[lldb] Fix FindDirectNestedType not working with class templates

[Endilll]

This patch attempts to fix lookup in class template specialization.

The first fixed problem is that during type lookup DeclContextGetName have been dropping template arguments. So when such a name was compared against a name in DW_AT_name, which contains template arguments, false mismatches have been occurring.

The second fixed problem is that LLDB's printing policy hasn't been matching Clang's printing policy when it comes to integral non-type template arguments. This again caused some false mismatches during type lookup, because Clang puts e.g. 3U in debug info for class specializations, but LLDB has been expecting just 3. This patch brings printing policy in line with what Clang does.

[llvmbot]

@llvm/pr-subscribers-lldb

Author: Vlad Serebrennikov (Endilll)

Changes

This patch attempts to fix lookup in class template specialization that have an integral non-type template parameter of non-int type. unsigned 3 might be printed as 3 or 3U depending on PrintingPolicy. This patch bring it in line with what Clang does (which addes the suffix).

---

Full diff: https://github.com/llvm/llvm-project/pull/81666.diff

3 Files Affected:

• (modified) lldb/source/Plugins/TypeSystem/Clang/TypeSystemClang.cpp (+7-2)
• (modified) lldb/test/API/python_api/type/TestTypeList.py (+17)
• (modified) lldb/test/API/python_api/type/main.cpp (+13)

diff --git a/lldb/source/Plugins/TypeSystem/Clang/TypeSystemClang.cpp b/lldb/source/Plugins/TypeSystem/Clang/TypeSystemClang.cpp
index a41e2c690853d2..e6a9d3f4f02836 100644
--- a/lldb/source/Plugins/TypeSystem/Clang/TypeSystemClang.cpp
+++ b/lldb/source/Plugins/TypeSystem/Clang/TypeSystemClang.cpp
@@ -2165,6 +2165,7 @@ PrintingPolicy TypeSystemClang::GetTypePrintingPolicy() {
   // (and we then would have suppressed them from the type name) and also setups
   // where LLDB wasn't able to reconstruct the default arguments.
   printing_policy.SuppressDefaultTemplateArgs = false;
+  printing_policy.AlwaysIncludeTypeForTemplateArgument = true;
   return printing_policy;
 }
 
@@ -9265,8 +9266,12 @@ ConstString TypeSystemClang::DeclContextGetName(void *opaque_decl_ctx) {
   if (opaque_decl_ctx) {
     clang::NamedDecl *named_decl =
         llvm::dyn_cast<clang::NamedDecl>((clang::DeclContext *)opaque_decl_ctx);
-    if (named_decl)
-      return ConstString(named_decl->getName());
+    if (named_decl) {
+      std::string name;
+      llvm::raw_string_ostream stream{name};
+      named_decl->getNameForDiagnostic(stream, GetTypePrintingPolicy(), /*qualified=*/ false);
+      return ConstString(name);
+    }
   }
   return ConstString();
 }
diff --git a/lldb/test/API/python_api/type/TestTypeList.py b/lldb/test/API/python_api/type/TestTypeList.py
index c267defb58edf9..f874e87771c624 100644
--- a/lldb/test/API/python_api/type/TestTypeList.py
+++ b/lldb/test/API/python_api/type/TestTypeList.py
@@ -150,6 +150,23 @@ def test(self):
         invalid_type = task_type.FindDirectNestedType(None)
         self.assertFalse(invalid_type)
 
+        # Check that FindDirectNestedType works with types from AST
+        pointer = frame0.FindVariable("pointer")
+        pointer_type = pointer.GetType()
+        self.assertTrue(pointer_type)
+        self.DebugSBType(pointer_type)
+        pointer_info_type = pointer_type.template_args[0]
+        self.assertTrue(pointer_info_type)
+        self.DebugSBType(pointer_info_type)
+
+        pointer_masks1_type = pointer_info_type.FindDirectNestedType("Masks1")
+        self.assertTrue(pointer_masks1_type)
+        self.DebugSBType(pointer_masks1_type)
+
+        pointer_masks2_type = pointer_info_type.FindDirectNestedType("Masks2")
+        self.assertTrue(pointer_masks2_type)
+        self.DebugSBType(pointer_masks2_type)
+
         # We'll now get the child member 'id' from 'task_head'.
         id = task_head.GetChildMemberWithName("id")
         self.DebugSBValue(id)
diff --git a/lldb/test/API/python_api/type/main.cpp b/lldb/test/API/python_api/type/main.cpp
index 98de9707d88654..b587acdd96c590 100644
--- a/lldb/test/API/python_api/type/main.cpp
+++ b/lldb/test/API/python_api/type/main.cpp
@@ -34,6 +34,15 @@ class Task {
     {}
 };
 
+template <unsigned Value>
+struct PointerInfo {
+  enum Masks1 { pointer_mask };
+  enum class Masks2 { pointer_mask };
+};
+
+template <unsigned Value, typename InfoType = PointerInfo<Value>>
+struct Pointer {};
+
 enum EnumType {};
 enum class ScopedEnumType {};
 enum class EnumUChar : unsigned char {};
@@ -71,5 +80,9 @@ int main (int argc, char const *argv[])
     ScopedEnumType scoped_enum_type;
     EnumUChar scoped_enum_type_uchar;
 
+    Pointer<3> pointer;
+    PointerInfo<3>::Masks1 mask1 = PointerInfo<3>::Masks1::pointer_mask;
+    PointerInfo<3>::Masks2 mask2 = PointerInfo<3>::Masks2::pointer_mask;
+
     return 0; // Break at this line
 }

[github-actions]

✅ With the latest revision this PR passed the C/C++ code formatter.

[Endilll]

This PR is a draft because the new test, which is simplified llvm::PointerIntPair and llvm::PointerIntPairInfo, doesn't pass even with the proposed fix.

[Michael137]

LGTM

Can you add a bit more detail into the commit message and PR description about what the underlying problem was? I.e., DeclContextGetName used to return type names without template parameters which FindTypes would then compare against the DW_AT_names from DWARF. But the names in DWARF do contain template parameters, causing the lookups to fail.

[JDevlieghere]

This breaks TestDataFormatterLibcxxChrono.py on macOS:

https://green.lab.llvm.org/green/view/LLDB/job/lldb-cmake/66400/
https://green.lab.llvm.org/green/view/LLDB/job/as-lldb-cmake/16178/

CC @rastogishubham

[Endilll]

Trying to take a look, but https://green.lab.llvm.org is extraordinary slow for me.

[JDevlieghere]

This is what the failure looks like:

======================================================================
FAIL: test_with_run_command_dsym (TestDataFormatterLibcxxChrono.LibcxxChronoDataFormatterTestCase)
   Test that that file and class static variables display correctly.
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/Users/jonas/llvm/llvm-project/lldb/packages/Python/lldbsuite/test/lldbtest.py", line 1676, in test_method
    return attrvalue(self)
  File "/Users/jonas/llvm/llvm-project/lldb/packages/Python/lldbsuite/test/decorators.py", line 150, in wrapper
    return func(*args, **kwargs)
  File "/Users/jonas/llvm/llvm-project/lldb/test/API/functionalities/data-formatter/data-formatter-stl/libcxx/chrono/TestDataFormatterLibcxxChrono.py", line 35, in test_with_run_command
    self.expect(
  File "/Users/jonas/llvm/llvm-project/lldb/packages/Python/lldbsuite/test/lldbtest.py", line 2447, in expect
    self.fail(log_msg)
AssertionError: Ran command:
"frame variable ss_tp"

Got output:
(std::chrono::time_point<std::chrono::system_clock, std::chrono::duration<long long> >) ss_tp = {
  __d_ = (__rep_ = 0)
}

Expecting sub string: "ss_tp = date/time=1970-01-01T00:00:00Z timestamp=0 s" (was not found)

[Endilll]

@JDevlieghere I have an idea what have caused this failure, but I guess I have to enable libcxx to replicate it locally.

[JDevlieghere]

Great. Let me know if you need help reproducing stuff on macOS.

[rastogishubham]

I am going to revert this patch for now, to make sure that green dragon is up and running again

[rastogishubham]

Hi @Endilll this patch has been reverted with 831ba95 to make sure the lldb bots are green.

[Endilll]

I'd appreciate if you gave me a slightly bit more time, because I've been testing the fix locally in the meantime.