将正则表达式从ruby转换为PHP

Question

I have been trying to convert a regular expression from ruby to PHP, however I have had little luck.

This is the ruby regular expression:

QUOTED_LITERAL = /"[^"\\]*(?:\\.[^"\\]*)*"/
UNQUOTED_LITERAL = /[^\s=,][^\s=,\\]*(?:\\.[^\s=,\\]*|=[^,>])*/
LITERAL = /(#{QUOTED_LITERAL}|#{UNQUOTED_LITERAL})/
PAIR = /#{LITERAL}\s*=>\s*#{LITERAL}/

And this is my go in PHP:

 const PAIR = '/("[^"\\]*(?:\\.[^"\\]*)*"|[^\s=,][^\s=,\\]*(?:\\.[^\s=,\\]*|=[^,>])*)\s*=>\s*("[^"\\]*(?:\\.[^"\\]*)*"|[^\s=,][^\s=,\\]*(?:\\.[^\s=,\\]*|=[^,>])*)/';

However when I run

$result = preg_match_all(self::PAIR, $input, $matches);

I get the error:

preg_match_all(): Compilation failed: unmatched parentheses at offset 62

---

However when run it through, http://www.phpliveregex.com/ with the test data:

"foo" => "bar", "foo" => bar, foo => "bar"

it seems to work fine.

Not sure whats going on.

Answer 1

The problem lies with your backslashes.

I managed to get it to compile after removing all backslashes.

Then, I replaced all double slashes with 4 of them, and preg_match_all() was able to compile the regex too.

const PAIR = '/("[^"\\\\]*(?:\\\\.[^"\\\\]*)*"|[^\s=,][^\s=,\\\\]*(?:\\\\.[^\s=,\\\\]*|=[^,>])*)\s*=>\s*("[^"\\\\]*(?:\\\\.[^"\\\\]*)*"|[^\s=,][^\s=,\\\\]*(?:\\\\.[^\s=,\\\\]*|=[^,>])*)/';

You might have to edit it to get the exact regex you want. You had the compilation error because \\\\ was fed to the regex engine as \\ , which escaped the immediate square brackets. To encode a literal backslash, you need to use \\\\\\\\ - once for the string, and once for the regex engine.

string '\\\\'  --becomes--> regex \\ --becomes--> literal \

Ruby doesn't have this problem because its regex syntax is separate from its string syntax.

(Related question: preg_match(): Compilation failed: unmatched parentheses .)