[Lldb-commits] [lldb] [LLDB] added getName method in SBModule (PR #150331)

[Jacob Lalonde via lldb-commits]

================
@@ -18,6 +18,26 @@ def tearDown(self):
         if self.background_pid:
             os.kill(self.background_pid, signal.SIGKILL)
 
+    def test_getname(self):
+        """Test the SBModule::GetName() method"""
+        self.build()
+        target, _, _, _ = lldbutil.run_to_source_breakpoint(
+            self, "// break here", lldb.SBFileSpec("main.c")
+        )
+
+        self.assertGreater(target.GetNumModules(), 0)
+        for i in range(target.GetNumModules()):
+            module = target.GetModuleAtIndex(i)
+            file_spec = module.GetFileSpec()
+            name = module.GetName()
+            if file_spec.IsValid() and file_spec.exists:
+#If file is valid and file exist, expect GetName() to be None
+                self.assertIsNone(name, f"Expected None for module with valid file {file_spec.GetFilename()}, got {name!r}")
+            else:
+#If no valid file, expect GetName() to be a non - empty string
+                self.assertIsInstance(name, str)
+                self.assertTrue(name, "Expected a non-empty name for module without a valid file")
----------------
Jlalond wrote:

@labath 

I think the best case for this API would be returning a valid string identifier when there is no backing file. VDSO is the example I have in mind but a module jitted in memory only could also make sense.

Per the code you linked
```
  ConstString m_object_name; ///< The name an object within this module that is
                                                    /// selected, or empty of the module is represented
                                                    /// by \a m_file.
```

Shouldn't this be the case? I'm very open to how we implement this but my goals align with the comments, which is obtaining a valid identifier when I can't fall back to the file for identification

https://github.com/llvm/llvm-project/pull/150331