Re: lovely regexs

Perhaps change the regular expression to turn off the default "greediness" of the "*", with a "?" quantifier, so that it gathers only up to the next quote character.

$moo =~ m/src="(.*?)"/;
[download]

However, I normally use HTML::TreeBuilder to parse and search html.

Comment on Re: lovely regexs Download Code

Replies are listed 'Best First'.
Re^2: lovely regexs by dsheroh (Monsignor) on Apr 12, 2009 at 22:20 UTC
Ignoring, for the moment, the wisdom of using the proper tool (which is generally not a regex) for parsing HTML... The issue here is not greediness. The issue is the misuse of ".". Making the "" non-greedy is just a band-aid which masks the fact that "." says "match any number of any* characters", when what you actually mean is "match any number of any non-double quote characters". The correct way to write that regex is: `$moo =~ m/src="([^"]*)"/;` [download] The non-greedy qualifier does have its legitimate uses, generally in cases where your target is terminated by a sequence of multiple characters. In cases where a negated character class can do the job, though, the character class will almost always be the better option.	[reply] [d/l]

Replies are listed 'Best First'.

Re^2: lovely regexs
by dsheroh (Monsignor) on Apr 12, 2009 at 22:20 UTC

The issue here is not greediness. The issue is the misuse of ".*". Making the "*" non-greedy is just a band-aid which masks the fact that ".*" says "match any number of any characters", when what you actually mean is "match any number of any non-double quote characters". The correct way to write that regex is:

$moo =~ m/src="([^"]*)"/;
[download]

The non-greedy qualifier does have its legitimate uses, generally in cases where your target is terminated by a sequence of multiple characters. In cases where a negated character class can do the job, though, the character class will almost always be the better option.

[reply]
[d/l]