ansaurus

Question

Answer 1

A:

Put brackets () around the bits of regex that you want to be stored in $errors.
You'll also want to use preg_match_all() rather then preg_match().

chigley 2010-10-08 16:19:27

That doesn't work for me

Kevin 2010-10-08 16:22:06

Answer 2

A:

Remember to escape your \ in strings.

preg_match_all('#<b>(.+?)</b>:(.+?) in <b>(.+?)</b> on line <b>([0-9]+)</b><br(?: /)?>#is',$string,$errors);

This code on ideone

Colin Hebert 2010-10-08 16:36:39

http://ideone.com/utL3K WORKS!

Kevin 2010-10-08 16:55:52

@Kevin, even so, you should read @Gumbo answer.

Colin Hebert 2010-10-08 16:57:23

Answer 3

A:

If this is your own website you can either: set the log levels and parse your log files (easier) or run your scripts from the command line with php -l.

ikanobori 2010-10-08 16:36:40

The problem is, it is the site of a client, so I can't use that method.

Kevin 2010-10-08 16:41:07

Answer 4

+2 A:

Forgive my language but it's quite foolish to attempt to parse HTML with regular expressions, especially potentially-malformed HTML. Use an HTML parsing library instead.

For HTML parsing and validation in HTML, I would refer to this answer; also check out the tidy extension.

Ether 2010-10-08 16:38:47

Well, in this case the HTML isn't really XML compliant, and moreover you can't really know where this error will show up so an XML parser (or HTML for what it worth) won't help.

Colin Hebert 2010-10-08 16:39:59

@Colin: there are HTML parsers that will identify errors, which is precisely what the OP wants to do. HTML is not regular, so using a regular expression will not be fruitful.

Ether 2010-10-08 16:44:35

That comment must be one of the most-linked ones here.

CanSpice 2010-10-08 16:55:36

You're prob. right, but how do I do this in this situation?

Kevin 2010-10-08 16:59:09

@Kevin: I've edited my answer with the best links I could find.

Ether 2010-10-08 17:16:34

Answer 5

+5 A:

Since – well, you know – you shouldn’t use regular expressions to parse HTML, try this using PHP’s DOM library:

libxml_use_internal_errors(true);
$doc = new DOMDocument();
$doc->loadHTML($str);
$messages = array();
foreach ($doc->getElementsByTagName('b') as $elem) {
    if (in_array($elem->textContent, array('Error', 'Warning', 'Notice'))) {
        $buffer = $elem->textContent;
        while ($elem->nextSibling !== null && strtolower($elem->nextSibling->localName) !== 'br') {
            $elem = $elem->nextSibling;
            $buffer .= $elem->textContent;
        }
        $messages[] = $buffer;
    }
}

This will search for B elements that’s content is one of “Error”, “Warning”, or “Notice” and take the textual contents from there up to the next BR element. The initial call of libxml_use_internal_errors will prevent that parsing errors will be reported.

Gumbo 2010-10-08 16:49:12

This works not entirely, how can I let this work the same as http://ideone.com/utL3K?

Kevin 2010-10-08 16:58:22

@Kevin: Ok, I have to admit that this might fail if the document is actually invalid HTML and is fragmented in such a way that parsing fails.

Gumbo 2010-10-08 17:10:52

No, it just does not list the errors correct. The while doesn't work. If I just delete the while it will list the errors... But not the texts

Kevin 2010-10-08 17:12:56

@Kevin: I fixed that bug.

Gumbo 2010-10-08 17:14:05

Thanks for the help!

Kevin 2010-10-08 21:15:46

ansaurus

tags:

views:

answers:

PHP regex that finds PHP errors

related questions