On 03/06/2012 11:28 AM, Eric Blake wrote:
On 03/06/2012 09:15 AM, Duncan Rance wrote:
> Hi,
>
> I'm building on OSX with no libnl. I had to do this to get src/util/virnetlink.c
to compile:
>
>
>
> diff --git a/src/util/virnetlink.c b/src/util/virnetlink.c
> index 1575bad..59f3e39 100644
> --- a/src/util/virnetlink.c
> +++ b/src/util/virnetlink.c
> @@ -545,9 +545,9 @@ int virNetlinkCommand(struct nl_msg *nl_msg ATTRIBUTE_UNUSED,
> */
> int virNetlinkEventServiceStop(void)
> {
> +# if defined(__linux__) && !defined(HAVE_LIBNL)
> netlinkError(VIR_ERR_INTERNAL_ERROR,
> "%s",
> -# if defined(__linux__) && !defined(HAVE_LIBNL)
> _("virNetlinkEventServiceStop is not supported since libnl was
not available"));
Oops - that's a blatant bug. ACK and pushed. I've also added you to
AUTHORS; let me know if you prefer an alternate spelling.
That fix is actually a bit off - it ends up emitting no error log on
failure for non-linux systems. (Of course these functions should never
even be called on non-linux systems, so the difference is mostly
academic :-)
I made a more comprehensive patch for all the stub functions in
virnetlink.c that collapses all of the nearly-identical log messages and
eliminates the use of preprocessor directives within macro invocations
for those functions:
https://www.redhat.com/archives/libvir-list/2012-March/msg00316.html