On Tue, Jun 18, 2002 at 08:10:56PM +0900, itojun@iijlab.net wrote:

 > 	(1)
 > 	why do we have gnu/dist/toolchain/bfd/peigen.c in our repository?
 > 	it can be generated at runtime at ease, and there's no real
 > 	build tool issue (it can be generated by simple sed replacement).
 > 	in fact, it is generated on-demand during src/tools/toolchain build.
 > 	if noone objects, i would like to propose to remove it.

It's there because binutils releases ship with the generated file already
there.

 > 	(3)
 > 	there seem to be pointer-casting bug (or portability issue) in
 > 	GNU toolchain.  they try to cast void * (or whatever pointer) to
 > 	unsigned long long, under certain configuration, and it fails to
 > 	compile.
 > 		void *p;
 > 		unsigned long long x;
 > 		x = (unsigned long long)p;
 > 	i've modified my local tree as below, basically doing:
 > 		x = (unsigned long long)(unsigned long)p;
 > 	is it safe to commit this?  if so, i would like to.

Err, are you building a cross-debugger?  If so, infptrace.c should
not be compiled.  The PTRACE_ARG3_TYPE is controlled by the "host",
which is always the "target" for a native GDB.

Also, kcore-nbsd.c has a similar problem ... it's only designed for
native build only.

-- 
        -- Jason R. Thorpe <thorpej@wasabisystems.com>