Re: PATCH: libffi vs. SPARC (again)

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Jeff" == Jeff Sturm <jsturm@one-point.com> writes:

Tom> In other words, suppose we change libffi to give an error if the
Tom> return type is FFI_TYPE_SINT8.  Once we've done that, why not
Tom> simply fix libffi to handle this piece of the ABI?

Jeff> ...except I fear it wouldn't be simple.  For one thing, any
Jeff> applications that depend on the current behavior could begin to
Jeff> fail.  (That includes ffitest.c, BTW.  Look at line 332 where a
Jeff> ffi_type_schar return type is given with a (int *) return
Jeff> value.)

Jeff> For another thing, changing libffi properly could not be done in
Jeff> any of the machine-independent sources, so you'd touch every
Jeff> port.

Ok.

Jeff> <sigh> I've already spent more time on this than I could afford
Jeff> right now.  I've been patching my local solaris builds for the
Jeff> last six months.  Nobody else has fixed it yet.  If nothing
Jeff> breaks, can we please go with this, at least for 3.1?

I'm inclined to agree.  Unless Bryce or Anthony object in the next day
or so, please check in your patch.

Could you make a quick patch for the libffi documentation explaining
this point?  If not tell me and I'll try to do it.

Tom