Fix handling of static imports from subclasses

[msridhar]

Fixes #764

[codecov]

Codecov Report

Attention: 5 lines in your changes are missing coverage. Please review.

Comparison is base (3fc9b8a) 86.98% compared to head (d615df2) 86.97%.
Report is 2 commits behind head on master.

Files	Patch %	Lines
...c/main/java/com/uber/nullaway/NullabilityUtil.java	62.50%	1 Missing and 2 partials ⚠️
...away/src/main/java/com/uber/nullaway/NullAway.java	90.00%	0 Missing and 1 partial ⚠️
...llaway/dataflow/AccessPathNullnessPropagation.java	85.71%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##             master     #904      +/-   ##
============================================
- Coverage     86.98%   86.97%   -0.02%     
- Complexity     1959     1965       +6     
============================================
  Files            77       77              
  Lines          6330     6348      +18     
  Branches       1223     1229       +6     
============================================
+ Hits           5506     5521      +15     
- Misses          420      421       +1     
- Partials        404      406       +2     

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

[yuxincs]

Changes LGTM, but I have one clarifying question due to my unfamiliarity with the rest of the codebase.

[yuxincs]

question: I'm a bit confused about the previous code. It suggests that we actually want the methodSymbol and we only resolve to baseSymbol under certain conditions (i.e., interface method corresponding to one from java.lang.Object). But it seems with the updated code that we actually always want baseSymbol?

Am I missing something?

[msridhar]

Yes, that's a good question, this was due to my misunderstanding before. Basically, always getting the base symbol is safe, since if there is no different base symbol the baseSymbol() method just returns this. I didn't understand this before; more detail in this comment. In the next EP version the getSymbol overload for MethodInvocationTree will always call baseSymbol(), so in some cases it may get called twice, but that is fine.