ERROR: Could not resolve variable c (/home/jenkins/workspace/Language-Tests/Java/job/ql/java/ql/src/semmle/code/java/dataflow/internal/DataFlowImpl.qll:2270,54-55)

That looks like a typo: c should be cc.

That looks like a typo: c should be cc.

Thanks.

Here are the profiling runs (all internal links):

• Java: https://jenkins.internal.semmle.com/job/Changes/job/Java-Differences/580/
• C++: https://jenkins.internal.semmle.com/job/Changes/job/CPP-Differences/836/
• C#: https://jenkins.internal.semmle.com/job/Changes/job/CSharp-Differences/21/

Performance for Java is unchanged, there is a massive improvement for C++, and C# is still running (performance expected to be unchanged) also unchanged.

Joining localFlowBigStep with getLocalCallContext creates a potential blowup - in ArithmeticTainted on jdk11 the blowup is more than a factor 10, but it doesn't show up as the cost ends up dominated by other predicates. It is, however, not the best join-order, so I've put up an alternative fix here: #2872