Re: [PATCH] newlib-stdint.h: Remove 32 bit longs

[Andy Ross <andrew dot j dot ross at intel dot com>]

Hans-Bernhard Bröker wrote:
> Let me boil this down:
>
> That code is _wrong_.  Trying to base any decision on its behaviour
> can only be wrong, too.

Guys, I assure you I understand how the standards work, and tried to
make that clear.  The concern was self-consistency with what newlib is
using internally, and the printf warning was the easiest way to show
that.  But it turns out that the problem is deeper (e.g. the PRI*
macros in newlib only work with longs even when it picks an int for
int32_t using its own header logic) and probably not fixable in gcc as
I thought.

Andy