Re: mips option parsing behaviour changed accidentally?

[cgd at broadcom dot com]

> > I note that your change on 2001-10-03 to tc-mips.c (rev 1.78 of that
> > file) seems to have had the effect of making the handling in
> > md_parse_option for:
> > 
> > 	OPTION_GP32
> > 	OPTION_GP64
> > 	OPTION_FP32
> > 	OPTION_MABI
> > 
> > ELF-specific.  These options used to be supported on non-ELF systems.
> 
> I doubt this for the -mabi option because it handles ELF ABI's.

So, I did a 'cvs update -D"2001-01-01"', and OPTION_MABI was most
definitely _not_ conditionalized on being ELF-only.  (That's also
evidenced by its location in the option list.)

The setting of the ELF header flags was obviously conditionalized on
ELF, and that appears to be that option's only effect, but the actual
-mabi option itself (i.e., the parsing/acceptance of it) was not
ELF-only.

Going forward, I don't know if that's The Right Thing or not, since I
don't know what the compiler has historically used.

If it _is_ going to become an ELF-only option, though, I believe that
it should be folded in with the rest of the options under
OPTION_ELF_BASE...  (yay, lots of renumbering! 8-)

Also, the tests that use -mabi should be explicitly excluded for ECOFF
if they're not already.



cgd