[Common-dev] Re: [Helix-dna-dev] Re: CR: New net API changes
Fredrik Hubinette hubbe at real.com
You can put <umake_includefiles>foo/bar</umake_includefiles> in a BIF
module. If a compilation includes this module, *all* umake invocations
in that build will read foo/bar and related pcf files before reading the
umakefile. (But after reading the .pcf file)
This make it possible to modify compiler settings for an entire build
without making changse to all modules. Since this is a very powerful tool,
it also has a bunch of pitfalls of course, but I think that in almost all
cases we should use this functionality instead of adding anything more to
our existing .cf files.
/Hubbe
Tom Marshall <tmarshall at real.com> writes:
> On Fri, May 14, 2004 at 10:53:58AM -0700, Fredrik Hubinette wrote:
> >
> > Are you sure that adding WIN32_LEAN_AND_MEAN to win32.cf won't break
> > any old builds?
>
> It did break the branch builds. Additionally, the TLC team is expressing
> resistance to using WIN32_LEAN_AND_MEAN in their builds. So our current
> proposed solution is to define it in hlxclib/windows.h just before including
> the system's windows.h.
>
> > If you are not sure, I have an alternate solution that might be better
> > using "global pcf files" instead of modifying the ribosome .cf file.
>
> That sounds interesting. Can you explain more?
>
> --
> Nothing is illegal if a hundred businessmen decide to do it, and that's true
> anywhere in the world.
> -- Andrew Young