preg_replace and POSIX

s0me0ne

Can you mix preg_replace and POSIX or do you have to use PCRE?

I am trying to clean up all non-printable items in 1 function I have

//get rid of non-printables
$string = preg_replace("/[^[[[:print]][[:space:]]]]/", '', $string);

Later on another function (that is optional) MIGHT have to clean it up more and if so it tries to do this

//done calling trim and strip_tags, now more cleaning

//stuff thats okay for email
$message = preg_replace("/[^a-zA-Z0-9_,#'%&\"\-\.\?\!\~\=\/\$\\n\r\s]/", '', $message);

//stuff not safe for webpage
$message = preg_replace("/[`;\*\^\|\\\<\>\f\t\v]/", '', $message);

I'm so confused, been trying to figure this out 😕

NogDog

If the goal is to filter text for output in HTML (either a web page or HTML email), I would probably just do:

$message = htmlentities(strip_tags($message));

s0me0ne

I've read that isnt safe enough, I was going to filter out more than just that 😕

NogDog

If this isn't "safe", I don't know what is:

$string = htmlentities(strip_tags(preg_replace('/[\x00-\x20\x7f-\x9f]/', '', $string)));

reddrum

NogDog wrote:
If this isn't "safe", I don't know what is:
$string = htmlentities(strip_tags(preg_replace('/[\x00-\x20\x7f-\x9f]/', '', $string)));

Hi NogDog,
If not too much trouble, I would appreciate a short explanation of what this expression represents: '/[\x00-\x20\x7f-\x9f]/'.

Thanks!

s0me0ne

its a regex that will filter out any unsafe characters, basically it only permits A-Z 0-9 and some other characters, compare it to an ascii chart

regex is quite complicated, thus i always have problems with it, i didnt even think of using the ascii table like this

reddrum

Hi sOmeOne,

So are you using it like this?:

$string = htmlentities(strip_tags(preg_replace('/[\x00-\x20\x7f-\x9f]/', '', $string)));
$string = preg_replace("/[`;\*\^\|\\\<\>\f\t\v]/", '', $string);

NogDog's string takes out the tags and spaces so you still need the last preg_replace(), is that correct?

NogDog

On further review, I would make a couple minor changes, so that any number of consecutive non-printing characters are converted to a single space (in case any of them such a NL or CR are needed to maintain word spacing):

$string = htmlentities(strip_tags(preg_replace('/[\x00-\x19\x7f-\x9f]+/', ' ', $string)));

s0me0ne

I'm using the following (thanks to nogdog)

$string = preg_replace('/[\x00-\x1f\x7f-\x9f]/', '', $string);//keep printables

$string = preg_replace('/[\x00-\x08\x0b\x0c\x0e-\x1f\x7f-\x9f]/', '', $string);//keep printables, new lines and carriage returns

Depending on if I need newlines or not, depends which one my function uses 🙂

reddrum

I get it now, this is tons better than what i was using, Awesome!
Thanks guys!!

deshetty

hmmmm