[PATCH] D60974: Clang IFSO driver action.

Saleem Abdulrasool via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Tue May 7 09:13:02 PDT 2019


compnerd added inline comments.


================
Comment at: clang/include/clang/Driver/Options.td:626
   HelpText<"Use the LLVM representation for assembler and object files">;
+def emit_ifso : Flag<["-"], "emit-interface-stubs">, Flags<[CC1Option]>, Group<Action_Group>,
+  HelpText<"Generate Inteface Stub Files.">;
----------------
plotfi wrote:
> compnerd wrote:
> > I thought that we were going to add `experimental` to this for the time being?
> Oh, I was specifying experimental in the -interface-stubs-version=experimental-ifo-elf-v1. Do you want the main flag to also be -emit-interface-stubs-experimental??
Okay, I can live with that.  I guess that was just unclear to me.  Doing that is nicer in that there is less churn in the actual flags.


================
Comment at: clang/include/clang/Driver/Types.def:91
 TYPE("ast",                      AST,          INVALID,         "ast",   "u")
+TYPE("ifs",                      IFS,          INVALID,         "ifs",   "u")
 TYPE("pcm",                      ModuleFile,   INVALID,         "pcm",   "u")
----------------
plotfi wrote:
> compnerd wrote:
> > What about `ifo` instead of `ifs`?
> I went with ifs because ifo implies the analog of a .o file and ifso implies the analog of a .so file. I want the tbe/ifs text files to just be thought of as something a little different. Like symbol listings that another tool can assemble into whatever format.  
Oh, I see.  Yeah .... torn on that tbh.  There are aspects of it as being analogs to `.o` or `.obj` as well.  But, having them be confused for one another is worse.  If this makes sense to you and @rupprecht, WFM.


================
Comment at: clang/lib/Frontend/FrontendActions.cpp:297
+        sema.LateTemplateParser(sema.OpaqueParser, LPT);
+        HandleNamedDecl(FD, Symbols, (FromTU | IsLate));
+      }
----------------
plotfi wrote:
> compnerd wrote:
> > Typo of `||`?  The field is boolean not a mask.
> It is a bitmask. 
Oh!


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D60974/new/

https://reviews.llvm.org/D60974





More information about the llvm-commits mailing list