test(utils): cover goal-objective and bash-exit input boundaries (#8048)

* test(utils): cover goal-objective and bash-exit input boundaries

* test(utils): make the string-comparison boundary case actually pin the contract

Review follow-up: 400 digits is below CPython's default int-max-str-digits cap (4300), so an implementation regressed to int(match.group(2)) would still pass. Derive the digit count from sys.get_int_max_str_digits() + 1 instead. Also document how this file differs from test_tool_result_meta.py (direct return contract vs message normalization). No production code changes.

---------

Co-authored-by: jack12178 <jack12178@users.noreply.github.com>

backend/tests/test_utils_bash_exit_boundaries.py

@@ -0,0 +1,51 @@
+"""Boundary coverage for ``deerflow.utils.bash_exit``.
+
+``backend/tests/test_tool_result_meta.py`` already drives the same markers
+through ``normalize_tool_message``. This file pins the direct return contract
+of ``bash_exit_status`` -- the exact ``(status, marker)`` tuple -- for the
+inputs where string handling, sign, or unbounded length matters. The few cases
+that overlap the existing suite are kept as the baseline for those assertions.
+"""
+
+from __future__ import annotations
+
+import sys
+
+from deerflow.utils.bash_exit import bash_exit_status
+
+
+def test_exit_code_tail_is_recognised_only_at_the_end() -> None:
+    assert bash_exit_status("output\nExit Code: 0") == ("success", "Exit Code: 0")
+    assert bash_exit_status("output\nExit Code: 1") == ("error", "Exit Code: 1")
+
+
+def test_embedded_phrase_in_ordinary_output_is_not_a_marker() -> None:
+    assert bash_exit_status("he said Exit Code: 0 then left") is None
+
+
+def test_signed_zero_is_success() -> None:
+    assert bash_exit_status("Command exited with code -0") == (
+        "success",
+        "Command exited with code -0",
+    )
+
+
+def test_negative_and_timeout_codes_are_errors() -> None:
+    for code in ("-1", "124", "137"):
+        assert bash_exit_status(f"Exit Code: {code}") == ("error", f"Exit Code: {code}")
+
+
+def test_unbounded_digits_are_compared_as_strings() -> None:
+    # CPython >= 3.11 refuses int() on digit strings longer than
+    # sys.get_int_max_str_digits() (4300 by default). Exceeding that cap makes
+    # an implementation regressed to int(match.group(2)) raise ValueError here
+    # instead of silently returning the same ("error", marker) tuple.
+    # A disabled cap (0) falls back to the default so the test still uses a
+    # length far beyond the standard conversion limit.
+    limit = sys.get_int_max_str_digits() or 4300
+    huge = "9" * (limit + 1)
+    assert bash_exit_status(f"Exit Code: {huge}") == ("error", f"Exit Code: {huge}")
+
+
+def test_trailing_whitespace_after_tail_is_allowed() -> None:
+    assert bash_exit_status("done\nExit Code: 0  \n") == ("success", "Exit Code: 0")

backend/tests/test_utils_goal_objective_boundaries.py

@@ -0,0 +1,32 @@
+"""Boundary coverage for ``deerflow.utils.goal_objective``."""
+
+from __future__ import annotations
+
+import pytest
+
+from deerflow.utils.goal_objective import MAX_GOAL_OBJECTIVE_CHARS, normalize_goal_objective
+
+
+def test_normalize_goal_objective_collapses_all_whitespace() -> None:
+    assert normalize_goal_objective("  a  \n\t b   c  ") == "a b c"
+
+
+def test_normalize_goal_objective_rejects_blank_only_input() -> None:
+    for blank in ("", "   ", "\n\t\r "):
+        with pytest.raises(ValueError, match="must not be empty"):
+            normalize_goal_objective(blank)
+
+
+def test_normalize_goal_objective_accepts_exact_limit() -> None:
+    text = "x" * MAX_GOAL_OBJECTIVE_CHARS
+    assert normalize_goal_objective(text) == text
+
+
+def test_normalize_goal_objective_rejects_one_over_limit() -> None:
+    with pytest.raises(ValueError, match="at most"):
+        normalize_goal_objective("x" * (MAX_GOAL_OBJECTIVE_CHARS + 1))
+
+
+def test_normalize_goal_objective_counts_characters_not_bytes() -> None:
+    text = "\u4e2d" * MAX_GOAL_OBJECTIVE_CHARS
+    assert normalize_goal_objective(text) == text