On 1/26/06, David Lee <[EMAIL PROTECTED]> wrote:
> On Thu, 26 Jan 2006, Alan Robertson wrote:
>
> > Andrew Beekhof wrote:
> > > [...]
> > > If its Solaris specific, I say do what you gotta do (but perhaps
> > > #ifdef it out for everyone else).
> >
> >
> > I disagree.
> >
> > ifdefs just clutter the code, and the string default won't appear except
> > if it means 0.0.0.0.  So, I'd say go ahead without the ifdefs.  Putting
> > comments explaining that this is needed for Solaris is fine though.
>
> Well I committed it a couple of hours ago.  It is mostly comment; the code
> is small.
>
> I had actually used a "#ifdef" around the code, but the relevant "#define"
> was set near the top, itself inside an "#if 1" (so always enabled)
> accompanied by a comment about restricting it that were ever to prove
> necessary.
>
> So I've ended up (before Alan's email!) with Alan's "always enable" and
> "comment" features, and Andrew's "have an ifdef available" feature.
>
> Do a diff between revs. 1.50 and 1.51 if you want to see it.  A reasonably
> decent solution, I think.
>

I'm happy to go along with Alan's no-ifdef rationale.
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/

Reply via email to