PATCH: driver/26885: [4.1/4.2 regression] -m64 -m32 no longer creates 32-bit object

H. J. Lu hjl@lucon.org
Sun Apr 23 01:12:00 GMT 2006


On Sun, Apr 23, 2006 at 12:44:29AM +0000, Joseph S. Myers wrote:
> On Sat, 22 Apr 2006, H. J. Lu wrote:
> 
> > The gcc driver has a generic problem that is if an option is canceled
> > by another one, gcc stills passes everything down to other programs.
> > When the order of all options are passed down unmodified, it isn't
> > a problem. But sometimes spec processing gets into our way, in which
> > case, the order of options processed by spec is diffeerent from the
> > command line. This patch implements Negative(XXX) to cancel -XXX in
> > the gcc driver, as suggested in PR 26885.
> 
> You need to update options.texi to explain the specification of what 
> Negative means in a .opt file.  If the way you seem to be chaining -g 

Sure.

> options in common.opt is intended to mean that each such option overrides 
> any of the others, that needs to be explained in the documentation.

Are you saying -gdwarf-2 doesn't override the previous -gstabs
and -gstabs+ currently?


H.J.



More information about the Gcc-patches mailing list