timshen added inline comments.

================
Comment at: clang/include/clang/Driver/Options.td:971-973
+def fexperimental_new_pass_manager_EQ : Joined<["-"], 
"fexperimental-new-pass-manager=">,
+  Group<f_clang_Group>, Flags<[CC1Option]>,
+  HelpText<"Enables an experimental new pass manager in LLVM.">, 
Values<"on,off,debug">;
----------------
chandlerc wrote:
> I don't think we want to expose this flag in the driver -- it should really 
> be a CC1-only thing for debugging.
> 
> As such, I think I'd just make it an independent flag:
> 
>   -fexperimental-new-pass-manager-debug-logging
> 
> Or some such. This also seems easier than having to define a new kind.
The two small disadvantages of your suggestion is that
1) It's more verbose to type: `clang -fexperimental-new-pass-manager 
-fexperimental-new-pass-manager-debug-logging`
2) it's the 3 states vs 4 states pattern: `clang 
-fno-experimental-new-pass-manager 
-fexperimental-new-pass-manager-debug-logging` doesn't make sense to me, and 
it's good to avoid that.

What do you think about these trade offs?


https://reviews.llvm.org/D34790



_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to