You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Benjamin Pflugmann <be...@pflugmann.de> on 2002/11/18 20:06:04 UTC

http://[user[:pass]@]host/ syntax (was: Re: svn merge fails in 0.15)

Hi.

On Mon 2002-11-18 at 03:20:47 -0800, Daniel Rall wrote:
> Peter Davis <pe...@pdavis.cx> writes:
> 
> > In addition to simply accepting it, Subversion should fully support the 
> > http://[user[:pass]@]host/ syntax, which it doesn't today.  That's an issue I 
> > raised (and had problems with) a long time ago, but I can't remember if it 
> > was decided to not support it or to simply delay implementation.
> 
> +1, I'm a big fan of that notation.  I've been using it for years with
> a variety of software.

Just want to mention that foo@bar is not defined for the HTTP scheme,
but only for the FTP scheme, AFAIK. At least, last time I studied the
RFCs. I do not mean to deny using this notation - just that this use
is not covered by the standards for what I know.

Bye,

	Benjamin.



Re: http://[user[:pass]@]host/ syntax (was: Re: svn merge fails in 0.15)

Posted by Peter Davis <pe...@pdavis.cx>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Monday 18 November 2002 14:00, Benjamin Pflugmann wrote:
> That is a completely different issue.  http://host/path@REV is allowed
> by the standard. In the path, almost everything is allowed. Subversion
> just has a special interpretation of the path.

Oops, looks like you're right.  I retract my comment.

- -- 
Peter Davis
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE92YbEhDAgUT1yirARAobRAJwISoWtHe0sJAF7FqRGnotg5WcF1QCeIWc4
SQx3wifMU324tPxqPdlGnM0=
=zQz3
-----END PGP SIGNATURE-----


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: http://[user[:pass]@]host/ syntax (was: Re: svn merge fails in 0.15)

Posted by Benjamin Pflugmann <be...@pflugmann.de>.
Hi.

On Mon 2002-11-18 at 13:46:44 -0800, Peter Davis wrote:
> On Monday 18 November 2002 13:41, Matt Quail wrote:
> > Allowing a super-set of the
> > http schema defined by RFC 1738 may still mean you are a valid 1738
> > parser...
> 
> It's not as if Subversion doesn't already support a superset with support for 
> http://host/path@REV syntax (as good or bad as that syntax might be).

That is a completely different issue.  http://host/path@REV is allowed
by the standard. In the path, almost everything is allowed. Subversion
just has a special interpretation of the path. This was always up to
the server to decide and it is just an (unfortunate?) convention that
the path is often mapped directly to the filesystem (but there are
also many systems which already have other interpretations).

> In any case, it's definitely wrong to successfully parse the user@host URL and 
> then completely ignore the "user@".  Right now, if you specify such a URL, 
> Subversion will not give an error and will just proceed as if you had not 
> specified the user at all.
> 
> Either it's valid syntax, or it's not.  It should be supported, or else there 
> should be an error saying the URL is not valid RFC 1738 (or at least a 
> warning).

I agree. If it's been decided that user/password specification is not
supported, the parser should complain.

> > In any case, I definitely find the "username:password@" syntax useful!
> 
> So would I :)

No comment. As I said, I just wanted to make the fact known.

Bye,

	Benjamin.

Re: http://[user[:pass]@]host/ syntax (was: Re: svn merge fails in 0.15)

Posted by Peter Davis <pe...@pdavis.cx>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Monday 18 November 2002 13:41, Matt Quail wrote:
> Allowing a super-set of the
> http schema defined by RFC 1738 may still mean you are a valid 1738
> parser...

It's not as if Subversion doesn't already support a superset with support for 
http://host/path@REV syntax (as good or bad as that syntax might be).

In any case, it's definitely wrong to successfully parse the user@host URL and 
then completely ignore the "user@".  Right now, if you specify such a URL, 
Subversion will not give an error and will just proceed as if you had not 
specified the user at all.

Either it's valid syntax, or it's not.  It should be supported, or else there 
should be an error saying the URL is not valid RFC 1738 (or at least a 
warning).

> In any case, I definitely find the "username:password@" syntax useful!

So would I :)

- -- 
Peter Davis
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE92V/EhDAgUT1yirARAr2VAJ9cpzwvNyWs1ELIEZrKKA+DWjYI5QCfZdq2
/4+vBvLSUCUIxPPR74ohpRQ=
=6IW7
-----END PGP SIGNATURE-----


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: http://[user[:pass]@]host/ syntax (was: Re: svn merge fails in 0.15)

Posted by Matt Quail <ma...@cortexebusiness.com.au>.
> Just want to mention that foo@bar is not defined for the HTTP scheme,
> but only for the FTP scheme, AFAIK. At least, last time I studied the
> RFCs. I do not mean to deny using this notation - just that this use
> is not covered by the standards for what I know.

The "//user:password@host:port/path" notation is defined as "Common 
Internet Scheme Syntax" in secion 3.1 of RFC 1738. This section does not 
specify that the "http" URL schema follows this syntax, merely:

> The scheme specific
> data start with a double slash "//" to indicate that it complies with
> the common Internet scheme syntax.

Secion 3.3 "HTTP" says that the syntax for the http schema is

    http://<host>:<port>/<path>?<searchpart>

and further states:

> No user name or password is allowed.


D'oh!

So if you are a "producer" of http URLs, you would be breaking RFC 1738 
if you put the "username:password@" in a http URL. But I suppose you 
could be more lenient as a "parser" of URLs. Allowing a super-set of the 
http schema defined by RFC 1738 may still mean you are a valid 1738 
parser...

In any case, I definitely find the "username:password@" syntax useful!

=Matt

-- 
Matt Quail -- Senior Software Engineer
http://www.cortexebusiness.com.au/
mailto:matt@cortexebusiness.com.au



---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org