Can you help me rewrite my javascript regex using php preg_replace?

Question

I have created a javascript regular expression in order to validate comments entered by users in my app. The regex allows letters, numbers some special symbols and a range of emojis

I received help here to correctly format my javascript regular expression and the final expression I am using is as follows:

Javascript Regex:

commentRegex =    /^(?:[A-Za-z0-9\u00C0-\u017F\u20AC\u2122\u2150\u00A9 \/.,\-_$!\'&*()="?#+%:;\<\[\]\r\r\n]|(?:\ud83c[\udf00-\udfff])|(?:\ud83d[\udc00-\ude4f\ude80-\udeff]))*$/;

I was advised to perform the same validation on the server side (with php) and so I am trying to perform a similar process using preg_replace().

So I would like to replace all characters (that are not contained in the regex), with the empty string. Here is my attempt however it is not working. thanks for any help

PHP

$commentText = preg_replace('#^(?:[A-Za-z0-9\u00C0-\u017F\u20AC\u2122\u2150\u00A9 \/.,\-_$!\'&*()="?#+%:;\<\[\]\r\r\n]|(?:\ud83c[\udf00-\udfff])|(?:\ud83d[\udc00-\ude4f\ude80-\udeff]))*$#', '', $commentText);

Edit:

After taking your advice in the comments I now have the following regex.

$postText = preg_replace('/^(?:[A-Za-z0-9\x{00C0}-\x{017F}\x{20AC}\x{2122}\x{2150}\x{00A9} \/.,\-_$!\'&*()="?\#\+%:;\<\[\]\r\n]|(?:\x{d83c}[\x{df00}-\x{dfff}])|(?:\x{d83d}[\x{dc00}-\x{de4f}\x{de80}-\x{deff}]))*$/', '', $postText);

However I am getting a warning

<b>Warning</b>:  preg_replace(): Compilation failed: character value in \x{} or \o{} is too large at offset 30 in <b>submit_post.php</b> on line <b>37

It would help us if you would add $commentText on what you specifically want replaced. — Xorifelse
– Xorifelse, Commented Jan 15, 2017 at 21:53
In PHP PCRE, you need to turn all \uXXXX to \x{XXXX}. And also, there is no need to write \r twice in [\r\r\n]. BTW, it is not working is a poor problem description, you should always provide the exact behavior you get. — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Jan 15, 2017 at 21:57
@Xorifelse I was hoping I could go with the approach where I have a whitelist of "allowed characters and then everything that is not in the whitelist gets replaced with the empty string. Is this possible? I can tell you the allowed characters then... — Sarah
– Sarah, Commented Jan 15, 2017 at 21:59
This is becoming interesting. Try $re = '/^(?:[A-Za-z0-9\x{00C0}-\x{017F}\x{20AC}\x{2122}\x{2150}\x{00A9} \/.,\-_$!\'&*()="?#+%:;<[\]\r\n]|[\x{1F300}-\x{1F3FF}]|[\x{1F400}-\x{1F64F}\x{1F680}-\x{1F6FF}])*$/u';. — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Jan 15, 2017 at 22:48

Graham · Accepted Answer · 2017-09-24 19:28:18Z

1

In short: use

$re = '/[^A-Za-z0-9\x{00C0}-\x{017F}\x{20AC}\x{2122}\x{2150}\x{00A9} \/.,\-_$!\'&*()="?#+%:;<[\]\r\n\x{1F300}-\x{1F3FF}\x{1F400}-\x{1F64F}\x{1F680}-\x{1F6FF}]+/u';
$text = 'test>><<<®¥§';
echo preg_replace($re, '', $text);

See the PHP demo.

A bit of an explanation:

Escape only special regex metacharacters inside the pattern AND the regex delimiters (if you choose a # as a regex delimiter, escape the # in the pattern, and then there is no need to escape /)
\uXXXX in PCRE must be replaced with \x{XXXX} notation
Since the text to be processed is Unicode and the chars you have in your pattern are out of the ASCII range, you have to use /u UNICODE modifier
As most emojis come outside the BMP plane, and the string now treated as a chain of Unicode code points, these symbols must be written using the extended \x notation, not as two byte notation used in JavaScript
Your 3 alternatives can be merged into 1 big character class and then you want to negated it by adding ^ at its start to make it a negated character class.

edited Sep 24, 2017 at 19:28

Graham

7,86020 gold badges67 silver badges92 bronze badges

answered Jan 16, 2017 at 20:58

Wiktor Stribiżew

631k41 gold badges502 silver badges632 bronze badges

Sign up to request clarification or add additional context in comments.

6 Comments

Sarah Over a year ago

Stribizew Hiya. I have a question related to this and thought i'd ask here before posting another question. This regex works perfect. When I type a message, it allows for example double quotes... However I have discovered that If I copy and paste text (containing double or single quotes) from another webpage into the textarea in my form and submit it, the quotes are replaced with empty string. I suspect its to do with the font that i copied and pasted in or the encoding or something else. Do you know anymore than me about this behaviour? thanks

Wiktor Stribiżew Over a year ago

That just means the quotes are matched with the regex and are removed with preg_replace.

Sarah Over a year ago

but remember our solution replaces everything BUT whats in the regex as I wanted. and it does allow quotes when I type them in but just not when I copy and paste. It's ok-I will read more about it.

Sarah Over a year ago

thanks.. if you see here I added in some characters to your text. (I copied and pasted the quotes from a web page). ideone.com/trSge1

Wiktor Stribiżew Over a year ago

The curly single quotes are \x{2019}\x{2018}. Add them to the negated character class to avoid removing them - ideone.com/oNIk9p.

|

Fabian N. · Accepted Answer · 2017-01-15 22:00:09Z

1

The regex in PHP has a character, which sourrounds the regex. In your case you are using the hash (#), but the character should not occour in the regex itslef, which it does...

You have to excape this character inside, or use another char. Why did you not use the same "/" as in the JS Version? The benefit is, it is already escaped.

I have not looked, if the rest would work, but I think so.

$commentText = preg_replace('/^(?:[A-Za-z0-9\u00C0-\u017F\u20AC\u2122\u2150\u00A9 \/.,\-_$!\'&*()="?#+%:;\<\[\]\r\r\n]|(?:\ud83c[\udf00-\udfff])|(?:\ud83d[\udc00-\ude4f\ude80-\udeff]))*$/', '', $commentText);

should work.

answered Jan 15, 2017 at 22:00

Fabian N.

1,28011 silver badges19 bronze badges

Comments

Scott Weaver · Accepted Answer · 2017-01-15 22:10:49Z

1

convert the \u.... sequences to \x{....}, and the result appears to be a valid PHP regular expression.

pattern: \\u(\w{4})

replace: \\x{$1}

regex101 demo

edited Jan 15, 2017 at 22:10

answered Jan 15, 2017 at 22:03

Scott Weaver

7,3832 gold badges33 silver badges45 bronze badges

2 Comments

Sarah Over a year ago

Thanks I have changed like you said but I am getting a warning now: Warning: preg_replace(): Compilation failed: character value in \x{} or \o{} is too large at offset 30 in submit_post.php on line 37

Scott Weaver Over a year ago

add utf-8 parsing with the u flag? preg_match('/pattern/u', input) check this answer stackoverflow.com/a/32375905/244811

Collectives™ on Stack Overflow

Can you help me rewrite my javascript regex using php preg_replace?

Javascript Regex:

PHP

Edit:

3 Answers 3

6 Comments

Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

Javascript Regex:

PHP

Edit:

3 Answers 3

6 Comments

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related