Bug #7746
Node registration update fails when <contactSubject> spans multiple lines
30%
Description
Node registration update fails when the string in the node registration document (attached) is on a separate line from the XML element tags. Apparently, parsing this field results in a subject string that does not match the LDAP record correctly, and the result is 401 not authorized exception:
CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org
results in:
<?xml version="1.0" encoding="UTF-8"?>
CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org
is not a Registered Subject
Correcting to:
CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org
succeeds.
DFC-node.xml 
(1.21 KB)
Subtasks
History
#1 Updated by Rob Nahf over 8 years ago
- Category set to d1_cn_noderegistry
The exception thrown is called when cnIDentify.getSubjectInfo(subject) returns a NotFound.
Calling https://cn-sandbox.test.dataone.org/cn/v2/accounts,
I see Lisa's subject string listed as
CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org
Lisa
Stillwell
true
without space characters leading the string.
In the Node document, after the new line, there are also 8 space characters preceding the non-whitespace part, so it is seeing the them as different subjects.
I think the following in the Node document would work as well.
CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org
Subject elements need to accept whitespace (internally at least), so I wonder if we need to tighten up validation to disallow leading and trailing whitespace, assuming it's possible?
#2 Updated by Rob Nahf over 8 years ago
- Status changed from New to In Progress
- % Done changed from 0 to 30
#3 Updated by Rob Nahf over 8 years ago
I tested all three subject strings and all of these pass when we call with libclient's CNode.getSubjectInfo(subject):
newlines and leading spaces: "\n CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org\n"
leading spaces: " CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org"
trimmed: "CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org"
newlines only: "\nCN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org\n"
current speculations:
Maybe the serialization / deserialization done by the client normalizes the string, where calling the identityService directly(?) does not?
Maybe the xslt parsing rules on subject and contactSubject are different?
#4 Updated by Rob Nahf over 8 years ago
cn-rest calls the identity manager directly, bypassing serialization of the request. (implementation class is org.dataone.service.cn.impl.v2.SystemMetadataDaoMetacatImpl in cn_identity_manager). The remaining question is figuring out how serialization process trims whitespace.
#5 Updated by Rob Nahf over 8 years ago
Using the TypeMarshaller, and TypeFactory.clone, I was not able to use serialization to remove leading newlines and other whitespace, as I thought it would.
It's unclear to me why calling cn.getSubjectInfo using a subject that has leading and trailing whitespace doesn't throw and exception when updateNode does...
Parking this issue for now. Below is code for a couple of the tests I did.
@Test
public void testSubjectWhitespace() throws ServiceFailure, ClientSideException, IOException {
org.dataone.client.v1.CNode cn = D1NodeFactory.buildNode(org.dataone.client.v1.CNode.class, new DefaultHttpMultipartRestClient(), URI.create("https://cn-sandbox.test.dataone.org/cn"));
Subject subject = TypeFactory.buildSubject("\n CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org\n");
try {
cn.getSubjectInfo(null, subject);
} catch (NotAuthorized | NotImplemented | NotFound | InvalidToken e) {
System.out.println(e.getDescription());
Assert.fail("Should not be not authorized");
}
}
@Test
public void testNodeParsing() throws InstantiationException, IllegalAccessException, IOException, JiBXException {
org.dataone.service.types.v1.Node node = TypeMarshaller
.unmarshalTypeFromFile(org.dataone.service.types.v1.Node.class, "/Users/rnahf/Downloads/DFC-node.xml");
org.dataone.service.types.v1.Node serializedNode = TypeFactory.clone(node);
List<Subject> sl = serializedNode.getContactSubjectList();
for (Subject s: sl) {
System.out.printf("{{%s}}\n",s.getValue());
}
TypeMarshaller.marshalTypeToFile(
TypeFactory.buildSubject(
"\n CN=Lisa Stillwell A15851,O=University of North Carolina at Chapel Hill,C=US,DC=cilogon,DC=org\n"),
"/Users/rnahf/Downloads/DFC-subject.xml");
Subject serializedSubject = TypeMarshaller.unmarshalTypeFromFile(Subject.class, "/Users/rnahf/Downloads/DFC-subject.xml");
System.out.printf("{{%s}}\n",serializedSubject.getValue());
}
