Make ParentMap work with explicit specializations of function templates.

For an explicit specialization, we first build a FunctionDecl, and then
we call SubstDecl() on it to build a second FunctionDecl, which has the
first FunctionDecl as canonical decl.

The address of an explicit specialization of function template used to be the
canonical decl of the FunctionDecl.  This is different from all the other
DeduceTemplateArguments() calls in SemaOverload, and since the canonical decl
isn't visited by ParentMap while the redecl is, it also made ParentMap assert
when computing the parent of a address-of-explicit-specialization-fun-template.

To fix, remove the getCanonicalDecl() call.  No behavior difference for clang,
but it fixes an assert in ParentMap (which is e.g. used by libTooling).

llvm-svn: 260159
This commit is contained in:
Nico Weber 2016-02-08 22:23:09 +00:00
parent 06dacd851f
commit 26911c7733
2 changed files with 7 additions and 1 deletions

View File

@ -10339,7 +10339,6 @@ private:
// Template argument deduction ensures that we have an exact match or
// compatible pointer-to-function arguments that would be adjusted by ICS.
// This function template specicalization works.
Specialization = cast<FunctionDecl>(Specialization->getCanonicalDecl());
assert(S.isSameOrCompatibleFunctionType(
Context.getCanonicalType(Specialization->getType()),
Context.getCanonicalType(TargetFunctionType)));

View File

@ -4333,6 +4333,13 @@ TEST(HasAncestor, NonParmDependentTemplateParmVarDeclRefExpr) {
declRefExpr(to(decl(hasAncestor(decl()))))));
}
TEST(HasAncestor, AddressOfExplicitSpecializationFunction) {
EXPECT_TRUE(matches("template <class T> void f();\n"
"template <> void f<int>();\n"
"void (*get_f())() { return f<int>; }\n",
declRefExpr(to(decl(hasAncestor(decl()))))));
}
TEST(HasParent, MatchesAllParents) {
EXPECT_TRUE(matches(
"template <typename T> struct C { static void f() { 42; } };"