Stripping Control Characters from PHP Strings

Q: Modifying Regular Expression for Control Character Removal

A PHP developer previously employed the following expression to purge control characters like STX from a string:

preg_replace("/[^a-zA-Z0-9 .\-_;!:?äÄöÖüÜß<>='\"]/","",$pString)

However, the result was overly restrictive, prompting the question: how can control characters be selectively removed?

A: Utilizing Specific Character Classes for Control Character Identification

To precisely target control characters, a more specific character class can be utilized:

preg_replace('/[\x00-\x1F\x7F]/', '', $input);

This expression matches characters with ASCII codes within the range \x00-\x1F and \x7F, encompassing the first 32 characters and \x7F, which includes characters like carriage returns.

Preserving Essential Characters like Line Breaks

If specific characters, such as line feeds or carriage returns, need to be preserved, their escape sequences can be excluded:

preg_replace('/[\x00-\x09\x0B\x0C\x0E-\x1F\x7F]/', '', $input);

In this modified expression, \x0A (line feed) and \x0D (carriage return) are excluded from the character class.

Modern Syntax and Deprecation

It is crucial to note that the deprecation of ereg_replace in PHP 5.3.0 and its subsequent removal in PHP 7.0.0 mandate the use of preg_replace over ereg_replace.

Finally, a Character Class for Control Characters

For a concise and portable alternative, the character class [:cntrl:] can be employed:

preg_replace('/[[:cntrl:]]/', '', $input);