xazax.hun added inline comments.
================
Comment at: clang/lib/StaticAnalyzer/Checkers/MoveChecker.cpp:556
+ bool IsLocal =
+ isa_and_nonnull<VarRegion, CXXLifetimeExtendedObjectRegion>(MR) &&
+ isa<StackSpaceRegion>(MR->getMemorySpace());
----------------
I think strictly speaking this might be "unsound" resulting in false positives
in scenarios like:
```
void f(bool reset) {
static T&& extended = getTemp();
if (reset) {
extended = getTemp();
return;
}
consume(std::move(extended));
f(true);
extended.use();
}
```
In case the call to `f(true)` is not inlined (or in other cases there is some
aliasing the analyzer does not know about), we might not realize that the
object is reset and might falsely report a use after move.
But I think this is probably sufficiently rare that we do not care about those.
================
Comment at: clang/lib/StaticAnalyzer/Checkers/StackAddrEscapeChecker.cpp:101
+ QualType Ty = LER->getValueType().getLocalUnqualifiedType();
+ os << "stack memory associated with temporary object of type '";
+ Ty.print(os, Ctx.getPrintingPolicy());
----------------
Is this a good wording for static lifetime extended temporaries?
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D151325/new/
https://reviews.llvm.org/D151325
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
