[C++ PATCH] [PR18354] Unary plus as template argument

[Giovanni Bajo]

Roger Sayle <roger@eyesopen.com> wrote:

> I believe that the real source of the problem is that INTEGER_CSTs
> should never be wrapped in NON_LVALUE_EXPRs, and indeed this is the
> only one of the five remaining places in GCC that create
> NON_LVALUE_EXPR nodes without checking whether it is actually
> needed or not.

Ah sorry, I had misread this paragraph. Your patch is surely a valid fix, but I
still prefer mine because it unifies the code for unary minus and plus
(avoiding duplication), and also removes one use of NON_LVALUE_EXPR. My
understanding is that we are trying to get rid of it.

Giovanni Bajo