2008/9/19, Dan Fandrich wrote:

> That makes the most sense to me. I've always wondered why this HAVE_*
> variable was treated differently.

Ok!, the check is now in CVS. For now, also removed all HAVE_ALARM
logic from setup.h.

2008/9/19, Jamie Lokier wrote:

> If there are known architectures which run configure and where alarm
> shouldn't be used, even though there's an alarm function, then I think
> it would be better to put that in the autoconf test. Something like
> "checking if alarm is usable" or something. That would make the
> autoconf more useful for other projects, and is similar to some other
> autoconf tests.

Yes, it could be done. Depending on the results of a first run of
autobuilds it can further be adjusted. Maybe NetWare builds done with
gcc might need it, or something else.

-- 
-=[Yang]=-