Compute a call's executable type once per generic inference

[ericschlarmann]

Since #1654, the executable type used for inference is the method's type as a member of its receiver, and computing it infers the receiver when the receiver is itself a generic call. inferCallType computed it twice for a call not yet inferred (constraints, then return type), and runInferenceForCall a third time for lambda argument types. Inference run from dataflow is never cached, so every generic call in a receiver chain doubled the work below it: checking a fluent chain such as r.path(..).entity(..).isEqualTo(..) repeated n times took O(4^n).

Compute the executable type once and pass it to runInferenceForCall and generateConstraintsForCall. At 9 links, dataflow inference runs drop from 524,259 to 315; a 96-link chain now checks in under 1s.

Adds GenericMethodTests.longChainOfGenericInstanceMethodCalls, which times out on master and passes in about 0.1s with the fix. No existing test's diagnostics change.

I used Claude Code to bisect the regression to #1654, find the cause and write the fix and test. I have reviewed and understood all the changes.

Assisted-by: Claude Code (claude-opus-5-5)

Summary by CodeRabbit

• Bug Fixes

  • Improved generic method-call inference consistency, including for nested calls and chained method calls.

• Tests

  • Added a regression test covering nullable-argument detection in a long chain of generic method calls.

[CLAassistant]

All committers have signed the CLA.

[coderabbitai]

Navigate logical layers of code changes, visualize relationships, and explore their blast radius.

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository: uber/NullAway/.coderabbit.yaml

Review profile: ASSERTIVE

Plan: Advanced

Run ID: 50652e1e-604c-446b-b063-90571a9abf24

📥 Commits

Reviewing files that changed from the base of the PR and between d5bcf3c and 664b66d.

📒 Files selected for processing (2)

• nullaway/src/main/java/com/uber/nullaway/generics/GenericsChecks.java
• nullaway/src/test/java/com/uber/nullaway/jspecify/GenericMethodTests.java

Included review availability: This review used your included allowance. Your plan provides up to 8 included reviews per hour; 7 remain after this review.

---

Walkthrough

Generic-call inference now computes executable types at call sites and reuses them for constraint generation, return-type substitution, and argument mapping. Nested generic calls and generic method-type substitution also pass computed executable types into inference. A new timeout-limited test checks that a 24-segment generic method chain reports a nullable argument at its final .path(null) call.

Suggested reviewers: msridhar

Priority: ⬇️ Low

Change: Bug fix

Merge Risk: ⚪ Minimal · up to 664b6

No merge-blocking issue is established; this change is ready for normal checks.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 62.50% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 8 functions across 2 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and concisely describes the main change: computing a generic call's executable type once during inference.
Description check	✅ Passed	The description directly explains the performance problem, the implementation change, and the added regression test.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

• Fix all pre-merge checks with AI

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create a new PR

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[msridhar]

Thanks for the great fix and the test! I had a feeling I missed some pathological cases like this.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 87.67%. Comparing base (51ef189) to head (664b66d).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #1896      +/-   ##
============================================
+ Coverage     87.62%   87.67%   +0.05%     
  Complexity     3499     3499              
============================================
  Files           110      110              
  Lines         11642    11641       -1     
  Branches       2403     2403              
============================================
+ Hits          10201    10206       +5     
+ Misses          664      657       -7     
- Partials        777      778       +1     

☔ View full report in Codecov by Harness.
📢 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]

@ericschlarmann how critical is this fix for your codebase? Do you need a release soon that includes it?

[ericschlarmann]

@ericschlarmann how critical is this fix for your codebase? Do you need a release soon that includes it?

Thanks for the quick review and merge!
We cannot upgrade before this is released, so a quick release would of course be appreciated. Until then, we can just coast on the old version, so I wouldn't say super critical ;)