Thus spake Frederick Bruckman ("FB> ") sometime Today...

FB> "-O" uses an older yet only slightly less obscure feature of
FB> "make", which is also described in the make(1) man page, yet even
FB> more tersely. That is, that objects will automatically be placed
FB> into a subdirectory named ${MAKEOBJDIR} or "obj" if it exists. The
FB> "make obj" phase creates the object tree in the designated directory,
FB> similarly to "-M" but with shorter paths, if possible, and also makes
FB> a single symlink into the corresponding place in the object tree in
FB> each source directory. If it's not possible to write to the object
FB> tree, (because it doesn't exist, say) "make obj" will simply create a
FB> directory directly in the source tree, rather than a symlink. I'd like
FB> to bet that that is what actually happenned in your case, and that the
FB> objects are all in sub-directories named "obj" or "obj.<arch>". The
FB> actual terminal value of the symlink or directory, that is, the value
FB> of ${MAKOBJDIR}, can be affected by other options, so I can't say
FB> exactly without more information. Whatever ${MAKOBJDIR} works out to
FB> be, there will either be a symlink or directory of that name in each
FB> source directory.
FB>
FB> If that's not clear, just try them both, and you'll see what they do.
FB> ;-)

To cast a long story into a
	(union { unsigned short story; different_t tack; } *)...

	-M works with automounted src directories, even though the paths
	   are insanely long.

	-O does not, and eventually punts on symlinks and starts putting
	   stuff under /usr/src directly.

FB> Frederick

				--*greywolf;
--
NetBSD: Stop, Drop, and Load