W3C home > Mailing lists > Public > www-validator@w3.org > February 2002

"+" chars in links as bad as ampersands?

From: Greg Roelofs <newt@pobox.com>
Date: Mon, 18 Feb 2002 14:43:14 -0500 (EST)
Message-Id: <200202181942.g1IJgtre028137@newbolt.sonic.net>
To: www-validator@w3.org
Cc: maps@activeworlds.com
[not a list member]

To the best of my knowledge--and according to the validator itself--a link
(anchor tag) may have a plus-character in the URL.  However, if the plus-
character is part of the "uri" argument passed to the validator itself, it
gets converted to "%20" (encoded space character), which seems entirely
bogus and inexplicable.  It's *possible* this is a browser bug (NN 4.7 under
Linux), but since the Opera 6.0 beta does exactly the same thing, I would
tend to suspect the validator.

Here's a test page, itself automatically generated:

	http://mapper.activeworlds.com/cgi-bin/map/map-aw/c0/aw+00810+01350.html?90,90

At the very bottom left is a period linked to the validator; its link is
as follows:

	http://validator.w3.org/check?uri=http://mapper.activeworlds.com/cgi-bin/map/map-aw/c0/aw+00810+01350.html%3F90,90

Either following that link or typing it in the browser's address bar results
in the following (from the validator's URI entry field):

	http://mapper.activeworlds.com/cgi-bin/map/map-aw/c0/aw%2000810%2001350.html?90,90

Since the requested page (with spaces) doesn't exist, the validator fails.
However, entering the "+"-enabled URL in the validator's URI entry field
directly works just fine, as does modifying the browser's address bar to
use %2b instead of plus-characters.  (Note that either approach results in
the validation of the original, unmodified page, not of a page itself modi-
fied to use %2b's.)

So, question:  _should_ I have to escape my plus-characters when they appear
as an argument to the validator's "check" script (is that a special CGI ex-
ception of some sort), or is the script in error?

Final note:  before we added an explicit character-encoding to the pages,
at least some versions of Internet Exploder would default to UTF-7, which
apparently treats "+[digits]" as some sort of encoded character.  Is that
something the validator might be doing, too (or UTF-8, more likely)?

Thanks,
-- 
Greg Roelofs            newt@pobox.com             http://pobox.com/~newt/
Newtware, PNG Group, Info-ZIP, AlphaWorld Map, Philips Semiconductors, ...
Received on Thursday, 21 February 2002 00:40:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:00 GMT