Serious Regular Expression problems!

redrabbit

AHHHH!

Aim
To convert HTML into XHTML code, e.g. make <br> <br />

Code
$text = preg_replace('/(<img .*)("|\'| )>/i','\1\2 />',$text);

The problem
It removes the / from the source. E.g.

<img src="../myimages/blah.gif"> turns that to:
<img src="..myimages/blah.gif" />

So it takes the first slash away meaning the image does not work.

Any suggestions?

Weedpacket

Search pattern:

#(<img.*?)(?<!/)>#is

Replacement text:

$1 />

The (?<!) negative lookbehind assertion is to stop a / from being added if there's one already there (don't want two, after all). The s modifier is so that . will match newlines, and the i is because tag names are case-insensitive in HTML (though of course you'll need to lowercase both them and attribute names to produce valid XHTML).

For a more robust HTML->XHTML conversion, the [man]tidy[/man] extension could be considered.

redrabbit

I have now found the problem but also have this problem with it:

http://www.phpbuilder.com/board/showthread.php?t=10323401

It seem to be replacing the tag of something else as well, as if it is not stopping checking but going through the whole text replacing ALL tags. Have I missed a termination or something?

Weedpacket

Well I do see one bug in what I wrote, and that is if the tag already does end with /> it will go haring off until it finds one that doesn't and puts the / there. Maybe

#(<img.*?)/?>#is

A lazy solution would be to turn /> into > and then put the /> back....

But then, I've used tidy for years 🙂

redrabbit

I was unsure how to install/utilise it.

config.m4
CREDITS
examples/
php_tidy.h
README
tests/
tidy.c
TODO

That's the files that tidy produced. Not really sure what to do with them. Another thing to remember is that they will be going on several servers, some of which I do not have root access to.