Re: authentication storage question
From: Ben Collins-Sussman <sussman_at_collab.net>
Date: 2001-09-18 20:10:21 CEST
Dale Thatcher <subversion@dalethatcher.com> writes:
> I've been trying to find a decent place for the proxy callbacks
Look at the new svn_ra.h -- it might be better for you now.
Here's the new system, as simply as I can put it:
* an administrator configures the svn server (apache) to require
* the svn client attempts to do something (checkout, commit, etc.)
* libsvn_client creates a table of callbacks that are able to
* if neon is never challenged by apache:
the callback table is ignored.
* if neon -is- challenged by apache:
neon uses the callback table to "pull" authentication data
- if authentication fails, neon returns an error immediately
- if authentication succeeds, the svn subcommand finishes,
So really, the old system made some erroneous assumptions -- it
Really, apache is in total control. Apache dictates exactly what auth
To make Subversion use a new authentication protocol called "FooAuth",
* make sure apache understands the protocol, and can issue a
* make sure neon knows how to respond to a FooAuth challenge
* make sure libsvn_client knows how to provide the specific FooAuth
---------------------------------------------------------------------
|
This is an archived mail posted to the Subversion Dev mailing list.
This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.