On Fri, Aug 10, 2007 at 08:11:23PM -0400, Greg Troxel wrote:
> So, with make replace, and I'd argue pkg_add -u, we'd set unsafe_depends
> if the above rule is violated, and otherwise not.
> 
> Does that seem ok to you?

I'm not even at the point of discussion changes in the scheme. Like I
said earlier in that thread -- we have too many issues to be covered
first before something can be implemented to fix this, without turning
pkg_install into an even great mess. Up to that point, I want to reduce
the number of pieces that mess with pkgdb as far as I can. I know that
make replace has to do some magic to get its functionality, but it
should be as small as possible.

I'd actually consider to
(a) Just recompute the dependency graph with pkg_admin rebuild-tree
and
(b) Use the normal pkg_admin interface to store the information.

While this is slower than the current approach, it is certainly a lot
less logic exposed. Only half-serious here :-)

Joerg