Multiline pattern matching

[Anonymous Monk]

How can I match the following?

foo {
...
...
...
};

Thanks in advance

[Beatnik]

Your description is rather vague... I'll assume

foo {
...
...
...
};

is in a string somehow... something like : $foo =~ /foo {[^};]+};/s; should be ok...

Greetz
Beatnik
... Quidquid perl dictum sit, altum viditur.

[tye]

If the part you left out can contain matching pairs of braces:

foo {
   bar {
   };
};

then I encourage you to try Parse::RecDescent. If the part you left out can also contain braces that are quoted or escaped in some way:

foo {
    bar "No closing } found!"
};

then I strongly encourage you to try Parse::RecDescent.
        - tye (but my friends call me "Tye")

[alfie]

Please read perlman:perlre. Your question is too vague to be answered but you should be able to find what you are looking for in this page.
--
use signature; signature(" So long\nAlfie");

[jeroenes]

Your problem description is somewhat minimal, but your title suggests that you are looking for the 's' modifier. So make your regex like:

$var =~ m/some text\nand more/s;

The last 's' makes sure all text is read like one string, even if it contains newlines.

Of course you could have read this in perlop and prelre.

Hope this helps,

Jeroen
"We are not alone"(FZ)

[blue_cowdawg]

Base Assumption: Your text you are trying to match is somehow in one single variable.

$str =~ m/foo\{(.*)\}\;/;

$1 is now going to contain everything between the curly braces. The flaw I see in my own pattern matching is the fact that if you have another "};" along the way you might not have what you are after.

Then again.. you were pretty vague about your specifications... ;-)

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Peter L. Berghold --- [e-mail]
"Those who fail to learn from history are condemned to repeat it."

[alfie]

Sorry, blue_cowdawg, you do a common mistake: You used dot star - please read Death to Dot Star! which explains why you should never ever use it :-)
--
use signature; signature(" So long\nAlfie");

[blue_cowdawg]

Your Death to Dot Star post is rather convincing!

I like merlyn's comment concerning inching along though. If I really had to write something to parse foo { ... }; I'd probably use that.

What I find really intriguing however is the original question posed however vague. Given a Perl'ish structure of text where you have a symbol followed by a open curly brace followed by more text with possibly more curly braces (open and closed) how do you go about matching the right pair?

My first instinct would be to look towards some sort of recursive parser and not a regex. What are your thoughts?

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Peter L. Berghold --- [e-mail]
"Those who fail to learn from history are condemned to repeat it."