ResolverScriptAttributeDefinition with jruby1.7.3 and idp-2.4.0

[Cantor, Scott]

On 7/26/13 4:24 AM, "Francesco Malvezzi" <francesco.malvezzi at unimore.it>
wrote:

>If I understood right, shibboleth-idp 2.4.0 does not need anymore to
>bridge execution of scripting languages in ScriptAttribute with jsr223.
>So for example the jars like: jruby-engine-20080611.jar are not
>included, and everythings simply works by dropping the jruby.jar in the
>expanded source lib directory.

No, what happened is we were shipping stale code that had supposedly been
incorporated directly in various scripting implementations themselves by
now. Whether any given engine works is beyond me, I don't know how any of
them work.

The other issue was that those stale classes essentially overlapped with
the ones you'd get in many newer jars, and so it would be random which one
was used.

>However, after upgrading to idp-2.4.0, I have noticed that now,
>sometimes, randomly, some scripted attribute definitions fail with a
>javax.script.ScriptException like:

That would seem to imply, with my naive understanding, that they have a
regression in their newer engine.

-- Scott