Re: Considering target file `Makefile'.

[Michael Sterrett -Mr. Bones.-]

On Thu, 8 Jun 2000, Paul D. Smith wrote:

> The way to have make not search for an implicit rule for a target is to
> define an explicit rule for that target.
> 
>   Makefile:;
> 
> There's some info on this in the GNU make manual sections on implicit
> rules.
> 
> HTH.

That sort of helps, but I'm not sure it gets me all the way to a solution.
Perhaps more specific information would be helpful.

We use a common included file called Makedefs.gm.  It's included in most
of our Makefiles at the top.  It would be really nice to be able to say
in Makedefs.gm:

Makedefs.gm:;
Makefile:;

and have all of our Makefiles then have the information gmake needs to
just skip over the implicit rule searches for those files (there are
others as well).

Unfortunatly, this then means that Makedefs.gm can't be included at the top
of the Makefile since these explicit rules then take the place of the first
and default rule which is done when make is invoked without a target on the
command-line instead of the usual "all" target (which is currently first).
Can Makedefs.gm (and other included files) add itself to MAKEFILES and
avoid this issue (since the info page says that the default rule won't
be taken from a file in MAKEFILES)?  I can't seem to make this work,
but maybe I'm doing it wrong.  It seems like it would be nice if the
default rule wouldn't be taken from an "included" Makefile either.

Putting this in Makedefs.gm doesn't seem to do what I was hoping it
would do:

    MAKEFILES += Makedefs.gm

Along this same line - is there a way to get gmake to skip the implicit
rule search for files that look like -lfoo in the dependencies and go
straight to the magic library dependency code?  I don't think that having
explicit and empty rules is what I'd want for that.

BTW, if this all seems pointless and anal...I'm just trying to squeeze
out more speed because we do lots of compiles per day on NFS mounted
disk on a Sequent/ptx machine where the NFS implimentation is horrible.
Skipping the large numbers of stat() calls would be nice.

Michael Sterrett
  -Mr. Bones.-
[EMAIL PROTECTED]