Stopping html encoding in URLs

[andrea_r]

I'm tracking down the issue with why timthumb doesn't work on WPMU setups. I've narrowed it down to this:

The generated URLs to images get specs with & in the string, and WPMU converts them to & instead.

Now - how to stop it? :D Any ideas, or has anyone run into something similar?

[andrea_r]

Doh - http://ca3.php.net/urldecode

[Otto]

... I'm not exactly seeing the problem. URL's are supposed to be entity encoded.

Or are you decoding them later into filenames or something?

[andrea_r]

Ugh, turns out that was a red herring. It was ultimately another issue, one not worth fixing for MU.

Which I knew, I was trying to discover *why* it wasn't working, and how it could be fixed.

[BinaryMoon]

What was the problem?

I am working on a new beta for TimThumb and I would like to make sure I have as many bugs fixed as possible, especially with relation to WPMU as I know a lot of people have problems with that.

You can grab the latest beta here : http://www.binarymoon.co.uk/2009/07/timthumb-beta-test/

Feel free to email me via the contact page on my site or to DM me on Twitter or something (in case I forget to check back here :) )

[andrea_r]

Oh bless your heart.... I had been replacing it with get_the_image but some of the queries were constructed in such a way it only worked half the time.

I'm gonna try this, and also see if I can untangle it enough to explain what was going on.

[andrea_r]

Okay, same problems.

Error message:

//wp-content/blogs.dir/1subd./files/2009/07/740px-coldplay_-_december_2008.jpg not found.

On:

domain.com/wp-content/themes/arthemia-premium/scripts/timthumb2.php?src=http://subd.domain.com/files/2009/07/740px-coldplay_-_december_2008.jpg&w=80&h=80&zc=1&q=100&bid=1

The issue is that timthumb is not coming up with the correct absolute path to the file. In this instance, the image is in the uploaded files of a sub-blog while the thumbnail is being shown on the main blog.

If we add the subdomain to the beginning of the url and provide the correct blog id, it does find the file.

Because the way people are using this is with the sitewide tags plugin (or whatever) to pull all member posts to the front page.

[BinaryMoon]

using the latest beta is "should" work with subdomains. Is there any way I can look at the site for real - ideally with ftp access (a big ask, i know)?