[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: CVS commit: src/usr.bin/make
David Holland <dholland-sourcechanges%netbsd.org@localhost> writes:
> On Tue, Mar 24, 2009 at 09:53:13AM -0400, Perry E. Metzger wrote:
> > > However, I think you should put back the change to main.c that sets
> > > MAKE_VERSION if passed on the command line; I suspect either pkgsrc
> > > and/or sjg's portable release probably use it, and probably if we
> > > aren't going to put a dynamic date into it from the makefile, we
> > > should put a static date into it.
> > Okay, I'll put that part of the change back. It causes no actual harm
> > to leave it.
> > > Maybe what we want to set it to is something like
> > > [snip]
> > That sort of date seems unobjectionable since it reflects the last
> > change made to the sources rather than the last compile date. However,
> > I'm not sure why one actively wants it given that you have all the
> > rcsid's already in the binary. Also, why would one want it more in
> > "make" than in any other program?
> Well, remember make is an interpreter; the date goes into the make
> variable $(MAKE_VERSION) so the information is available to Makefiles.
Yah, but it is largely worthless. Consider that different branches
have non-linear relationships in dates and features being present.
> I don't know to what extent that variable is used, standardized, or
> historically expected, but given that Unix make has a large body of
> lightly documented "standard" and historical behavior, it's usually
> best not to remove anything like that unless one is quite sure it's
> *not* needed.
It isn't used in our source tree, that's for sure, and if anyone is
depending on it, they probably are going to get into trouble because
(again) the relationship between dates and features isn't consistent
If there is a desire to have a version that can be used for feature
testing, it should probably be hand set.
Perry E. Metzger perry%piermont.com@localhost
Main Index |
Thread Index |