[LLVMdev] TableGen related question for the Hexagon backend

[Sebastian Pop]

On Fri, Aug 3, 2012 at 1:10 PM, Jakob Stoklund Olesen <stoklund at 2pi.dk> wrote:
>
> On Aug 3, 2012, at 10:11 AM, "Villmow, Micah" <Micah.Villmow at amd.com> wrote:
>
>> [Villmow, Micah] Just curious as we have a similar approach for some code in our backed, but if the higher level source and the scripts to produce the intermediates were provided, would that be sufficiently acceptable to using generated td files?
>
> No.
>
> It's perfectly fine if the initial version of the td files are generated by a script, but td files in the LLVM tree need to be the primary source you work on going forward.
>

Ok, I think we are reaching here a middle ground.  I think
that our scripts can detect redundancies and can zip the zillion
classes into multiclass hierarchies that would enable us to move
forward: evolving the code by hand and the result would also be
more readable for humans.

> It think you would be better served by having a way of verifying the correctness of td files using your canonical architecture description. See for example Richard Barton's talk from the April developer meeting.
>

Very good point: we can still use our scripts for verification purposes.
Thanks for your useful remarks and guidance.

Sebastian
-- 
Qualcomm Innovation Center, Inc is a member of Code Aurora Forum