Source-Changes archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: CVS commit: basesrc



>
>On Tue, 23 Oct 2001, James Chacon wrote:
>
>: Neither distrib or etc should be descended to on a build unless explictly
>: asked to. Now your assuming that anytime a rules change in bsd.*.mk which
>: may change internal rule names in some way will be propoaged to etc and
>: distrib.  I guess whomever is doing that will "remember" to come to
>: etc/distrib and fix it here when that happens.
>
>The Makefiles should pull in one of the standard <bsd.*.mk> files that
>define all of the ${TARGETS} -- <bsd.subdir.mk> is adequate, which the sparc
>distrib was using when I fixed it, but <bsd.prog.mk> with NOPROG= set is
>better since you automagically get objdir support and etc.
>
>If this policy is followed, there is no need to "fix" it again in the
>future, since you get all the targets defined automatically.  Fixing these
>now will keep these Makefiles more standardized, and will keep them working
>in the future.

Keep them more standardized???

etc/Makefile currently has:

all clean cleandir depend distclean etc includes install lint:


in it as a null rule. Until it's fixed, leaving it and distrib off the main
subdir list is safer...You don't end up with accidents then..

You don't want to recurse into distrib unless etc is done and even then the
only time would be on a make release.

I can agree this should be better laid out long term, but opening everyone
who builds trees right now up to these pitfalls rather than doing it after
those Makefiles are fixed makes no sense.

You want to leave something in which arguablly does nothing today, opens up
potential problems, etc in order to have it in place ahead of time for a 
future change? Just put it in place then, when the changes are all ready...

James



Home | Main Index | Thread Index | Old Index