Posted by Harlan Messinger on 12/20/07 16:32

David Segall wrote:
> I want to encode a string that will be used as a GET data parameter
> but the algorithm I have can produce the characters "/", "+" and "="
> in addition to alphanumeric characters. Those characters don't have a
> named entity in my HTML text book so I believe they can be used
> without further encoding. Can they? In other words, is the URL
> <http://example.com?myparam=four/two+3=5> valid? For extra credit :),
> where should I have looked to find the definitive answer to this
> question?

URLs are not HTML. They have their own syntax. In particular, the plus
and equals signs have special meanings in a query string: the plus sign
is interpreted as a space character and the equals sign is used to
create a key/value association as you did in your own example,
associating the key "myparam" with the value "four/two+5=5".

The characters that have special meaning in a query string or that
delimit the query string from other parts of the URL are the ones in the
set {=?&;#%+}. When you use want to use any of these as an ordinary
character, encode it as %nn where nn is the hexadecimal ASCII code for
the character. An embedded space can be encoded as either %20 or a plus
sign.

See http://en.wikipedia.org/wiki/Percent-encoding.

[Back to original message]