Hello Max,
mmm, I would need to get my hands dirty again with this. This reminds me
that I had an issue with the logout verification signature, see here [1].
Disable the signature for the auth request would be acceptable for your
system? Our security team it is OK with this, or maybe they never realized
;)
Cheers,
Luis
[1]
http://lists.jboss.org/pipermail/keycloak-user/2018-September/015420.html
El vie., 8 feb. 2019 a las 9:34, <max(a)mascanc.net> escribió:
Hi,
On Wed, Feb 06, 2019 at 02:13:46PM +0100, Luis Rodríguez Fernández wrote:
> May I ask you what is the client implementation? For my dev environment,
> using the tomcat saml adapter in the SP side and Keycloak
> 4.8.2.Final-SNAPSHOT in the IdP one is working:
It is strange: going in remote debug with eclipse (running in local in my
MacOS),
I have been able to obtain a succesful redirect, and I did not see any
trivial
points on how the assertion signature could be damaged.
I'll investigate for encoding issues on the Linux machine.
In the code, the only point in which the assertion is marshalled to DOM is
through
a call to parse() on the inputstream.
The DOM builder factory is assigned to the threadlocal: why? Can it be
a threading issue, knowing the no thread safety of the dom implementation?
--
"Ever tried. Ever failed. No matter. Try Again. Fail again. Fail better."
- Samuel Beckett