Re: How to capture backwards using regex?

in reply to How to capture backwards using regex?

It looks like you reinvented a poor man's pos()

Different approaches come to mind:

The simple one is to decrement your "pointer" by the max length of a possible integer.
Another, is to also capture the (.{$P}) prelude and to post-process any trailing digits inside
A "pure" regex approach is to use a positive look-behind assertion, see perlretut#Looking-ahead-and-looking-behind

Update

I like Corion's solution best, in terms of pure regex solutions.

Cheers Rolf
_{(addicted to the Perl Programming Language :)

see Wikisyntax for the Monastery}

Comment on Re: How to capture backwards using regex? Download Code

Replies are listed 'Best First'.
Re^2: How to capture backwards using regex? by LanX (Saint) on Nov 10, 2024 at 13:10 UTC
For completeness > A "pure" regex approach is to use a positive look-behind assertion use v5.12; use warnings; use Test::More; say "Version: $]"; sub test { my ($msg, $str, $exp) = @_; my $pos = index($str, ""); die "NO POINTER <$msg> in <$str>" if $pos == -1; substr($str,$pos,1) = ""; die "MULTIPLE POINTERS <$msg> in <$str>" if index($str,"") >-1; no warnings 'experimental'; # no better category available + ??? my $match = join "", ($str =~ m/^.{$pos}(plb:(\d{0,254}))(\d+)/) +; is ($match,$exp,$msg); } test "In the middle", 'World 123456 Hello 789 ' => "123456" ; test "At the start" , 'World 123456 Hello 789 ' => "123456" ; test "At the end" , 'World 123456 Hello 789 ' => "123456" ; test "Before" , 'World 123456 Hello 789 ' => "" ; test "Right before" , 'World 123456 Hello 789 ' => "" ; test "After" , 'World 123456 Hello 789 ' => "" ; test "Right after" , 'World 123456 Hello 789 ' => "" ; test "In next Nr" , 'World 123456 Hello 789 ' => "789" ; test "Nr at start" , '123456 Hello' => "123456" ; test "Nr at end" , 'World 123456' => "123456" ; test "Pos at end" , 'World 123456' => "" ; done_testing; [download] `Version: 5.038002 ok 1 - In the middle ok 2 - At the start ok 3 - At the end ok 4 - Before ok 5 - Right before ok 6 - After ok 7 - Right after ok 8 - In next Nr ok 9 - Nr at start ok 10 - Nr at end ok 11 - Pos at end 1..11` [download] Remarks Unfortunately with variable length this still seems experimental with 5.38 ... `Variable length positive lookbehind with capturing is experimental in regex;` and I had to use a hard limit of `254` instead of `*` `Lookbehind longer than 255 not implemented` Strangely enough I couldn't find a dedicated warnings category and had to disable all "experimental" warnings locally. Updates had to fix tests because of C&P errors leading to multiple pointers, fixed test() to catch those errors. Cheers Rolf _{(addicted to the Perl Programming Language :) see Wikisyntax for the Monastery}	[reply] [d/l] [select]
Re^2: How to capture backwards using regex? by harangzsolt33 (Deacon) on Nov 10, 2024 at 17:05 UTC
Yeah, I've played around with positive look-behind regex, and it just didn't want to work. Of course, I'm using TinyPerl 5.8, so I'm a decade or two behind. lol But surprisingly, Corion's solution works flawlessly on TinyPerl 5.8, so that's great!	[reply]
Re^3: How to capture backwards using regex? by LanX (Saint) on Nov 10, 2024 at 17:53 UTC
Look behind used to only work with fixed length patterns. IIRC. That's most likely also related to the "experimental" status of my code. > But surprisingly, Corion's solution works flawlessly Doesn't surprise me that's a very old feature, could be even Perl4. The idea of `Pos` and related stuff is very sed/awk-ish, and text processing was one of the main objectives of Perl4. Cheers Rolf _{(addicted to the Perl Programming Language :) see Wikisyntax for the Monastery}	[reply]

In Section Seekers of Perl Wisdom