backreference regex help

[bobbinsbro]

hi all

how would i go about using a backreference in a character class in a pcre regex pattern?

 

example of something similar to what i want to do:

$pattern = '/(abc)([^\1]).*/';

 

thanks in advance

bob

[cags]

Erm... like that. I don't understand the question. Given an example input with expected output.

[bobbinsbro]

the problem is that doesn't work.

i'm trying to get the contents of a string delimited by either double quotes ("...") or single quotes ('...'). the regex i'm using is:

$pattern = '/(\'|")([^\1]*)(?:\1)/s'

that doesn't work. however, changing it to

$pattern = '/(\'|")([^\']*)(?:\1)/s'

does work (assuming the string is delimited by single quotes).

 

i checked and made sure that \1 contains a single or double quote (used preg_match() with the optional &$matches argument and print_r()'ed $matches to see what is in the sub-patterns).

 

as a result, i assume either i can't use a backreference in a character class or i'm doing something wrong.

[gizmola]

Taking you at your word, if I was going to construct this regex I would use something like:

 

$pattern = '/(\'|")(.*?)(\1)/s';

 

Not sure why you are trying to match things that aren't the backreference when you will already use the backreference to delimit your match.

 

 

[bobbinsbro]

the pattern i gave is part of a much longer pattern, that is supposed to match long blocks of text containing many quote-delimited strings. i originally tried what you suggested, but preg_match() didn't stop at the first quote (the one on which i need to end the sub-pattern) but rather continued to the very last quote, counting all intermediate quotes as part of the (.*) sub-pattern.

so i tried using matching any character that isn't the backreference to force preg_match() to stop on the first quote found. i need to use the back reference because i need to allow both single and double quotes in all places, and i can't know which in advance.

 

example of a string i want to match:

'the brown fox' 'jumped over the lazy dog'

and the pattern

$pattern = '/(\'|")([^\1])(\1)/s';

should match only

the brown fox

but doesn't match anything.

using

$pattern = '/(\'|")(.*)(\1)/s';

matches

the brown fox' 'jumped over the lazy dog

which is not what i want

[gizmola]

You didn't look at my pattern carefully. 

 

What you're talking about is the default of greedy.  To make the match non-greedy I added the '?' to the match all characters  ->  '.*?'

[bobbinsbro]

thank you very much, that did the trick. i didn't know about greedy and non-greedy. nice to learn new things.