This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [patch] Fix PR middle-end/18029 and PR middle-end/18030 (4.1material)


Hi Roger,

> > 2004-10-16  Kazu Hirata  <kazu@cs.umass.edu>
> >
> > 	PR middle-end/18029 and PR middle-end/18030
> > 	* expr.c (expand_assignment): Add a special case to handle
> > 	BIT_IOR_EXPR and BIT_XOR_EXPR.
> 
> This patch is OK for 4.1 (though I gave serious thought to approving
> it for mainline).

Thanks!  To my surprise, this triggered only once in GCC, real.c.

> I also agree that we should factor either expand_component_assignment
> or expand_bitfield_assignment out of expand_assignment, and that this
> will be easier after your proposed clean-up to avoid expand_assignment
> returning a rtx.  Depending upon how invasive/safe these clean-ups are,
> they may also need to wait until 4.1.

I can clean up expand_assignment in a series of small patches.  For
example, I could remove all the requests for expand_assignment to
return an rtx (as I have proposed), and I could remove
expand_assignment's ability to return an rtx in a separate patch.

Kazu Hirata


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]