urlencode vs rawurlencode

[Joshua Beall]

Hi All,

I can see from the manual that the difference between urlencode and
rawurlencode is that urlencode translates spaces to '+' characters, whereas
rawurlencode translates it into it's hex code.

My question is, is there any real world difference between these two
functions? Or perhaps another way of asking the question: *why* are there
two different functions? In what situation would you need one, and not be
able to use the other?

Thanks!

-Josh

[John Dunlop]

Joshua Beall wrote:

I can see from the manual that the difference between urlencode and
rawurlencode is that urlencode translates spaces to '+' characters, whereas
rawurlencode translates it into it's hex code.

My question is, is there any real world difference between these two
functions?
I don't know.
Or perhaps another way of asking the question: *why* are there two
different functions?
A good question. I don't know the answer to that either.

A plus sign is reserved in the query component. A reserved character
may be used for its reserved purpose or, if it doesn't conflict with
the reserved purpose, as data.

Spaces encoded as plus signs is specific to form encoding. The
HTML4.01 specification describes the encoding process: "[i]f the
method is 'get' and the action is an HTTP URI, the user agent takes
the value of action, appends a `?' to it, then appends the form data
set, encoded using the 'application/x-www-form-urlencoded' content
type" (HTML4.01, sec. 17.13.3). So, here, spaces are encoded as plus
signs; elsewhere, spaces are encoded as "%20", as explained in
RFC2396, section 2.4.

Consider:

1. <http://domain.example/?baz=foo+bar>
2. <http://domain.example/?baz=foo%20bar>
3. <http://domain.example/?baz=foo%2Bbar>

All three are syntactically valid URIs. The first could be a URI
generated from an HTML form, where the action specified was
<http://domain.example/>, the method GET and the form data set
consisting of a control named "baz" with current value "foo bar". The
space in the current value is replaced with a plus sign.

Reading Björn Höhrmann's explanation of reserved characters in

"Re: Good/Bad - URI encoding in HTML editor",
http://lists.w3.org/Archives/Public/...2May/0032.html

we see that numbers one and two are *not* equivalent.

Also related is Terje Bless' request for clarification

"Ambiguity of Allowed/Recommended URI Syntax and Escaping",
http://lists.w3.org/Archives/Public/...2Nov/0014.html
In what situation would you need one, and not be able to use the other?

That depends on the URI generator, I think.

The documentation for urlencode says "[t]his function is convenient
when encoding a string to be used in a query part of a URL" [1]. I
don't see any reason to favour it over rawurlencode, however, which
encodes as per section 2.4 of RFC2396 (modulo the fact it always
encodes certain unreserved characters [2]).

Refs.:

"Uniform Resource Identifiers (URI): Generic Syntax", 1998,
http://www.ietf.org/rfc/rfc2396.txt

"Uniform Resource Locators (URL)", 1994,
http://www.ietf.org/rfc/rfc1738.txt
[1] "PHP: urlencode - Manual",
http://www.php.net/manual/en/function.urlencode.php

[2] Section 2.3 of RFC2396 says:

| Unreserved characters can be escaped without changing the semantics
| of the URI, but this should not be done unless the URI is being used
| in a context that does not allow the unescaped character to appear.

--
Jock

[John Dunlop]

If I sound confused, that's because I am.

John Dunlop wrote:

Consider:

1. <http://domain.example/?baz=foo+bar>
2. <http://domain.example/?baz=foo%20bar>
3. <http://domain.example/?baz=foo%2Bbar>
[ ... ]
Reading Björn Höhrmann's explanation of reserved characters in

"Re: Good/Bad - URI encoding in HTML editor",
http://lists.w3.org/Archives/Public/...2May/0032.html

we see that numbers one and two are *not* equivalent.

Actually, I think, numbers one and two are equivalent. Hopefully I've
got this straight in my head now. :-)

RFC1630, which I hadn't read before, sums up Tim BL's original intent:

| Within the query string, the plus sign is reserved as shorthand
| notation for a space. Therefore, real plus signs must be encoded.
| This method was used to make query URIs easier to pass in systems
| which did not allow spaces.

According to RFC1738, sec. 3.3, however, plus signs weren't reserved
in the query component ("searchpart ") of an HTTP URL. That means they
had no reserved purpose, so a plus sign meant a plus sign, not a
space, and they didn't need encoded.

Then came along RFC2396 and the plus sign became reserved in the query
component again. Real plus signs must now be encoded. It doesn't say
what the reserved purpose is for plus signs. I guess, then, plus
signs are shorthand for spaces.

Previously, I was under the impression that a question mark mustn't
appear in query components. It seems I was wrong. A URI may contain
more than one question mark, although URI generators are discouraged
from generating such URIs. The second "?" should always be treated as
data by parsers. See

Roy T. Fielding, 2002-11-17, "Re: Ambiguity of Allowed/Recommended URI
Syntax and Escaping",
http://lists.w3.org/Archives/Public/...2Nov/0015.html

Refs.:

RFC1630 (informational) , 1994-06, "Universal Resource Identifiers in
WWW: A Unifying Syntax for the Expression of Names and Addresses of
Objects on the Network as used in the World-Wide Web",
http://www.ietf.org/rfc/rfc1630.txt

RFC1738 (proposed standard), 1994-12, "Uniform Resource Locators
(URL)",
http://www.ietf.org/rfc/rfc1738.txt

RFC2396 (draft standard), 1998-08, "Uniform Resource Identifiers
(URI): Generic Syntax",
http://www.ietf.org/rfc/rfc2396.txt

--
Jock

[Adriaan]

"John Dunlop" wrote

1. <http://domain.example/?baz=foo+bar>
2. <http://domain.example/?baz=foo%20bar>

Note that these are also the same as for using $_GET['baz'] (which by design
holds the decoded values). But when explicitely (manually) decoding
$_SERVER['QUERY_STRING']: http://php.net/rawurldecode does *not* convert +
characters into spaces!

Adriaan