regex question - can't get unclosed quote marks inside of HTML tags

[lawrence k]

I do not know much about regex.

I'm worried about lines like this:

<a href="myFile>my file</a>

There is only one quote mark in that html.

I wanted to fix this problem, so I tried this:
function command($string=false) {
$pattern = '/(.*)<a (.*)"(.*)>/i';
$replacement = '$1<$2"$3">';
$newString = preg_replace($pattern, $replacement, $string);
return $newString;
}
This finds no matches. Even when I feed it the above as a test line.

What have I done wrong.

[dennis.alund]

This isn't really a solution to your problem... just a hint of what's
wrong; a quote is also a character... i.e. in the expression (.*) will
also match a quote.
What you want to look into is negative lookarounds; (?<!x)y matches an
'y' not preceeded by a 'x' and x(?!y) matches a 'x' not followed by an
'y'.
But if you're not a regexp-ninja I'd recommend you to find an easier
solution... negative lookarounds aint trivial.

[lawrence k]

de**********@gmail.com wrote:

This isn't really a solution to your problem... just a hint of what's
wrong; a quote is also a character... i.e. in the expression (.*) will
also match a quote.
What you want to look into is negative lookarounds; (?<!x)y matches an
'y' not preceeded by a 'x' and x(?!y) matches a 'x' not followed by an
'y'.
But if you're not a regexp-ninja I'd recommend you to find an easier
solution... negative lookarounds aint trivial.

I see. So if I have a string like this:

<a href="myfile">myfile</a>

and I feed it to this function:

function command($string=false) {
$pattern = '/(.*)<a (.*)"(.*)>/i';
$replacement = '$1<$2"$3">';
$newString = preg_replace($pattern, $replacement,
$string);
return $newString;
}

I will get a match? But does that mean I'll end up with this:

<a href="myfile"">myfile</a>

With an extra quote mark? That is not the problem I was having.

But you are saying that I need to replace the final (.*) with something
that says "everything but a quote mark"?

[lawrence k]

de**********@gmail.com wrote:

This isn't really a solution to your problem... just a hint of what's
wrong; a quote is also a character... i.e. in the expression (.*) will
also match a quote.
What you want to look into is negative lookarounds; (?<!x)y matches an
'y' not preceeded by a 'x' and x(?!y) matches a 'x' not followed by an
'y'.
But if you're not a regexp-ninja I'd recommend you to find an easier
solution... negative lookarounds aint trivial.

So I want this?

function command($string=false) {

$pattern = '/(.*)<a (.*)"[^"]+>/i';
$replacement = '$1<a $2"$3">';
$newString = preg_replace($pattern, $replacement, $string);
return $newString;
return $string;
}

[lawrence k]

de**********@gmail.com wrote:

This isn't really a solution to your problem... just a hint of what's
wrong; a quote is also a character... i.e. in the expression (.*) will
also match a quote.
What you want to look into is negative lookarounds; (?<!x)y matches an
'y' not preceeded by a 'x' and x(?!y) matches a 'x' not followed by an
'y'.
But if you're not a regexp-ninja I'd recommend you to find an easier
solution... negative lookarounds aint trivial.

Thanks for the reply. I just created this file and I uploaded it to my
server for testing:

<?php

echo "hey";
flush();

function command($string=false) {
echo "hey";
$pattern = '/(.*)<a (.*)"[^"]+>/i';
$replacement = '$1<a $2"$3">';
// $newString = preg_replace($pattern, $replacement, $string);
return $newString;
}

$string = "<p><a href=\"myfile>myfile</a> ";
echo $string;
flush();
$string = command($string);
echo $string;

?>
When I don't comment out the preg_replace line then I get one "hey"
echoed to the screen and then the script apparently dies without error.

[lawrence k]

de**********@gmail.com wrote:

This isn't really a solution to your problem... just a hint of what's
wrong; a quote is also a character... i.e. in the expression (.*) will
also match a quote.
What you want to look into is negative lookarounds; (?<!x)y matches an
'y' not preceeded by a 'x' and x(?!y) matches a 'x' not followed by an
'y'.
But if you're not a regexp-ninja I'd recommend you to find an easier
solution... negative lookarounds aint trivial.

Well, okay, for anyone interested, I worked it out and found the
correct pattern is this:

<a ([^">]+)"([^">]+)>