On Wed, Mar 31, 2004 at 03:22:11PM -0800, Jeff Newmiller wrote: > > /* bar.h > > I do things with alcohol (get it? "bar?" hahaha...) */ > > > > #include "bar.h" /* My header */ > > #include "zzz.h" /* Contains some #define's for compile-time options */ > > > > ... > > > > > > Should I make "bar.h" depend on "zzz.h"? Or "foo.o" depend on it? Or...? > > You have bar.h including bar.h... not a good idea. I am not really sure > where your question was leading... assuming we omit the recursion...
Heh, whoops, no ignore that, sorry. I thought I was doing 'bar.c', but yeah... bar.h #include's zzz.h, but NOT bar.h itself. :) No recursion. Sorry for the typo... > * the foo.c dependency should include foo.h, bar.h, and zzz.h > > * any bar.c dependency should include bar.h and zzz.h Okay, so Rod's idea for a Makefile variable seems best, since there will be .h files that include other .h's, and are themselves included by other .c files... <snip> > In general, for a given .c.o dependency, you need to include all of the > quoted includes that it depends on... directly OR indirectly. This is > generally not done for <> includes (library headers). Heh, in the environment I'm coding for, there ARE no "<>" includes. Lame. >:^P -bill! [EMAIL PROTECTED] Have you visited the Linux Users' Group http://newbreedsoftware.com/bill/ of Davis yet!? http://www.lugod.org/ _______________________________________________ vox-tech mailing list [EMAIL PROTECTED] http://lists.lugod.org/mailman/listinfo/vox-tech