<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<br>
<br>
<div class="moz-cite-prefix">On 2/4/16 4:56 PM, Liam Hoekenga wrote:<br>
</div>
<blockquote
cite="mid:CAH4ZtKRCdPVMhJRW9Hg83K_g4zAXEu8VAO521ifvBsSS+Q4F+w@mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote"><br>
<div>My bad. I like pretty printing. Here's the
un-altered..</div>
<div><br>
</div>
<div><br>
</div>
</div>
</div>
</div>
</blockquote>
<br>
That looks good. I even threw that into a file and OpenSAML
ParserPool can parse just fine. So if that's literally the entirety
of the request body, it ought to work...<br>
<br>
<blockquote
cite="mid:CAH4ZtKRCdPVMhJRW9Hg83K_g4zAXEu8VAO521ifvBsSS+Q4F+w@mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">
<div> <br>
</div>
</div>
As downloaded, it does use the wrong content type. I took a
hint from John Dennis's message and added </div>
<div class="gmail_extra">-H "Content-Type: text/xml;
charset=utf-8"<br>
</div>
<div class="gmail_extra"><br>
</div>
<div class="gmail_extra">The unmodified script does use
application/x-www-form-urlencoded. The SOAP request looks the
same...</div>
</div>
</blockquote>
<br>
Ah, ok. That makes more sense.<br>
<br>
<br>
<blockquote
cite="mid:CAH4ZtKRCdPVMhJRW9Hg83K_g4zAXEu8VAO521ifvBsSS+Q4F+w@mail.gmail.com"
type="cite">
<div dir="ltr"><br>
<div class="gmail_extra">The headers look the same, save the
content-type...</div>
<div class="gmail_extra">
<div class="gmail_extra">> POST /idp/profile/SAML2/SOAP/ECP
HTTP/1.1</div>
<div class="gmail_extra">> Host: <a moz-do-not-send="true"
href="http://shib-idp-test.www.umich.edu">shib-idp-test.www.umich.edu</a></div>
<div class="gmail_extra">> User-Agent: curl/7.46.0</div>
<div class="gmail_extra">> Accept: */*</div>
<div class="gmail_extra">> Authorization: Basic
*************************</div>
<div class="gmail_extra">> Content-Length: 735</div>
<div class="gmail_extra">> Content-Type:
application/x-www-form-urlencoded</div>
<div class="gmail_extra"><br>
</div>
</div>
<div class="gmail_extra">...and both content types result in the
IdP throwing a 500 (tho maybe for different reasons?)<br>
</div>
</div>
<br>
</blockquote>
<br>
<br>
Yeah, so with the right content type, does the IdP still log the
same XML parsing error? If so, I'm at a loss. Need to confirm, b/c
it's entirely possible that it's getting past the parsing but then
throwing a fatal error which results in the 500 and SOAP fault you
previously posted, for example something specific to the authN or
ECP case. For example, not sure off-hand what that would return if
for example basic authN isn't configured correctly and so there's no
REMOTE_USER visible in the request. <br>
<br>
<br>
</body>
</html>