[PATCH] D14265: DI: Reverse direction of subprogram -> function edge.

[David Blaikie via llvm-commits]

dblaikie added a subscriber: dblaikie.
dblaikie added a comment.

Generally looks good to me - one or two minor comments.


================
Comment at: lib/Bitcode/Writer/BitcodeWriter.cpp:1027
@@ -1026,3 +1026,3 @@
   Record.push_back(N->isOptimized());
-  Record.push_back(VE.getMetadataOrNullID(N->getRawFunction()));
+  Record.push_back(0);
   Record.push_back(VE.getMetadataOrNullID(N->getTemplateParams().get()));
----------------
Could we use the size of the record to determine whether we're in the old format or the new one? Rather than including an empty field?

================
Comment at: lib/IR/DebugInfoMetadata.cpp:364
@@ -364,8 +363,3 @@
 
-Function *DISubprogram::getFunction() const {
-  // FIXME: Should this be looking through bitcasts?
-  return dyn_cast_or_null<Function>(getFunctionConstant());
-}
-
 bool DISubprogram::describes(const Function *F) const {
   assert(F && "Invalid function");
----------------
I wonder if this function is even needed anymore... hopefully we'll just not encounter DISubprograms for a Function that aren't attached to the same Function.


http://reviews.llvm.org/D14265