================
@@ -161,6 +161,7 @@ bool CoreEngine::ExecuteWorkList(const StackFrame *SF,
unsigned MaxSteps,
return MaxSteps - Steps;
};
const unsigned STUSteps = ProcessWList(MaxSteps);
+ STUHadWorkRemaining = WList->hasWork();
----------------
marco-milanese-sonarsource wrote:
> Ideally, we would keep the remaining STU work at the end of the CTU work
> list, to continue STU exploration if CTU turns out fruitless. However, that
> would be a significantly more complex and risky change.
I'm not a huge fan of that: if we finish the CTU work we may end up doing STU
work using the CTU budget; this could complicate a lot of things and may not be
very intuitive: imagine extending finding that extending the CTU budget results
in a larger STU exploration, wouldn't it be strange?
> Would it make sense to move it right next to where the STU remaining work
> gets discarded (l.173)?
The problem I see is that in non-CTU mode we would not set the flag, and it may
not be obvious from users that don't know about this detail; they may assume
that it is always set.
https://github.com/llvm/llvm-project/pull/219225
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
