[PATCH] D136176: Implement support for option 'fexcess-precision'.

[Zahira Ammarguellat via Phabricator via cfe-commits]

zahiraam added inline comments.


================
Comment at: clang/include/clang/Basic/LangOptions.h:301
 
+  enum Float16ExcessPrecisionKind { FPP_Standard, FPP_Fast, FPP_None };
+
----------------
rjmccall wrote:
> rjmccall wrote:
> > You can leave this named `ExcessPrecisionKind` — if we introduce excess precision for other types, they'll have the same set of options.
> There was a miscommunication here.  Please leave this enum named `ExcessPrecisionKind`, since its values could be applied to any FP type.  However, please set up the language options so that different types are independently controlled.  So there should be a option called `Float16ExcessPrecision` whose type is `ExcessPrecisionKind`. If we add a similar option for `__bf16`, it will be called `BF16ExcessPrecision` and will also have type `ExcessPrecisionKind`, and so on.
Not sure I understand this comment.  The enum is already named ExcessPrecisionKind!
I have added a cc1 option at line 1580 of Options.td called ffloat16-excess-precision whose type is ExcessPrecisionKind .
Am I misunderstanding something?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D136176/new/

https://reviews.llvm.org/D136176