Jakob Stoklund Olesen <jolesen at apple.com> writes:> On Oct 10, 2011, at 7:21 AM, David A. Greene wrote: > >> Bill Wendling <wendling at apple.com> writes: >> >>> As the release enginerd, I would encourage you to hold off any changes >>> to TableGen until after Friday, if they can wait. :-) Of course, Jakob >>> gets the final say on whether they go in or not. >> >> If I can get the for loop done, it would benefit all targets even if >> it's just used for top-level defs. It would be a nice 3.0 feature, I >> think. >> >> What do you think, Jakob? > > Absolutely not.Ok..... Now I'm baffled. Do you object to the for loop in general or just doing it before 3.0? -Dave
On Oct 10, 2011, at 2:15 PM, David A. Greene wrote:> Now I'm baffled. Do you object to the for loop in general or just doing > it before 3.0?In light of your recent embafflement, I think it is best if we take small steps. Please confirm that you have understood the message from Evan and me that we disagree with the general direction of removing redundancy from instruction definitions, and that your patches to that effect will be rejected. /jakob
Jakob Stoklund Olesen <stoklund at 2pi.dk> writes:> I think it is best if we take small steps.Small steps. Yes.> Please confirm that you have understood the message from Evan and me > that we disagree with the general direction of removing redundancy > from instruction definitions, and that your patches to that effect > will be rejected.Yes, I get it. I think I have said that before. I'm talking about the for loop and paste support needed to implement loops as first described by Che-Liang and as worked out by he and I to improve the syntax. This would be used for top-level defs only. For example, in the register definition example Che-Liang gave originally. I will not change anything before the 3.0 branch. But I am working on this functionality with the understanding that it is desired. Please tell me now if we've changed our minds. -Dave