🚀 go-pugleaf

Last post, I had problems with the '-c68020' option problem. Got that fixed;
however, I've come upon a new problem and some 'bugs' with the make process:

1) put the m68k binutils (as, ld, etc) in the build directory. This wasn't
	in the INSTALL file. (Probably isn't necessary for every compile; but it
	probably doesn't hurt?) (this fixed the option problem)

2) call to ar uses /usr/m68k-vxworks, not /usr/local/m68k-vxworks (std location
	where it says it looks for things; might have been the same path with
	as option problem?).

3) when compiling in objc dir, had a problem with BOOL definitions being
	mult. typedef'ed. BOOL is def'ed in <types/vxTypesOld.h>, and unfortunately,
	is not ifdef'ed properly (if !defined (YES) || !defined (TRUE), but no
	def'ing of YES or TRUE w/in the if to properly ifdef the objc/objc.h BOOL
	def'n). Also the objc/objc.h is unsigned char whereas the vxWorks BOOL is
	just int.

4) Have a similar problem with wchar_t, ptrdiff_t, and size_t in 
	include/stddef.h and <types/vxTyeps.h>.

Of Note: I believe I'm using vxWorks 5.2 include files w/ gcc 2.7.1

And, of course, there's yet another challange:
Things are going along ok, until the libgcc2.a build. Oh, it makes it fine,
and then decides that it needs to make it again. And again. And again. And...
get the picture?

Doing a make -d it seems to indicate that xgcc is 'newer' than libgcc2.ready
(which it 'obviously' isn't!) and runs thru the build again.  Tried to trace
the problem in the Makefile, but couldn't seem to track it down.

Thanks in advance,
K'vin
--
K'vin De Vries
  Deep Freeze Consulting
** Really Cool Software **