buildbot failure in MacPorts on buildports-snowleopard-x86_64

[Michael Dickens]

True; it is possible. It is my understanding that hackrf is meant to be
small and with minimal dependencies (just 2: pkg-config and libusb), so
that it can be run on a wide variety of host devices. Thus, while it
works on Mac OS X and Windows, it is also meant to run on embedded
devices.  When targeting some of these devices, setting the standard to
use is required; but, I highly doubt it is on any 10.4+ Mac OS X system,
which use pretty-standard GCC or (now) clang.  Hence my reasoning for
just removing it when installing it via MacPorts, which I'll stick with
unless someone complains :) - MLD

On Fri, Jun 7, 2013, at 02:03 AM, Joshua Root wrote:
> It's possible (if hopefully unlikely) there are subtle semantic
> differences that they rely on. Safest would be to use -std=gnu89
> instead, which gcc-4.2 does accept, and in later gcc versions specifies
> the same behaviour as gnu90. (The reason for the synonym is that ISO C90
> is identical to ANSI C89.)
> 
> The flag of course only changes anything when using clang, as it
> defaults to c99 but (llvm-)gcc-4.2 already uses gnu89 by default.