Matching place holders starting/ending with {{ }}

bricktop

Hi All,

I have been haing problems getting a regular expression working to find all matches of the following pattern burried inside XHTML markup inside a variable.

{{param1:123_xyz^{param2:xyx^param3:123}}}

I need to match all place holders like the above, including the curly brackets at the start and end. The params inside can and will vary each and every time and contain mainly alphanumerical characters but also carrats/underscores/colons etc.

Could anyone possibly give me some guidance please?

Kind regards,
Brick...

NogDog

preg_match_all('/\{\{.*\}\}/Us', $text, $matches);

nrg_alpha

Here's my take:

$str = 'this is garble {{param1:123_xyz^param2:xyx^param3:123}} and so is this!
This is secondary garble {{param1:345_adr^param1:xyx^param6:788}} and this is soo much more garble!';
preg_match_all('#\{{2}[^}]+\}{2}#s', $str, $matches);
echo '<pre>';
print_r($matches[0]);
echo '</pre>';

Output:

Array
(
    [0] => {{param1:123_xyz^param2:xyx^param3:123}}
    [1] => {{param1:345_adr^param1:xyx^param6:788}}
)

Utilising negated character classes avoids the need to have the regex engine do any backtracking.

NogDog

Could be a problem though if "}" is allowed within the place-holder, e.g.:

Some text {{param:123^param:example {text} in brackets^param:234}} for example

All depends on the functional requirements and how flexible/portable you want this code to be. But if it is a given that "}" cannot be part of the data within the double-bracketed patterns, then your solution would, indeed, probably be better.

nrg_alpha

True enough. I was going by what the OP mentioned:

The params inside can and will vary each and every time and contain mainly alphanumerical characters but also carrats/underscores/colons etc.

So in this case, I am assuming curly brace characters would not be found inside the surrounding brace set. If there are to be any, then definately, my solution would not work.