Error for Orchestrator intent with entities is inconsistent & interferes with some scenarios

[scheyal]

Repro

• Create a dialog with Orchestrator recognizer (or use root)
• Add an intent trigger with ML entities --> Error (see below)
• Remove ML entities, go to editor right pane for intent, add ML entity --> no error

Expected:

• Consistent errors/warnings across
• Since Orchestrator simply ignores ML entities, it is useful to allow it for use cases when copying it from other LUIS based dialog for routing.
• Consider: In add trigger UI, just display a non-blocking warning. Do not display in right intent pane cause it may be by user design.

Visuals

When adding intent we get an error:

But in another pane it is just fine:

[darrenj]

I see this when Orchestrator is enabled (warnings). This is on LU trigger editing on a dialog

However as Eyal has found if you edit the LU on the new trigger dialog you don't get the same consistent warnings/errors

[mewa1024]

I get the same behavior as Darren on 2.0.0-nightly.255657.9d2aeef, namely if I switch from the default recognizer to Orchestrator my existing trigger phrases with entities do not show a warning. The warning is only triggered when I start editing the trigger phrases.

The reverse is also true: If I am using Orchestrator and have entities in my trigger phrases, I see the warnings. If I then switch back to the default recognizer, the warnings persist until I start editing:

https://user-images.githubusercontent.com/39317542/124047204-76844a80-d9c8-11eb-9536-88e276d6216e.mov

It seems like we should run validation on the trigger phrases when the recognizer is changed, not just when phrases are edited.

[mewa1024]

(removing UX design label)