fix: separate out EntityMutationValidator

[wschurman]

Why

Now that these are used for both triggers and validators, they should have a distinct name. I considered the following, but am open to others suggestions as well:

• EntityMutationCallback - callbacks usually don't prevent something from running, so it didn't make sense for validation
• EntityMutationHandler - handlers generally refer to overriding behavior of the framework rather than hooking into it.
• EntityMutationExecutable- the method inside it is called executeAsync so this was the natural name
• EntityMutationLambda - not really a JS term
• EntityMutationHook - The term hook refers more to the when it is run rather than the what is run IMO.

How

TS renames.

Test Plan

yarn tsc

[codecov]

Codecov Report

Merging #68 into master will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #68      +/-   ##
==========================================
+ Coverage   94.55%   94.57%   +0.02%     
==========================================
  Files          59       59              
  Lines        1488     1494       +6     
  Branches      167      168       +1     
==========================================
+ Hits         1407     1413       +6     
  Misses         79       79              
  Partials        2        2              

Flag	Coverage Δ
#integration	94.57% <100.00%> (+0.02%)	⬆️
#unittest	94.57% <100.00%> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
packages/entity/src/EntityCompanion.ts	100.00% <ø> (ø)
...s/entity/src/EntityMutationTriggerConfiguration.ts	100.00% <ø> (ø)
packages/entity/src/EntityCompanionProvider.ts	100.00% <100.00%> (ø)
packages/entity/src/EntityMutator.ts	98.02% <100.00%> (+0.03%)	⬆️
packages/entity/src/EntityMutatorFactory.ts	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fc4377d...3334412. Read the comment docs.

[ide]

"Executable" sounds like a program. Here is a proposal:

Make two named types: a trigger type and a validator type. Both might be aliases for the same underlying type.

If we never intend to treat triggers and validators the same, they don't need to share an interface. In fact it might be desirable for them not to share and interface so that we don't imply that triggers and validators are interchangeable. Put another way, TS interfaces don't let us declare intent, and the intents behind triggers and validators differ.

I also think it's worth considering just using a plain function for the type unless there is a need to add more methods to the trigger interface.

[wschurman]

That's good reasoning for separating out the types. I updated the PR to reflect that.

As for the class vs function, see this comment: #65 (comment). I'll definitely consider making it a function in the future if usage seems to not make use of the benefits of a class, but having it as a class has turned out to be reasonably helpful for privacy policy rules.