W3C home > Mailing lists > Public > www-validator@w3.org > September 1999

Re: [PATCH] use CGI.pm;

From: Terje Bless <link@tss.no>
Date: Wed, 29 Sep 1999 14:21:28 +0200
Message-Id: <199909291224.OAA22199@vals.intramed.rito.no>
To: Matthew Wickline <listbox@wickline.org>
cc: Gerald Oskoboiny <gerald@w3.org>, W3C Validator <www-validator@w3.org>
On 29.09.99 at 06:02, Matthew Wickline <listbox@wickline.org> wrote:

>I haven't tested this code, so pardon typos :) What if you pre-process the
>query string before cgi.pm gets called?

Yeah, that would work, but that brings us dangerously close to the old way
where we process the query string ourselves. The very thing which CGI.pm
was supposed to avoid.

>Above won't work for POSTs. Does anyone use POST to call the validator?

Yes[0]! And when file upload is added, the form on
<URL:http://validator.w3.org/> will of necessity use POST.

>I suppose that sort of strategy could be finessed to work with POSTS as
>well, if necessary.

Yes, but it would be ugly as sin and slow as molasses. Believe me, you do
*not* want to try parsing nested multipart MIME messages by hand!

>Just pre-process any post request until it looks like a (properly munged)
>GET request by the time it gets to cgi.pm


>If that works, perhaps that old funky style of passing parameters in the
>PATH_INFO could be supported [...]

Yeah, but we could do that anyway. It was removed on purpose because IMO
that is a horrible interface to encourage. It serves no purpose, it's
confusing and it makes the code bloat into a big hairy mess. I say good

[0] - Well, it's probably only me at the moment, as that requires the
      CGI.pm patch which hasn't landed on the trunk yet, but that is
      likely to change once it does. :-)
Received on Wednesday, 29 September 1999 08:25:28 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 14:17:25 UTC