regex grouping issue

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

sadly, i need to perform some regex operations on a myspace page
i'm running into a wall with tuning a regex to parse a user id link
This works :

my $RE_myspace_user= qr/myspace\.com\/([\d\w]*)(:?\/|$)/;
[download]

However, that could match some unwanted things, so I updated it... This works (limit out their index and browse page ):

my $RE_myspace_user= qr/myspace\.com\/[^(?:index\.cfm|browse)([\d\w]*)
+](:?\/|$)/;
[download]

But thats' messy - sure they only use index.cfm, but something would be better to block all .cfm pages, so... This doesn't work:

my $RE_myspace_user= qr/myspace\.com\/[^(?:[\w\d\_]\.cfm|browse)([\d\w
+]*)](:?\/|$)/;
[download]

it just generates an erro i guess i can't put a character class in that nesting, which is fine -- does anyone have a suggestion to approximate it though?

Comment on regex grouping issue Select or Download Code

Replies are listed 'Best First'.
Re: regex grouping issue by Fletch (Bishop) on Aug 31, 2006 at 19:14 UTC
Erm, `[^(?:index\.cfm\|browse)([\d\w]*)]` doesn't mean what you think it means. You're conflating a complemented character class with negative lookahead; look in perlre for `(?!)`.	[reply] [d/l] [select]
Re^2: regex grouping issue by nmerriweather (Friar) on Aug 31, 2006 at 19:21 UTC
i totally forgot about negative lookaheads... that should cut down on a lot of my non-capturing parenthesis. thanks for the tip!	[reply]
Re^3: regex grouping issue by japhy (Canon) on Aug 31, 2006 at 20:28 UTC
The primary problem is that a character class is for characters, not sequences of characters. You can't use a character class to match any of a group of strings, and you can't use a negated character class to NOT match any of a group of strings. 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]
Re: regex grouping issue by Velaki (Chaplain) on Sep 01, 2006 at 10:40 UTC
Something you might wish to consider is the restructuring of the logic to take advantage of DeMorgan's Laws, which I found myself using the other day on a particularly funky regex at work. Essentially, DeMorgan's Laws in my own words are DeMorgan's Laws: The compliment of a conjunction is the disjunction of the compliments. and The compliment of a disjunction is the conjunction of the compliments. In other words, not(P and Q) = (not-P or not-Q), et vice versa. In my application, rather than using alternate lists of certain characters with positive lookarounds, I looked for the match text to have none of the characters with the negative lookarounds. (Which is an "AND" condition, and much easier to perform in my case.) Hope this helped a little, -v. "Perl. There is no substitute."	[reply]