[PATCH] GCC compatibility: pass -z linker options to the linker

[Reid Kleckner]

================
Comment at: include/clang/Basic/DiagnosticDriverKinds.td:122
@@ -121,1 +121,3 @@
 
+def warn_linker_args_take_Wl : Warning<"linker argument -z should be -Wl,-z">,
+  InGroup<InvalidCommandLineArgument>;
----------------
Maybe "linker argument %0 should be escaped with -Wl, and commas"?  Specifically, I think many users won't realize that they need to escape the second space after -z.

Alternatively, I would be happy if we removed this diagnostic, but I know Joerg won't.  My preferred resolution is, again, that we ask GCC to document -z as a linker option and then we support it without equivocation.

================
Comment at: include/clang/Driver/Options.td:312
@@ -311,1 +311,3 @@
   Flags<[DriverOption, CoreOption]>;
+def z_Flag : Separate<["-"], "z">, Flags<[LinkerInput, RenderAsInput]>,
+  HelpText<"Pass -z <arg> to the linker">, MetaVarName<"<arg>">;
----------------
This can be `def z :`.  For -Z we use Z_Flag and Z_Joined because I think they have different semantics.

================
Comment at: lib/Driver/Tools.cpp:207
@@ +206,3 @@
+      // Pass -z prefix for gcc linker compatibility.
+      if (A.getOption().getName().equals(z_Flag)) {
+        D.Diag(diag::warn_linker_args_take_Wl);
----------------
Why do this check this way instead of .matches(options::OPT_z)?

================
Comment at: lib/Driver/Tools.cpp:209
@@ +208,3 @@
+        D.Diag(diag::warn_linker_args_take_Wl);
+        CmdArgs.push_back("-z");
+      }
----------------
Why not `A.claim(); A.render(Args, CmdArgs)`?  Then you can split this whole clause into it's own 'else if' bucket.

================
Comment at: lib/Driver/Tools.cpp:7588
@@ +7587,3 @@
+      // Pass -z prefix for gcc linker compatibility.
+        if (Input.getInputArg().getOption().getName().equals(z_Flag)) {
+          D.Diag(diag::warn_linker_args_take_Wl);
----------------
ditto, why not .matches?

http://reviews.llvm.org/D4393