|
As noted in the code review, the test case seems to be showing the effect of a previous bug. Please update the test case to correctly show the bug.
|
|
After adjusting the test case, it is obvious the fault lies with mdsal-binding-dom-codec, which fails to perform proper encoding in both directions, as it is using no-op codec.
|
|
The test model also demonstrates that there is something wrong with base codegen, as the target type is 'java.lang.Enum' whereas it should be OutputA. That may actually be the root cause for this behavior.
|
|
So the codec problem actually stems from the codegen problem – since the getter method has the wrong return type, we are ending up not looking up the proper codec and defaulting to no-op.
|
|
We seem to be using a different codepath (or ordering?), as a plain container with the same type of leaf works correctly, whereas RPC input ends up going to BaseYangTypes.javaTypeForYangType(), which is returning the java.lang.Enum we end up using.
|
Generated at Wed Feb 07 20:10:05 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.