Les is correct that my first and second questions were exact opposites.
Jeremy answered my first question and the solution is completely
satisfactory for handling the weird condition that occurs only in devel
mode.
My second question came up when I realized what Subject.logout() is
discarding the "remember me" cookie as a side effect of logout. I wanted
a way to have the user "de-authenticate" without being forgotten. Upon
further review I guess it makes little sense to do this. So I withdraw
my request and apologize for the all the noise.
Thanks,
Brad
Jeremy Haile wrote:
Brad,
My main concern is altering the Subject interface for what may be a rare
case and one that has a pretty easy workaround (just invalidate the
session). I think that usually when a user logs out, they expect their
remembered identity to be logged out as well. Currently, logout() marks
the user unauthenticated, forgets their remembered identity, and ends
their session - which is the behavior of most web apps that I'm familiar
with.
As Les mentioned, I'd love to hear more detail about your use case. If
this seems to be a common case, then we could definitely consider adding
the functionality to Subject.
Jeremy
On Thu, 31 Jul 2008 19:29:32 -0400, "Les Hazlewood" <[EMAIL PROTECTED]>
said:
Hi Brad,
Could you please summarize the exact end result that you're trying to
achieve? I'm a tad confused at the moment. Your first question was to
see
if the remembered identity could be explicitly forgotten and/or ignored
after a fresh DB install during development. The second question asks if
you can logout a user but still retain the remember me identity, which is
almost an exact opposite of the first question :)
Perhaps if you could explain exactly what you're trying to achieve, then
it
would be easier for me to think of how this might apply to our API.
I'm a little hesitant to alter the Subject interface (at least at the
moment) because this is one of the very core interfaces of JSecurity,
used
everywhere by everyone (in JSecurity's 'guts' as well as end-users. I
just
try to think long and hard before doing these kinds of things.
So, in getting clarification, I might be able to say "Oh, this is easy,
we
could modify some behavior in the SecurityManager implementation to
achieve
what you want" (template methods or whatever), which would be much more
digestible than changing such a critical API interface.
I'm definitely willing to try to make your life easy, I'm just confused
at
the moment by the polarity of the two questions ;)
Thanks!
Les
On Thu, Jul 31, 2008 at 6:03 PM, Brad Whitaker <[EMAIL PROTECTED]>
wrote:
I've been trying to use subject.getSession().stop() but I seem to be having
some problems with simply creating a new session. I'm not sure exactly what
the problems are (perhaps Grails or some other part of my framework has
manipulated the session in some way the gets lost when the session is
stopped.)
Anyway, the more I think about it the more I think this approach is really
taking a very long way around. Why not add a parameter to Subject.logout()
that gets passed to (Default) SecurityManager so that
DefaultSecurityManager.rememberMeLogout() isn't called? Would this work? It
seems much cleaner and direct than trying to kill and create a session
because we don't want the subject to be remembered.
Thanks,
Brad
Brad Whitaker wrot
I'll let you continue the implementation discussion on the other thread,
but will add my two cents from a user's perspective:
I'd like to see a method on the Subject that can logout the subject out
but preserves "remember me". This could be accomplished with either a new
method or a new parameter to logout(). Subject just seems like the right
place to put this functionality (although I don't understand the
implementation issues).
Les Hazlewood wrote:
I think it might be more 'correct' to do this in JSecurity via
subject.getSession().stop() method instead. If in an HTTP environment,
HttpSession.invalidate() will be called on your behalf. If not using
HTTP
container sessions (for whatever reason), it also does the appropriate
invalidation on the underlying implementation.
But this surfaces an interesting question for the development team:
If someone calls subject.getSession().stop(), should they be able to then
immediately call subject.getSession() and have it return a brand new
session?
Currently that doesn't happen. Any calls on that returned session would
throw an InvalidSessionException. Going back to the desire to prevent
these
exceptions from occurring when possible, isn't it a good idea to create a
new one?
I can't think of any reasons at the moment to not allow a new session to
be
created as described. I like the idea of making this possible. What do
you
guys think?
On Thu, Jul 31, 2008 at 1:28 PM, Jeremy Haile <[EMAIL PROTECTED]>
wrote:
Well, the way JSecurity works an explicit logout removes the "remember
me"
cookie. A session timeout will of course not remove the remember me
cookie.
So if the user doesn't log out and their session times out then when
they
go back to the site they are remembered. However, if the explicitly log
out
and return to the site, they will have to re-authenticate.
If you want to simply un-authenticate them, but not remove the remember
me,
you could just invalidate their current HTTP session by calling
HttpSession.invalidate(), but don't call Subject.logout(). Their next
request would start a new session which would be remembered, but not
authenticated.
Hope this helps!
Jeremy
On Jul 31, 2008, at 1:21 PM, Brad Whitaker wrote:
Thanks for the response -- I didn't realize that Subject.logout() would
remove the remember me cookies.
This behavior surprises me a little bit and leads to a different
question:
is there a way to "un-authenticate" a user? It seems it would valuable
to be
able to log a user out but still remember them. Am I missing this in
the API
or does this capability not currently exist?
Brad
Jeremy Haile wrote:
Hey Brad,
The usual way of forcing JSecurity to "forget" a subject is to call
Subject.logout() - this should remove any remember me cookies as well.
Perhaps you could auto-logout subjects in your development
environment upon
first access? You could also just bookmark the /logout URL and click
the
bookmark when you start a new development session.
This would be difficult to do on the server side (i.e. without a web
request from a browser), since it involves actually clearing the
cookie from
a user's machine.
Please let me know if you have any ideas about how JSecurity could
make
this process easier.
Jeremy
On Jul 31, 2008, at 12:11 PM, Brad Whitaker wrote:
Is it possible to force JSecurity to "forget" a subject that has
previously been remembered?
This is an issue for me only in "development" mode and shouldn't
occur
in a production environment. The problem is that I often start a
development
session with an empty user database but the browser comes to the site
with a
cookie. I end up getting a Principal that I don't know. I would like
to
discard the cookie at this point. Is this possible? Or is there a
better way
to deal with this issue (other than clearing the cache on the
browser)?
Thanks,
Brad
