From f1c13377a3979f7732ad5ff693ecb549855aba49 Mon Sep 17 00:00:00 2001
From: Teknium <127238744+teknium1@users.noreply.github.com>
Date: Sat, 8 Aug 2026 11:57:45 -0700
Subject: [PATCH] test(cron): regression coverage for Windows encoding cluster

- CJK/emoji round-trip + human-readable jobs.json (PRs #52302/#29754)
- emoji through no_agent script stdout capture (issue #42384)
- truncated/invalid UTF-8 script stdout must not raise (#47393)
---
 tests/cron/test_cron_script.py | 45 ++++++++++++++++++++++++++++++++++
 tests/cron/test_jobs.py        | 24 ++++++++++++++++++
 2 files changed, 69 insertions(+)

diff --git a/tests/cron/test_cron_script.py b/tests/cron/test_cron_script.py
index 3e77293ea7..dbe6bf9f6b 100644
--- a/tests/cron/test_cron_script.py
+++ b/tests/cron/test_cron_script.py
@@ -200,6 +200,51 @@ class TestRunJobScript:
         assert "encoding" not in captured["kwargs"]
         assert "errors" not in captured["kwargs"]
 
+    def test_emoji_stdout_round_trips_through_script_capture(self, cron_env):
+        """Emoji in script stdout must reach the caller intact (#42384).
+
+        On Windows the fix is the utf-8 + errors='replace' popen kwargs
+        (asserted above); on POSIX the UTF-8 locale default must already
+        carry emoji through. Either way the delivery content is the real
+        text, never an exception.
+        """
+        from cron.scheduler import _run_job_script
+
+        script = cron_env / "scripts" / "emoji.py"
+        script.write_text(
+            'import sys\n'
+            'sys.stdout.buffer.write("backup done \\N{PARTY POPPER} 日次".encode("utf-8"))\n',
+            encoding="utf-8",
+        )
+
+        success, output = _run_job_script("emoji.py")
+
+        assert success is True
+        assert "backup done 🎉 日次" == output
+
+    def test_invalid_utf8_stdout_does_not_raise(self, cron_env):
+        """Truncated/invalid UTF-8 in script stdout must never escape as an
+        exception (#47393) — a raised UnicodeDecodeError higher up would
+        silently drop the whole delivery (#42384). The run may fail, but it
+        must fail as a (False, message) result the scheduler can deliver.
+        """
+        from cron.scheduler import _run_job_script
+
+        script = cron_env / "scripts" / "bad_bytes.py"
+        # b'\xe6\x97' is the first two bytes of a three-byte CJK sequence —
+        # a truncated write, exactly the shape reported in #47393.
+        script.write_text(
+            "import sys\n"
+            "sys.stdout.buffer.write(b'partial \\xe6\\x97')\n",
+            encoding="utf-8",
+        )
+
+        success, output = _run_job_script("bad_bytes.py")  # must not raise
+
+        assert isinstance(success, bool)
+        assert isinstance(output, str)
+        assert output  # a message is always produced, never a silent drop
+
 
 class TestBuildJobPromptWithScript:
     """Test that script output is injected into the prompt."""
diff --git a/tests/cron/test_jobs.py b/tests/cron/test_jobs.py
index 17b4347be4..0416bdcc7c 100644
--- a/tests/cron/test_jobs.py
+++ b/tests/cron/test_jobs.py
@@ -190,6 +190,30 @@ def tmp_cron_dir(tmp_path, monkeypatch):
 
 
 class TestJobCRUD:
+    def test_cjk_and_emoji_round_trip_readable_in_jobs_json(self, tmp_cron_dir):
+        """CJK/emoji job text must round-trip AND stay human-readable on disk.
+
+        With json.dump's default ensure_ascii=True, every non-ASCII char in
+        jobs.json is written as \\uXXXX escapes, which users reported as
+        unreadable garbage when inspecting their job store (#52302, #29754).
+        ensure_ascii=False + the existing encoding="utf-8" writer keeps the
+        text literal; the utf-8-sig reader must parse it back identically.
+        """
+        name = "日次レポート 🎉 café"
+        job = create_job(prompt=f"Summarize {name}", schedule="30m", name=name)
+
+        # Round-trip through save/load is lossless.
+        fetched = get_job(job["id"])
+        assert fetched["name"] == name
+        assert name in fetched["prompt"]
+
+        # On-disk representation is literal UTF-8, not \uXXXX escapes.
+        from cron.jobs import JOBS_FILE
+        raw = JOBS_FILE.read_text(encoding="utf-8")
+        assert "日次レポート" in raw
+        assert "🎉" in raw
+        assert "\\u65e5" not in raw
+
     def test_create_and_get(self, tmp_cron_dir):
         job = create_job(prompt="Check server status", schedule="30m")
         assert job["id"]