Global variables

[iceman22]

Hi everyone.

I've been trying to do this for a while and have not been able to successfully do it or find any good info on Google. I have a script that submits a form, and all the variables that aren't $_POST variables will be cleared.

Is there any way to preserve an ordinary variable? The solution I have is pretty dirty but it works, I just save it as a text file and open it in the code executed after the form has been submitted, to avoid conficts using tmpfile() would work except it generates a different string before and after the form is submitted, so I use the md5 algorithm and multiply the user agent, ip and host. Clearly this isn't the way to code.

I've got another little problem, with regular expressions.

I want to parse URLs from within some text, I want to hyperlink regular [example.com...] and also http://www.example.com [example.com]. I have regular expressions that work for either, but they conflict when I use both of them.

I use this for a regular URL:
$text = preg_replace("~(http://\S+¦www\.\S+)~i", "<a href=\"\\1\">\\1</a>",$text);

And this for UBB URLs:
$text = preg_replace("/\[url=(.*?)\](.*?)\[\/url\]/i","<a href=\"\\1\">\\2</a>",$text);

I believe they could work in harmony if the top regex didn't parse URLs that have an equals sign preceding them, eg. =http://... because all [url] formatted URLs have this.

I found it surprisingly difficult to find a regex for the [url] formatted link that actually worked, I had to take a few in the end and build it with them.

Thanks.

[edited by: jatar_k at 7:02 pm (utc) on Jan. 4, 2005]
[edit reason] examplified [/edit]

[mcibor]

With the regex there's a simple answer:

You're replacing first [......] with <a
and the www... with <a, so you are replacing the same url twice (i don't think that's what you want).

Eg. $text = "http://www.bla.com is a nice page"

after first replace you get:
<a href="http://www.bla.com">http://www.bla.com</a> is a nice page

after second replace you get:
<a href="http://<a href="www.bla.com">www.bla.com</a>">http://<a href="www.bla.com">www.bla.com</a></a> is a nice page

That's completely horrendous!

To achieve goal leave replace [..,...] but the second you should replace urls that must be without [....]
I don't know how to do that, but it's possible.

With preserving variables it's a bit tricky. You can use session to preserve them, or database, or if the data is not very relevant place it into <input name="variables" type="hidden" value="<?echo $preserve_variables;?>" />

[Salsa]

Regarding the ordinary variables, as mcibor suggested, couldn't you put them in hidden form fields, or is the issue more complicated than that?

As for the regex problem, try adding an assertion, like:

$text = preg_replace("~(?<!=)(http:// ... ;

...to your regular url pattern. That should exclude cases where http is preceeded by an =

I hope this helps.

[iceman22]

Ok I found a solution to the regex problem. The old regex that parsed regular URLs wasn't any good, it also messed up text that didn't have the http:// in front of them. These two expressions work properly in this order:

$text = preg_replace('/(?<!=)(?<!])(http¦ftp)+(s)?:(\/\/)((\w¦\.)+)(\/)?(\S+)?/i','<a href="\0">\0</a>',$text );
$text = preg_replace('/\[url=(.*?)\](.*?)\[\/url\]/i','<a href="\\1">\\2</a>',$text);

The first expression stripped the "http://" from the visible output, but changing <a href="\0">\4</a> to <a href="\0">\0</a> fixed that.

I had success using the hidden fields in forms, but for private information, I've left using files. What about session variables?