[PATCH] Enhancements for the DynTypedMatcher system.

Samuel Benzaquen sbenza at google.com
Thu Jun 20 07:14:18 PDT 2013



================
Comment at: include/clang/ASTMatchers/ASTMatchersInternal.h:188
@@ -187,1 +187,3 @@
 
+template <typename T> class Matcher;
+
----------------
Manuel Klimek wrote:
> Is this still needed?
Removed.

================
Comment at: include/clang/ASTMatchers/ASTMatchersInternal.h:257
@@ -248,1 +256,3 @@
 
+  /// \brief Returns \c true if the pass DynTypedMatcher can be converted to \c
+  ///   Matcher<T>.
----------------
Manuel Klimek wrote:
> s/pass/passed/
Done

================
Comment at: include/clang/ASTMatchers/ASTMatchersInternal.h:258
@@ +257,3 @@
+  /// \brief Returns \c true if the pass DynTypedMatcher can be converted to \c
+  ///   Matcher<T>.
+  ///
----------------
Manuel Klimek wrote:
> and I'd probably say to "a" Matcher<T>
Done

================
Comment at: include/clang/ASTMatchers/ASTMatchersInternal.h:267
@@ +266,3 @@
+
+  /// \brief Construct a Matcher<T> interface around the dynamic \c Other.
+  ///
----------------
Manuel Klimek wrote:
> dynamic "matcher" \c Other.
Done

================
Comment at: include/clang/ASTMatchers/ASTMatchersInternal.h:274
@@ +273,3 @@
+    assert(canConstructFrom(Other));
+    return Matcher<T>(new WrappedMatcher(Other));
+  }
----------------
Manuel Klimek wrote:
> I'm wondering whether there's a simpler way - but probably not, as we don't have access to the internals of the "real matcher" behind the DynTypedMatcher 'Other'.
You don't have source and destination types together in the same function, so you have to go through the generic interface to stitch them together.


http://llvm-reviews.chandlerc.com/D815

BRANCH
  dyn_type_check

ARCANIST PROJECT
  clang



More information about the cfe-commits mailing list