<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<br>
<div class="moz-cite-prefix">On 11/12/14 1:45 PM, Brent Putman
wrote:<br>
</div>
<blockquote cite="mid:5463AAB2.8020301@georgetown.edu" type="cite">
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
<br>
<br>
<br>
I imagine it may be too late to change every place where we
instantiate a PRC to use a (new) 2-arg constructor - and that's
assuming the message context instances are even available at that
point. But you guys are more familiar. If it's just changing a
few places, etc, then maybe worth doing. <br>
</blockquote>
<br>
Actually, just realized, the MC are not necessarily typically
available, I think, in our current model. The inbound is generated
by the Decoder usually, after the PRC is ctor-ed. My memory is
hazy... I think this is accounted for by a difference in processing
models. I think I originally designed: 1) decode, so get an inbound
MC 2) create your InOutOperationContext with that (and a new
outbound MC), as context for the middle "business logic" processing
3) run "business logic" 4) encode outbound.<br>
<br>
So I think the actual design with the PRC (which Chad added in the
IdP, and was later moved to OpenSAML) diverged there. Now there's an
impedance mismatch (I think) - we assume you create a PRC upfront,
and then hang stuff off of it as you go.<br>
<br>
Have to think more - maybe there's a way to solve - like just
overriding and making the setParent(BaseContext) public on
MessageContext. Then wherever we set the MC's on the PRC, just also
set the parent explicitly. <br>
<br>
Unless this isn't a big deal, in which case we can just leave as-is
for now.<br>
<br>
</body>
</html>