Dependencies satisfied, build stops anyway.

[Chris H]

On Wed, 22 Nov 2017 21:12:28 -0800 "bob prohaska" <fbsd at www.zefox.net> said

> On Wed, Nov 22, 2017 at 08:32:53PM -0800, Kevin Oberman wrote:
> > 
> > Stale ports SHOULD not be an issue. Normally old ports are updated
> > automatically. It is important that your ports tree is consistent and,
> > usually that it is current. Assuming yours is, the only issue is the
> > oddball case like this one. You won't hit them often. Usually it is when a
> > file that is tested for is missing. Usually a library. pkgdb says that the
> > port is installed, but the make looks for the actual library and triggers
> > an install, as opposed to a reinstall, when it is not found. Unled this
> > systm has been massively mangled, it is a very unusual development, at
> > least in my experience.
> > --
> 
> Aye, there's the rub 8-)
> 
> Some time ago I accidentally destroyed most of /usr and ended up
> reconstructing it from a recent snapshot. Since buildworld and
> buildkernel seemed to work in /usr/src, I thought /usr/ports would
> be likewise recovered after a fresh checkout.
> 
> For now I've set
> FORCE_PKG_REGISTER=yes 
> in /etc/src.conf and that seems to allow make to proceed without
> interruptions for dns/bind910, but I gather that's not a long-term
> fix.
I think you might have been better off going with the suggested
cd /offending/port
make deinstall
make reinstall

The choice to pollute your environment with FORCE_PKG_REGISTER
will simply compound the problem you're already experiencing. :(
Just saying.

In the end, you're probably going need to decide on a revision,
and check it out for both src && ports. Then perform a fresh
build(world/kernel) && install world/kernel. Else you'll be
fighting  this and other woes for eternity.

I'm only saying all this in an effort to help. :)

All the best, and best of luck. :)

--Chris

> 
> Thanks very much!
> 
> bob prohaska
>