On 22/09/2007, at 11:25 PM, Jonathan Weiss wrote:
Tim Lucas wrote:
How about simply removing #authentication from the #query_revision
svn command args?
I need local auth as there are cases where you do not have an auth
cache
(do not want to have one), best example ist Webistrano.
What
I believe I am just going to put the --password switch back in. It is
known to work in all relevant cases, and doesn't overcomplicate the
code. If security concerns you, it's not the end of the world: it
would be relatively straightforward to subclass the Subversion module
and add your own
Tim Lucas wrote:
If you still needed the flexibility of specifying local auth details
maybe you could have #query_revision use a #local_authentication
which drew from two local-only vars: :scm_local_username
and :scm_local_password ?
Would that suffice?
If it would default to
Thanks Jamis.
On 19/09/2007, at 2:50 AM, Jamis Buck wrote:
I'm *this* close to putting the --password option back in, and to heck
with security concerns. Pulling that option has opened a can of worms.
:(
Unfortunately, handling local command streams similarly to remote
streams is
Tim Lucas wrote:
How about simply removing #authentication from the #query_revision
svn command args?
I need local auth as there are cases where you do not have an auth cache
(do not want to have one), best example ist Webistrano.
Jonathan
--
Jonathan Weiss
http://blog.innerewut.de
On 02/09/2007, at 10:27 AM, Jamis Buck wrote:
*Feature*: The subversion SCM module no longer uses the --password
option to authenticate. Instead, it waits until the server prompts for
the subversion password and send it that way. This is more secure
(ps won't show the password anymore), but