Re: [local-bind] local-bind: Support binding to local interface/IPs.

On Fri, 16 Jul 2010, Ben Greear wrote:

> I did a grep for AF_INET6 and it's used many places already, so maybe that's
> not a big deal?

Ah right, I forgot we have this snippet in ares_setup.h:

#if !defined(HAVE_AF_INET6)
#if defined(HAVE_PF_INET6)
#define AF_INET6 PF_INET6
#else
#define AF_INET6 AF_MAX+1
#endif
#endif

So using that define unconditionally is indeed fine.

> In general, the style of c-ares seems quite random. I assume you want it to
> be like curl?

I think so. I think I'll go over it at some point and unify the style
somewhat... I think I originally intended to not change it where it wasn't
needed to get changed so that we would maintain similarities with the original
ares as far as posssible, but as time passes and we've done so many changes I
think we can now officially stop doing that and instead make the code more
readable.

> Do you know any git-fu to make it show where the trailing whitespace is for
> changes already committed (ie, in my tree)?

I saw them with 'git diff' when having applied the patch locally, and I tend
to use emacs' show-trailing-whitespace feature for that (combined with my
code-cleanup function that removes them for me).

And 'git apply' generally complains about them (controlled by the
core.whitespace configuration variable).

More than so, I don't know...

> Well, for what it's worth, we do system-test this stuff, and I plan to stay
> current with c-ares (and curl).

Great!

-- 
  / daniel.haxx.se