JSpecify: infer generic method type arguments for assignments

[haewiful]

This PR infers nullability of generic method type arguments for the case where the result of the call is assigned to a variable, e.g.:

  class Foo<T extends @Nullable Object> {
    Foo(T t) {}
    static <U extends @Nullable Object> Foo<U> make(U u) { return new Foo<>(u); }
    void test() {
      Foo<@Nullable Object> f1 = Foo.make(null); // legal
      Foo<Object> f2 = Foo.make(null); // error reported; arg cannot be @Nullable
    }
  }

See the tests for (many) more examples.

[msridhar]

@haewiful thanks for this! It looks like CI fails on the :nullaway:buildWithNullAway task:

https://github.com/uber/NullAway/actions/runs/12978381939/job/36192901792?pr=1131#step:5:606

I suggest looking at the code it is crashing on and trying to write a unit test that causes the same failure, which will make it easier to debug.

[codecov]

Codecov Report

Attention: Patch coverage is 91.13924% with 7 lines in your changes missing coverage. Please review.

Project coverage is 88.16%. Comparing base (baf8f77) to head (fe8702d).
Report is 3 commits behind head on master.

Files with missing lines	Patch %	Lines
.../InferSubstitutionViaAssignmentContextVisitor.java	79.31%	3 Missing and 3 partials ⚠️
...ava/com/uber/nullaway/generics/GenericsChecks.java	96.87%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff            @@
##             master    #1131   +/-   ##
=========================================
  Coverage     88.16%   88.16%           
- Complexity     2281     2300   +19     
=========================================
  Files            87       88    +1     
  Lines          7461     7531   +70     
  Branches       1491     1503   +12     
=========================================
+ Hits           6578     6640   +62     
- Misses          445      448    +3     
- Partials        438      443    +5     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[msridhar]

Hrm, this seems suspect. If you create a new GenericsChecks object here, then the inferred types will immediately disappear. Do we need to instead somehow pass in the GenericsChecks object from the NullAway instance into this class?

[haewiful]

Yes, I do think the GenericsChecks object should be passed from NullAway because getGenericReturnNullnessAtInvocation() method call uses what is in inferred types but doesn't add to it. If the returned type contains some type variable in the inferredTypes, it should be used here.
In my examination, the AccessPathNullnessPropagation class is used in AccessPathNullnessAnalysis which is used in the NullAway file. I think we should pass the GenericsChecks instance from NullAway to AccessPathNullnessAnalysis to AccessPathNullnessPropagation.

[msridhar]

Ok done in bfccfad

[msridhar]

@haewiful this logic here is incorrect. We shouldn't put the upper bound in the map; we should put lhsType, except that if lhsType is @Nullable, we should strip that annotation off of it. (I think you can do that by calling stripMetadata().) See the new test inferNestedNonNullUpperBound() that I added; we should get an error there, but we do not right now.