[LLVMdev] Can insertvalue ignore its last operand?

[Duncan Sands]

Hi Jianzhou,

> Some of my code is compiled to
>
>    %insert2 = insertvalue { float, float } %insert, float %33
>
> When the index value (the last operand) is omitted, does it mean it
> uses some default value, say 0?
> The manual does not allow this syntax though.
>    http://llvm.org/docs/LangRef.html#i_insertvalue

I don't think this is valid.  Does it pass the verifier?  If so, please
open a bug report with a complete testcase.

Thanks, Duncan.