This is looking a lot better.
In particular, I feel like the testing needs a bit more work; I'm not seeing
the connection between the added tests and the functional changes in this patch.
I'd like espindola to give a second opinion on the code, and he can probably
suggest improvements to the testing.
================
Comment at: include/clang/Driver/Job.h:103
@@ +102,3 @@
+ bool useResponseFile() const {
+ return ResponseFile != nullptr && NeedsResponseFile;
+ }
----------------
Is there ever any situation where NeedsResponseFile == true but ResponseFile !=
nullptr? In general, what is the meaning of the 4 different possibilities for
(ResponseFile {=,!}= nullptr) x (NeedsResponseFile {=,!}= true)
http://reviews.llvm.org/D4897
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
