Cache ExplicitCallGraph.ExplicitNode.hashCode

[liblit]

The old implementation recomputed getMethod().hashCode() * 8681 + getContext().hashCode() on every call. However, CGNode objects are canonical per (method, context) pair, both fields are final and never change, and equals() is identity-based. Therefore, caching the hash in a final field is safe and avoids repeated (and re-derived) hash computations. Every node is created through findOrCreateNode, which rejects null methods, so eager computation in the constructor cannot NPE.

Effect on JMH benchmarks (JDK 17; 3 forks × 6 one-second iterations per configuration after warm-up; comparisons against master using per-fork means as the unit of analysis, Welch's t-test, one-sided in the improvement direction):

• DataflowBenchmark.analyzeTestContextInsensitive
  • baseline: 0.2934 ms/op (99% CI [0.2792, 0.3077])
  • patched: 0.2158 ms/op (99% CI [0.2027, 0.2288])
  • 26.5% faster, p = 0.005
• DataflowBenchmark.analyzeTestContextSensitive
  • baseline: 0.2233 ms/op (99% CI [0.2145, 0.2321])
  • patched: 0.2136 ms/op (99% CI [0.1993, 0.2278])
  • 4.4% faster, p = 0.046

No other benchmark showed a statistically significant difference in either direction.

A note on multiple comparisons: we ran 19 hypothesis tests per branch, each at the nominal alpha = 0.05 level. Even if this change had no real effect anywhere, uncorrected testing would give roughly a 62% chance of at least one spurious "significant" result (1 - 0.95^19), so raw p-values overstate our confidence. Applying a Holm step-down correction across those 19 tests, the context-insensitive improvement has an adjusted p ("q") of 0.19: above the conventional 0.05 cutoff, though its large ~27% effect size and disjoint 99% confidence intervals leave little practical doubt that it is real. The context-sensitive result does not survive any correction and should be treated as suggestive rather than conclusive.

These gains are consistent with async-profiler data showing ExplicitNode.hashCode as a significant fraction of dataflow-analysis CPU; the same hash chain also feeds LocalPointerKey.hashCode during pointer analysis.

No API change; backward compatible.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 50.81%. Comparing base (3762b68) to head (1b2d5de).

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #2048      +/-   ##
============================================
- Coverage     50.81%   50.81%   -0.01%     
  Complexity    12762    12762              
============================================
  Files          1368     1368              
  Lines         83498    83499       +1     
  Branches      14414    14414              
============================================
- Hits          42432    42431       -1     
- Misses        36382    36383       +1     
- Partials       4684     4685       +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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[msridhar]

Nice!