PATCH: Consistent use of the noduplicate attribute
eliben at google.com
Mon Mar 17 09:26:18 PDT 2014
Thanks for the review - committed in r204049.
James - I revived the setter per your request (and put it in the invoke
too, for consistency with the call).
On Mon, Mar 17, 2014 at 3:21 AM, James Molloy <james at jamesmolloy.co.uk>wrote:
> Hi Eli,
> Sorry I didn't get to this very quickly. The patch in general looks fine
> but I don't agree with removing setCannotDuplicate(). Many users of
> noduplicate (for example OpenCL stacks) are out of tree, and I don't think
> it's right to make an interface unorthogonal (getter but no setter) when
> it's so simple just because there are no users in tree.
> I'd appreciate if that could be added back in!
> On 15 March 2014 22:12, Duncan P. N. Exon Smith <dexonsmith at apple.com>wrote:
>> On 2014 Mar 14, at 16:04, Eli Bendersky <eliben at google.com> wrote:
>> > Hello,
>> > The "noduplicate" attribute of call instructions is sometimes queried
>> directly and sometimes through the cannotDuplicate() predicate. This patch
>> streamlines all queries to use the cannotDuplicate() predicate. It also
>> adds this predicate to InvokeInst, to mirror what CallInst has.
>> > In addition, setCannotDuplicate() is removed because it's not being
>> used anywhere in the codebase.
>> > PTAL,
>> > Eli
>> > llvm-commits mailing list
>> > llvm-commits at cs.uiuc.edu
>> > http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>> llvm-commits mailing list
>> llvm-commits at cs.uiuc.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-commits