regex help required

ghosh123 has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: regex help required by Corion (Patriarch) on Jun 24, 2013 at 07:14 UTC
What are the relevant parts of all these combinations? `/.*/` will match all these above possible combinations. So maybe you want to tell us what combinations are invalid.	[reply] [d/l]
Re: regex help required (question help too) by Anonymous Monk on Jun 24, 2013 at 07:55 UTC
I am just guessing like Corion `m{ \Qabc\E \| \Qabc-def\E \| \Qabc_def.12\E \| \Q.abc\E \| \Q.abc_123\E \| \Q.abcd-def\E \| \Qabcd.123.def\E }x;` [download] Read perlintro#More complex regular expressions or perlrequick#Matching this or that See also Regexp::Assemble/Regexp::Trie/Regex::PreSuf/Regexp::Wildcards P.S. At some point "guessing" stops being fun and help does not follow; I know what I mean. Why don't you? , How to ask questions the smart way.	[reply] [d/l]
Re: regex help required by hdb (Monsignor) on Jun 24, 2013 at 07:49 UTC
`/abc\|abc-def\|abc_def\.12\|\.abc\|\.abc_123\|\.abcd-def\|abcd\.123\.def/` [download]	[reply] [d/l]
Re: regex help required by space_monk (Chaplain) on Jun 24, 2013 at 09:36 UTC
You have inadequately said what you are looking for in your answer. The following will match all the above possible combinations in `$result`, but I strongly suspect that it is not what you want `my $result = $string;` [download] If you spot any bugs in my solutions, it's because I've deliberately left them in as an exercise for the reader! :-)	[reply] [d/l] [select]
Re: regex help required by locked_user sundialsvc4 (Abbot) on Jun 24, 2013 at 13:16 UTC
Indeed. You must not only determine what it is you do want to match (hence the reply #1: “what are the relevant parts?”), but also what rules will filter out what you do not want. Seriously consider using a `sub`routine to allow for some conditional logic to be used ... (pseudocode follows) `sub my_filter { my $str = shift; my $TRUE = 1; my $FALSE = 0; # FOR CLARITY return $FALSE if ($str =~ /reject_pattern_#1/); return $FALSE if ($str =~ /reject_pattern_#2/); return $TRUE if ($str =~ /accept_pattern_#1/); return $TRUE if ($str =~ /accept_pattern_#2/); return $FALSE; // DOESN'T MATCH ANYTHING }` [download] Then, write a test-program using Test::More which puts your subroutine through many dozens of tests, using a carefully-chosen set of “correct,” “almost correct,” “barely incorrect” and “outright bogus” strings, to `prove` that your logic is actually trustworthy and reliable. (Tip: you will spend a lot more time ferreting-out the little bugs that were lurking in your original version, hoping to remain un-detected, than you originally thought possible.)