refactor(tests): consolidate test tooling under tests/tools - #499

tests/tools/compile_commands.py (1)

149-171: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value

Consider adding a std parameter to write_entries for parity with write_cdb.

write_cdb allows overriding the C++ standard, but write_entries hardcodes "-std=c++17". Minor inconsistency between the two sibling helpers; low priority since current callers only need c++17.

♻️ Optional parity fix

-def write_entries(workspace, entries):
+def write_entries(workspace, entries, *, std: str = "c++17"):
     """Write a compile_commands.json with per-file extra arguments.

     Args:
         workspace: Root directory of the workspace.
         entries: List of (file_name, extra_args) pairs; a file may appear
             multiple times to model multi-configuration projects.
+        std: C++ standard version (default: c++17).
     """
     data = [
         {
             "directory": str(workspace),
             "file": str(workspace / f),
             "arguments": [
                 "clang++",
-                "-std=c++17",
+                f"-std={std}",
                 "-fsyntax-only",
                 *args,
                 str(workspace / f),
             ],
         }
         for f, args in entries
     ]
     (workspace / "compile_commands.json").write_text(json.dumps(data))

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@tests/tools/compile_commands.py` around lines 149 - 171, Add an optional std
parameter to write_entries, defaulting to "c++17", and build the compiler
argument from it instead of hardcoding "-std=c++17". Keep the existing behavior
for current callers and align the parameter naming and handling with the sibling
write_cdb helper.

tests/tools/lifecycle.py (1)

121-145: 🩺 Stability & Availability | 🔵 Trivial | ⚡ Quick win

Log swallowed exceptions during best-effort shutdown.

Static analysis flags the three try/except Exception: pass blocks in shutdown_client. The best-effort teardown intent is reasonable (a crashed/already-dead server shouldn't blow up teardown), but silently discarding every exception makes flaky CI teardown failures hard to diagnose.

🔍 Suggested logging on swallowed exceptions

 async def shutdown_client(c: CliceClient, *, verbose: bool = False) -> None:
     """Gracefully shut down a client, force-kill if needed."""
     try:
         await asyncio.wait_for(c.shutdown_async(None), timeout=10.0)
-    except Exception:
-        pass
+    except Exception as exc:
+        print(f"[shutdown_client] shutdown_async failed: {exc!r}", flush=True)

     try:
         c.exit(None)
-    except Exception:
-        pass
+    except Exception as exc:
+        print(f"[shutdown_client] exit() failed: {exc!r}", flush=True)
     ...
     try:
         await assert_server_exited_cleanly(c.server)
     finally:
         try:
             await c.stop_io()
             await asyncio.sleep(0.1)
-        except Exception:
-            pass
+        except Exception as exc:
+            print(f"[shutdown_client] stop_io failed: {exc!r}", flush=True)

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@tests/tools/lifecycle.py` around lines 121 - 145, Update shutdown_client to
log exceptions from each best-effort shutdown operation instead of silently
passing: shutdown_async, exit, and stop_io. Preserve the existing teardown flow
and exception swallowing, but emit concise diagnostic messages including the
caught exception and identify which operation failed.

Source: Linters/SAST tools