[YANGTOOLS-489] NPE in AbstractStringInstanceIdentifierCodec

Sun, 10 Apr 2022 18:35:36 +0000

I have a data model with an instance-identifier leaf however the YangInstanceIdentifier value did not conform to a valid schema. This resulted in an NPE in AbstractStringInstanceIdentifierCodec#serialize at line 36:

@Override
public final String serialize(final YangInstanceIdentifier data) {
StringBuilder sb = new StringBuilder();
DataSchemaContextNode<?> current = getDataContextTree().getRoot();
for (PathArgument arg : data.getPathArguments()) {
current = current.getChild(arg); <--- line 36

if(current.isMixin())

{ /* * XML/YANG instance identifier does not have concept * of augmentation identifier, or list as whole which * identifies mixin (same as paretn element), * so we can safely ignore it if it is part of path * (since child node) is identified in same fashion. * */ continue; }

...
}
}

It looks like the rest of the code would work OK w/o the isMixin() check since the ID doesn't have a backing schema so it could check for null "current" and ignore. However if the intent is for leaf instance-identifier values to correspond to a valid path/schema in the tree, then we should emit an exception (eg IllegalArgumentException) with useful context in the message.

OpenDaylight JIRA

[YANGTOOLS-489] NPE in AbstractStringInstanceIdentifierCodec