Problem with preg_replace!

crewcialdotorg

I am trying to take a url...

and take away everything after the first attribute... so it would become <a href="linkhere">

but here's the twist.. i want to do it if the input is double quoted, single quoted, alternate single/double or no quotes or just one quote.. because in the replacement i just use a callback for it.

here is my match strong: $replace = "/<(\w+) (\w+)=(\"|'|)(.*?)(\"|'|)\s|\S>/msi";

i've tweaked it a bunch and this is the latest i can come up with.. here's what i'm doing with the replace function.

$new_string = htmlentities(preg_replace($replace, "<\1 \2=\"\3\">", $_GET[shit]));

I'm adding double quotes for the user.

Thanks!

xblue

I think your problem originates here: [\s|\S] means whitespace character or non-whitespace character, in other word: anything.

let's break it down and see what is matched:

<a href="linkhere" name="link 1"> 

<(\w+) (\w+)=
matches '<a href=', this is fine.

(\"|'|) which could be rewritten as [\"']?
matches the quotation mark, fine as well.

(.*?) matches nothing, since it consumes as little as possible.

(\"|'|) matches nothing, since at this position there is no quotation mark, so it uses the last branch.

[\s|\S] matches the 'l' in 'linkhere'

(.*?)> matches all the rest until the closing bracket.
:(:(:(

so, let's start over:

<(\w+) (\w+)= seems okay.

then we have that attribute value, which may be not quoted (then it mustn't contain spaces) or in double or single quotes (where spaces would be legal) - I'd actually recommend to treat these cases separately in an alternation, otherwise it's difficult to deal with spaces properly (since a space will probably be what you are going to identify the next part, see below).

if that's all, we don't need to match the whole tag. otherwise, a space, followed by anything except a closing bracket, would match the other attributes that need to be removed.

I hope that gets you on the track? putting the replacement argument with the backreferences together may be a little tricky, but it's possible.

crewcialdotorg

That's pretty helpful.

Basically.. the replacement i'm creating looks like <\1 \2=\"\4\">..

so i'm creating the output to be proper html(or xhtml, in my case)

the only problem i'm having is handling the user not quoting the first argument, because i don't care about the rest of them.

if you want to see what i'm doing this from.. http://crewcial.org/ereg

just put in a img,href,or font tag.. actually, any take with an attribute. i'm going to change it later to fit certain tags as i want to handle img tags differently than anchor tags.

thanks for your help so far.

crewcialdotorg

top!

xblue

sorry to say that, but your current pattern

/<(\w+) (\w+)="'"'>/msi

fails in some situation:

before replace: <a name="Let's go">
after replace: <a name="Let">

again, I'd recommend to treat the cases separately. you should be able to put up the patterns for doublequoted strings, singlequoted strings and not quoted strings I think, and put them into an alternation?

if you need help with the replacement argument, just post what you've got so far.

btw, drop the i and m modifier, you don't need them.