>I think the effort is better spent encouraging bind-4.9.3 to be
>incorporated in current.

>Paul Vixie really seems to want to release 4.9.3B24.

Every time I think about this, I think that the resolver
code  should  be in its *own* shared library, not in libc.
Is there some good reason to put the resolver in libc ?

I know the designs of NetBSD dynamically-linked shared libraries and
SunOS shared libraries have subtle and important differences (no Silly
Archives?); but repeatedly reading the documentation in bind-4.9.2 and
4.9.3-beta{3 through 24} leaves me wth a knee-jerk response that there
are good reasonsto NOT merge resolver code into the libc shared
library.

Changing this would of course mean changing the Makefiles of any
binaries that link against resolver code in libc.  Is there any additional
downside?