings in @strings to process, I leave all the lines from my file joined together as one giant string with embedded \n chars. From this angle, since I only have to use each regex once across all the strings via them being 'joined' into one string, I won't benefit from qr.

    my ( $crummy, $good );
    foreach my $crummy_good_ar ( @corrections_to_make ) {
        ( $crummy, $good ) = @$crummy_good_ar;
        $file_in_string_form =~ s/\b(\Q$crummy\E)\b/$good/ig;
    }
[download]

However, as you can see (?) from the example above, I have lots of crummy/good switchouts to do, and is my plodding approach above the best that can be expected?

P.S. Can you clarify/update what you meant by:
the benefit of qr// objects is lost if there is additional text in the pattern match

I think you are saying that a precompiled/qr regex used in a follow-on regex will have to be recompiled if you snap additional text on to the qr'd variable, because the overall text of the new regex will be different. Although at least one would still have the benefit of 'concentrated regex logic' within the qr'd variable?

Comment on Re^4: Tokenizing and qr// <=> /g interplay Download Code

Replies are listed 'Best First'.
Re^5: Tokenizing and qr// <=> /g interplay by japhy (Canon) on Apr 25, 2005 at 12:36 UTC
I can answers your first question by answering your second one. If you made all the first elements of your array references Regexp objects: `$_->[0] = qr/\b(\Q$_->[0]\E)b/ for @corrections_to_make;` [download] then you could do your loop as `for my $crummy_good_ar (@corrections_to_make) { my ($crummy, $good) = @$crummy_good_ar; $file_in_string_form =~ s/$crummy/$good/ig; }` [download] This way, even if you end up looping over THAT code, you'd still be dealing with already-compiled regexes. As soon as you put additional text into a regex with qr// in it: `my $rx = qr/abc/; if ($str =~ /^($rx)$/) { ... }` [download] Perl has to do the "compare physical regex forms" test. Only if the qr// object is all alone will it have the entire benefits it was made for. Jeff `japhy` Pinyan, P.L., P.M., P.O.D, X.S.: Perl, regex, and `perl` hacker How can we ever be the sold short or the cheated, we who for every service have long ago been overpaid? ~~ Meister Eckhart	[reply] [d/l] [select]
Re^6: Tokenizing and qr// <=> /g interplay by ff (Hermit) on Apr 25, 2005 at 13:37 UTC
Okay, so an array of `[ qr, $good ]` combos could help me if my example was something like: `for my $file_in_string_form ( @all_files ) { for my $crummy_good_ar (@corrections_to_make) { my ($crummy, $good) = @$crummy_good_ar; $file_in_string_form =~ s/$crummy/$good/ig; } }` [download] And in the meantime, if I have no additional files to process, I might as well compile the regexen as needed? Thanks! I think I understand now. :-)	[reply] [d/l] [select]