[10230] in www-talk@info.cern.ch
Re: PUT method
daemon@ATHENA.MIT.EDU (jna)
Sat Aug 17 15:33:04 1996
Date: Sat, 17 Aug 1996 13:34:40 +0500 (GMT+0500)
From: jna <jna@retina.net>
To: Jeremey Barrett <jeremey@forequest.com>
Cc: www-talk@w3.org
In-Reply-To: <Pine.BSI.3.93.960817102001.25759A-100000@descartes.forequest.com>
Resent-From: www-talk@w3.org
On Sat, 17 Aug 1996, Jeremey Barrett wrote:
> On Sat, 17 Aug 1996, I stupidly wrote (jna):
>
> >
> > What exactly is a PUT method? At last check, there was no such method.
>
> PUT is part of the HTTP/1.0 spec in the Appendix on 'Additional Features',
> with the caveat that it may or may not be implemented or implemented
> correctly. Is is used for replacing a particular URI. For example,
> let's say you are remotely editing /blah/blah.html, and wish to replace
> it. In theory, a PUT header followed by the document would replace
> the existing one. There is a bunch of server config necessary to make this
> really happen most likely.
I do remember this now, I guess I spaced the other day; It's supported by
a few modules for spinner that allow you to remotely edit pages. I also
believe that illustra put support for it in their web server as well.
> PUT does in fact work, if passed to a CGI script. Using netscape gold,
> you can 'Publish' a document via HTTP, which sends a PUT request followed
> by the full HTML text of the document, as it should. A CGI reads this from
> stdin, and does receive the document text. My problem is that the
> netscape browser converts forms with PUT methods to GET requests before
> sending them to the server, and I have no idea why.
Yep. Okay. Thanks for clearing this up. :) Has an RFC been established to
account for the handling of user authorization and authentication when
modifing files via a PUT method? It's probably an issue best left to
particular implementations, but some form of standard would be nice. A
typical implementation might rely on basic authentication first, and then
use the REMOTE_USER id to evaluate permissions on the file being modified.
(just a guess)
--john
