Now to work out how to upload v2 of the patch...
================
Comment at: unittests/Tooling/RecursiveASTVisitorTest.cpp:414-415
@@ +413,4 @@
+ EXPECT_TRUE(Visitor.runOver(
+ "struct WithCtor { WithCtor(); }; \n"
+ "struct Simple { Simple(); WithCtor w; }; \n"
+ "int main() { Simple s; Simple t(s); }\n"));
----------------
Richard Smith wrote:
> A comment explaining the purpose of the declarations of the default
> constructors would be useful. Alternatively, how about passing Simple &s as
> an argument to your function below, or using an "extern Simple s" global?
The latest version of the patch (soon to be uploaded) drops the copy from s to
t and describes the role of the default constructors.
================
Comment at: unittests/Tooling/RecursiveASTVisitorTest.cpp:413-416
@@ +412,6 @@
+ // to use them in order to force them to appear in the AST.
+ EXPECT_TRUE(Visitor.runOver(
+ "struct WithCtor { WithCtor(); }; \n"
+ "struct Simple { Simple(); WithCtor w; }; \n"
+ "int main() { Simple s; Simple t(s); }\n"));
+}
----------------
Manuel Klimek wrote:
> I'd probably quickly add a EXPECT_FALSE for everything but the copy
> constructor, so that we are actually sure the test passes for the right
> reason.
The test passes because the visitor visits the implicit (un-written)
construction of a WithCtor object -- nothing else satisfies the expectations.
I've added another test that a suitable visitors does *not* visit it (because
it's implicit, and that visitor skips implicit code).
http://llvm-reviews.chandlerc.com/D26
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
