On Sat, 20 Feb 1999, Scott Reynolds wrote:

> By examining the history of the build target....

That's a lot more work than just reading a paragraph that describe
what build is intended to do. And most developers don't do it.

>  There are others, however, that do daily and weekly builds.
> It would be Really Bad if a daily build took longer than a day. :-)

Well, then it's really bad on a fair number of architectures already.
A daily build can be done in under a day on a Sparc 5 if you don't
get any compile errors, but I don't think a Sparc 1 will do it.
Nothing in the Sun 3 line will build a snapshot in 24 hours or
less. The same is true of the VAX line.  It may well be true of
others that I'm not familar with.

> Interesting, should work, but how about something that's far simpler to
> implement and doesn't require a tone of infrastructure?  Create an
> `upgrade' target that builds the prerequisites, then does a `${MAKE}
> build' as the last step.  (This target would probably not do anything for
> a non-root DESTDIR.)

Sounds like a good idea to me. Heck, I'd even go for a target that
would ignore DESTDIR for building the toolchain, but use it for
rebuilding the world after the new toolchain is installed.

cjs
--
Curt Sampson  <cjs@cynic.net>   604 801 5335   De gustibus, aut bene aut nihil.
The most widely ported operating system in the world: http://www.netbsd.org