Hi,

There were some patches I circulated on this list in March for some funky
bin-install behavior:

1. real-su-bin-install forgets it's doing a bin-install, so you wind up
   getting a source install when it's not what you asked for

2. tgz's cached in packages/All actually harmful, because they prevent the
   search of BINPKG_SITES for their preqrequisites, causing bin install to
   fail unless all the prerequisites are cached in packages/All too.

3. given DEPENDS+= foo>=1.2nb2:../../bar/foo
   and the bar/foo Makefile provides 1.2nb3
   even if a 1.2nb2 binary exists, which would satisfy the requirement,
   it is ignored and a source build of 1.2nb3 is done.

   The needed information on which package version was actually required
   was already present in PKGNAME_REQD, a variable set but never used within
   bsd.pkg.mk; the fix was as simple as ... using it.

   The patch does not change what happens if you explicitly cd bar/foo and
   run make; you're still assured of building the current 1.2nb3 in that case.

The patches are in http://mail-index.netbsd.org/tech-pkg/2005/03/20/0025.html

More complete explanations and rationale for the changes are in
http://mail-index.netbsd.org/tech-pkg/2005/03/20/0023.html

I think I made a mistake at first using a subject: line that seemed more
niche-related so it may have gone under many people's radar, and starting off
with such complete explanations that it just seemed like too much to read.

So this is just a short note to mention again the patches are there, and IF
there's anything unclear about the rationales, they can be found too in the
earlier message.

Should I make a PR do you think just to keep them visible? I'm curious if
anyone sees effects in the patches that look like they would break something
else.

-Chap