GkvJwa wrote:
> @GkvJwa I don't understand the added message in your last comment.
> [Previously](https://github.com/llvm/llvm-project/pull/172287#discussion_r2635384617)
> you already identified the call stack which tells us how pushCleanup is
> called. `NormalAndEHCleanup` does really mean it is a `NormalCleanup` +
> `EHCleanup` so yes the code is executed -> thats really how the problematic
> `seh.scope.begin` gets emitted for your case which causes the issue:
> https://godbolt.org/z/hKW6vroeY .
What I'm trying to say is that with SEH enabled, `EmitCXXTemporary` in the
`__try/__except` block should not actually be called.
`EmitCXXTemporary` executes `pushDestroy` to pass `NormalAndEHCleanup`
```
void CodeGenFunction::EmitCXXTemporary(const CXXTemporary *Temporary,
QualType TempType,
Address Ptr) {
pushDestroy(NormalAndEHCleanup, Ptr, TempType, destroyCXXObject,
/*useEHCleanup*/ true);
}
```
However, if we consider handling this within `pushCleanup`, Could it be
misjudged as a normal use of `(pushCleanup + NormalAndEHCleanup)`?
This led me to wonder before if it was possible to make a judgment directly in
this way.
```
void *EHScopeStack::pushCleanup(CleanupKind Kind, size_t Size) {
...
if (CGF->getLangOpts().EHAsynch && IsEHCleanup && !IsLifetimeMarker &&
CGF->getTarget().getCXXABI().isMicrosoft() && CGF->getInvokeDest()) {
if (IsNormalCleanup) {
CGF->CGM.getDiags().Report(CGF->CurFuncDecl->getLocation(),
diag::err_seh_expected_handler);
}
}
CGF->EmitSehCppScopeBegin();
}
```
I may communicate via code in the future (you can do that too), as there might
be some misunderstandings in the text.
https://github.com/llvm/llvm-project/pull/172287
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
