debugger: Fix gdb adapter and logger (#28280)

There were two bugs that caused the gdb adapter not to work properly,
one on our end and one their end.

The bug on our end was sending `stopOnEntry: null` in our launch request
when stop on entry had no value. I fixed that bug across all dap
adapters

The other bug had to do with python's "great" type system and how we
serialized our unit structs to json; mainly,
`ConfigurationDoneArguments` and `ThreadsArguments`. Gdb seems to follow
a pattern for handling requests where they pass `**args` to a function,
this errors out when the equivalent json is `"arguments": null`.

```py
@capability("supportsConfigurationDoneRequest")
@request("configurationDone", on_dap_thread=True)
def config_done(**args): ### BUG!!
    ...
```

Release Notes:

- N/A

crates/debugger_ui/src/tests/debugger_panel.rs

@@ -270,7 +270,7 @@ async fn test_we_can_only_have_one_panel_per_debug_session(
                     .clone()
             });
 
-            assert_eq!(client.id(), active_session.read(cx).session_id(cx).unwrap());
+            assert_eq!(client.id(), active_session.read(cx).session_id(cx));
             assert_eq!(
                 ThreadId(1),
                 running_state.read(cx).selected_thread_id().unwrap()
@@ -307,7 +307,7 @@ async fn test_we_can_only_have_one_panel_per_debug_session(
                     .clone()
             });
 
-            assert_eq!(client.id(), active_session.read(cx).session_id(cx).unwrap());
+            assert_eq!(client.id(), active_session.read(cx).session_id(cx));
             assert_eq!(
                 ThreadId(1),
                 running_state.read(cx).selected_thread_id().unwrap()