正则表达式 - 逃避负面前瞻

Question

I have the following JSON-encoded string:

$json = '"|\t|\n|\\\u0027|\\\u0022|"';

What is the most efficient way to escape all the (already) escaped chars / codepoints except \\\\" or \\\\\' ? I though about using preg_replace() with a negative lookahead regular expression but it's not working as I expected, the output should be:

$json = '"|\\\t|\\\n|\\\u0027|\\\u0022|"';

I'm feeling lost in this ocean of JSON-PHP-PCRE escaping, can someone help me out?

Answer 1

Something like this may work with the help of negative lookahead:

<?php
  $json = '"|\t|\n|\\\u0027|\\\u0022|"';
  $s = preg_replace('~(\\\\)(?!(\\1|u002[27]))~', '$1$1$1', $json);
  var_dump($json);
  var_dump($s);
?>

OUTPUT

string(25) ""|\t|\n|\\u0027|\\u0022|""
string(29) ""|\\\t|\\\n|\\u0027|\\u0022|""

Answer 2

I'm a bit confused by exactly what you are trying to do but I can transform your input to your output with this:

preg_replace('/\|\\([^\\])\|/', '\\\\\\$1|', $json);

Note: I'm not at my computer so I can't verify that this is perfect but it looks good from here.

Answer 3

Try

$result = preg_replace('/(?<!\\\\)\\\\(?!\\\\)/', '\\\\\\\\\', $subject);

This finds a \\ only if it is single (ie neither preceded nor followed by another \\ ) and replaces it with \\\\\\ .