danix800 wrote:
> > Another PR would fix the crash when this one is landed.
>
> Can you land the fix for the crash first? If the other PR gets reverted it
> creates a lot of churn, so it would preferable to land that first. Assuming
> it is not somehow dependent on this fix.
This new crash is a false alarm caused by the first incorrect fix. Sorry about
that!
Could you also take a look at the new fix for the first crash?
Previously we think that lambda without trailing return couldn't
have return type defined inside its body, but for
```c
[] { // no parentheses for parameters
struct X {};
return X();
};
```
the proto has trailing return deduced to the type defined inside.
So I think we should remove the trailing return test for lambdas.
https://github.com/llvm/llvm-project/pull/101031
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
