<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<br>
<div class="moz-cite-prefix">On 4/3/14 2:36 PM, Cantor, Scott wrote:<br>
</div>
<blockquote cite="mid:CF63219A.4C89E%25cantor.2@osu.edu" type="cite">
<pre wrap="">
</pre>
<pre wrap="">
But I actually mispoke in my last suggestion, we can't actually wire in a
SecurityConfiguration with the key because we need the resolver to work
for that, which is why the auto-wired code now is poking in a Parameters
object directly.
But once you have the resolver working we can make a little more progress.</pre>
</blockquote>
<br>
I haven't looked at any of this, but couldn't you just wire in a
Static*Resolver for now and change it when I have the Basic- one
working? Of course you'd have to summon up the signing credential,
etc, to inject in to the static resolver, so that would obviously
have to be faked up. But you could a least move "more forward" that
way. <br>
<br>
I guess it's just a question of which bits you go back in and change
later.<br>
<br>
<br>
<blockquote cite="mid:CF63219A.4C89E%25cantor.2@osu.edu" type="cite">
<pre wrap="">
I did want to note that Rod's going to need classes to build the
*Configuration objects to serve to the rest of the IdP, and right now you
have some Basic*Confguration beans in the xmlsec-impl module. We might
want to move those into -api or we'll just end up duplicating them.</pre>
</blockquote>
<br>
Yeah, frack. I was thinking/hoping that those would all just be
beans wired from plain Spring beans files, etc, and the IdP wouldn't
have to actually know about them. Or maybe they'd be in -spring
modules, where I thought we relaxed the no-depending-on-impls
policy. But if that's not true... yeah. I'm wondering whether
there are other bits like this where the -impl dependency policy is
waiting to bite us.<br>
<br>
</body>
</html>