W3C home > Mailing lists > Public > www-validator@w3.org > August 2001

Re: Warn about unknown parameters

From: Terje Bless <link@pobox.com>
Date: Wed, 1 Aug 2001 06:48:45 +0200
To: Nick Kew <nick@webthing.com>
cc: W3C Validator <www-validator@w3.org>
Message-ID: <20010801064851-r01010700-7d2f8e5a-0910-010c@>
On 01.08.01 at 05:06, Nick Kew <nick@webthing.com> wrote:

>>Waitaminute... We end up requesting a different page and that page
>>happens to be invalid?

Ok. On the same page so far...

>>So the extra params aren't affecting _our_ behaviour, it's the _lack_ of
>>params that makes the _validated_ server change _it's_ ´behaviourª?
>No.  Validator requests wrong page.  Validator's fault.

GIGO: Garbage In, Garbage Out! :-)

>>   IOW, it's not an error in the Validator -- the error lies with
>> the user failing to escape CGI params
>By what standard is ;1 a CGI param?  Not by anything official.

«Anything official» specified «&» as the separator.

I rest my case... :-)

>Indeed there is only one issue: does a character need URLEscaping?
>*  If yes, then we should complain of the URL, not the document
>*  If no, then the author is fully entitled to use it and expect
>   user agents to cope.
>Both the Valet and the WDG Validator are happy with

So is the Validator if you enter the URI in the form interface at
<URL:http://validator.w3.org/> so that the browser properly escapes those
characters (I think. See below). It's only when you try to link directly to
it and don't escape the «;» that it ends up fetching the wrong page. This
would merit a warning that the may user need to escape the character in
question, but nothing more I don't think.

Of course, it may be that what's /really/ happening is that the browser
contructs an URL using «&» as the param separator and CGI.pm thus doesn't
get confused by the «;» as it's now ignoring them in favour of the «&».

Question: do your tools do anything to _explicitly_ deal with this issue or
does it just happen to work? You've rolled your own CGI handler for Page
Valet, no? What separators does it accept and what if anything does it do
to prevent this problem? IIRC, Liam used CGI.pm for the WDG Validator; does
it do anything in particular to deal with this?

It sure don't look like it:
Received on Wednesday, 1 August 2001 00:48:51 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 14:17:30 UTC