Help with regex please...

[cranjled]

I have to look over a string and remove anything that's not within the following ranges:

a-z
A-Z
0-9
_ (underscore)

So any punctuations, spaces, or generally non-filename-acceptable characters would be removed from the string. Underscores would be permitted, however.

I know it only takes a few minutes with this regex stuff once you get the hang of it...but until then...your genius is greatly appreciated!



Cranjled

[Mandes]

^[a-zA-Z0-9_]*$

[santanu]

if(eregi('^[a-zA-Z0-9_]+$',$string))
return true;
else
return false;

[stereofrog]

I have to look over a string and remove anything that's not within the following ranges:

a-z
A-Z
0-9
_ (underscore)

Try this

preg_replace("~\W+~", '', $a); 

[cranjled]

Thanks you guys! Sheesh...twenty characters can really get you bangin' your head... I really appreciate the responses.

Question: Is one of these methods preferred over another? Any why?

Thanks again!

[cranjled]

stereofrog, that was exactly the trick I needed. It keeps letters, numbers and underscores and strips out all the rest. Most excellent!

Thanks to all of you!

Cranjled

[cranjled]

Embarassingly, I can't figure out how to add the dot character to the allowed characters...and furthermore, I'd do a sad smily here, but I don't even know the right character combo for that...

Using the example that stereofrog posted, worked nicely for the purpose, but now I need to add the dot character:

Can someone amend the following so that it allows the dot character too? (Just that one character, though!)

 preg_replace("~\W+~", '', $a);

Any help is greatly appreciated!

Thanks,

Cranjled

[logic_earth]

preg_replace('~[^\w\.]+~', '', $a); 

[cranjled]

Boy, do I feel goofy ... !

Thanks so much!



Cranjled

[cranjled]

This regex works nicely for the intended purposes. With a bit of tweaking it could maybe be used in something else that I'm trying.

The following code removes many characters that are not needed:

 preg_replace("~\W+~", '', $a);

So, I decided to try it for image verification purposes. What I'm doing is reading an image file via file_get_contents(), and stripping out unneeded characters. The regex above does most of the work, but does leave the string with many unwanted characters (such as ÿ, Ø, à, ÿ, þ, etc...). I guess these /are/ technically word characters, but since they are unneeded, can the pattern be amended to remove any non-english-word characters?

Essentially, the goal is to read the file in and determine if it is a "real" gif, jpg or png image. I have performed extension and mimetype checks, but these are easily spoofed. To add an additional layer of assuredness, I would like to verify the filetypes by actually reading them into a string and determining from there. Can this also be spoofed? (And if so, what IS the best method to assure a valid image where only an image should be?)

Thanks so much for following this thread!

- Cranjled

[logic_earth]

http://us.php.net/manual/en/function.exif-imagetype.php

[cranjled]

Now, that's just what the doctor ordered! Thanks much!