[Lldb-commits] [lldb] [lldb-dap] Support StackFrameFormat (PR #137113)

[John Harrison via lldb-commits]

================
@@ -178,14 +179,45 @@ void StackTraceRequestHandler::operator()(
   llvm::json::Array stack_frames;
   llvm::json::Object body;
 
+  lldb::SBFormat frame_format = dap.frame_format;
+
+  if (const auto *format = arguments->getObject("format")) {
+    const bool parameters = GetBoolean(format, "parameters").value_or(false);
+    const bool parameter_names =
+        GetBoolean(format, "parameterNames").value_or(false);
+    const bool parameter_values =
+        GetBoolean(format, "parameterValues").value_or(false);
+    const bool line = GetBoolean(format, "line").value_or(false);
+    const bool module = GetBoolean(format, "module").value_or(false);
+    const bool include_all = GetBoolean(format, "includeAll").value_or(false);
----------------
ashgti wrote:

I think this is maybe a different meaning than the spec meant for `includeAll`.

It says:

> Includes all stack frames, including those the debug adapter might otherwise hide.

Which to me, sounds like its about hiding frames that we might not find interesting (e.g. frames from system libraries).

I think if this is `false` you could skip showing the extended backtraces.

https://github.com/llvm/llvm-project/pull/137113