A html space is showing as %2520 instead of %20

Question

Passing a filename to the firefox browser causes it to replace spaces with %2520 instead of %20.

I have the following HTML in a file called myhtml.html:

<img src="C:\Documents and Settings\screenshots\Image01.png"/>

When I load myhtml.html into firefox, the image shows up as a broken image. So I right click the link to view the picture and it shows this modified URL:

file:///c:/Documents%2520and%2520Settings/screenshots/Image01.png
                    ^
                    ^-----Firefox changed my space to %2520.

What the heck? It converted my space into a %2520. Shouldn't it be converting it to a %20?

How do I change this HTML file so that the browser can find my image? What's going on here?

Answer 1

A bit of explaining as to what that %2520 is :

The common space character is encoded as %20 as you noted yourself. The % character is encoded as %25.

The way you get %2520 is when your url already has a %20 in it, and gets urlencoded again, which transforms the %20 to %2520.

Are you (or any framework you might be using) double encoding characters?

Edit: Expanding a bit on this, especially for LOCAL links. Assuming you want to link to the resource C:\my path\my file.html:

• if you provide a local file path only, the browser is expected to encode and protect all characters given (in the above, you should give it with spaces as shown, since % is a valid filename character and as such it will be encoded) when converting to a proper URL (see next point).
• if you provide a URL with the file:// protocol, you are basically stating that you have taken all precautions and encoded what needs encoding, the rest should be treated as special characters. In the above example, you should thus provide file:///c:/my%20path/my%20file.html. Aside from fixing slashes, clients should not encode characters here.

NOTES:

• Slash direction - forward slashes / are used in URLs, reverse slashes \ in Windows paths, but most clients will work with both by converting them to the proper forward slash.
• In addition, there are 3 slashes after the protocol name, since you are silently referring to the current machine instead of a remote host ( the full unabbreviated path would be file://localhost/c:/my%20path/my%file.html ), but again most clients will work without the host part (ie two slashes only) by assuming you mean the local machine and adding the third slash.