searxng-use-cli/tests/test_utf8_and_winpath.py

"""Tests for v1.8.1 Windows compatibility fixes.

Covers:
* ``force_utf8_stdout()`` — reconfigures stdout/stderr to UTF-8, idempotent,
  tolerates non-UTF-8 encodings, never raises on any input.
* ``_windows_appdata_config_dir()`` — returns APPDATA-based path on Windows,
  empty Path on POSIX.
* ``resolve_instances()`` — discovers config under ``%APPDATA%/searxng-cli/``
  when set (simulating Windows), still works when APPDATA is unset (POSIX).
* ``fetch.py`` failure diagnostics — error output includes ``status_code=``,
  ``cause=``, ``url=`` fields for programmatic diagnosis.
"""
import io
import os
import sys
from pathlib import Path
from unittest.mock import patch

import pytest

# Ensure scripts/ is on sys.path
sys.path.insert(0, str(Path(__file__).resolve().parent.parent / "scripts"))

from common import force_utf8_stdout


# ----- force_utf8_stdout -----

def test_force_utf8_stdout_is_idempotent():
    """Calling twice should not raise and should leave stdout usable."""
    force_utf8_stdout()
    force_utf8_stdout()  # second call: encoding already utf-8 or close — no-op
    # stdout must still be writable
    print("test", end="")
    sys.stdout.flush()


def test_force_utf8_stdout_replaces_non_utf8_stream():
    """A non-UTF-8 stdout should be reconfigured to UTF-8."""
    # Simulate a GBK stream by creating a fresh TextIOWrapper with GBK
    original = sys.stdout
    try:
        buf = io.BytesIO()
        # Use errors='replace' so writing non-GBK chars doesn't crash the test
        sys.stdout = io.TextIOWrapper(buf, encoding="gbk", errors="replace")
        force_utf8_stdout()
        # After force_utf8_stdout, encoding should now be utf-8 (or utf-8-sig)
        enc = (sys.stdout.encoding or "").lower().replace("-", "")
        assert enc in ("utf8", "utf8sig"), f"expected utf-8, got {enc}"
    finally:
        sys.stdout = original


def test_force_utf8_stdout_handles_missing_buffer():
    """A stream without .buffer should not crash the function."""
    class FakeStream:
        encoding = "gbk"
        # No .buffer, no .reconfigure

    original = sys.stdout
    try:
        sys.stdout = FakeStream()
        # Should not raise
        force_utf8_stdout()
    finally:
        sys.stdout = original


def test_force_utf8_stdout_skips_already_utf8():
    """An already-UTF-8 stdout should be untouched."""
    original = sys.stdout
    try:
        buf = io.BytesIO()
        sys.stdout = io.TextIOWrapper(buf, encoding="utf-8")
        force_utf8_stdout()
        # Still utf-8
        assert (sys.stdout.encoding or "").lower().replace("-", "") == "utf8"
    finally:
        sys.stdout = original


def test_force_utf8_stdout_allows_non_ascii_print():
    """End-to-end: after force_utf8_stdout, printing non-ASCII chars works.

    Regression test for the Windows GBK crash where ``print('\\xa0')`` raised
    ``UnicodeEncodeError``. With UTF-8 enforced, it must succeed silently.
    """
    force_utf8_stdout()
    # This must not raise. On a broken GBK stdout it would.
    print("\xa0中文测试\u2014em-dash", end="")
    sys.stdout.flush()


# ----- _windows_appdata_config_dir -----

def test_windows_appdata_dir_returns_path_when_set(monkeypatch):
    """When APPDATA is set, returns APPDATA/searxng-cli."""
    import search
    monkeypatch.setenv("APPDATA", "/fake/appdata")
    result = search._windows_appdata_config_dir()
    assert result == Path("/fake/appdata") / "searxng-cli"


def test_windows_appdata_dir_returns_empty_when_unset(monkeypatch):
    """When APPDATA is unset (POSIX), returns a sentinel Path that never exists."""
    import search
    monkeypatch.delenv("APPDATA", raising=False)
    result = search._windows_appdata_config_dir()
    # Sentinel path must never exist on disk — safe to append to candidate lists
    assert not result.exists()


# ----- resolve_instances with APPDATA -----

def test_resolve_instances_finds_appdata_config(monkeypatch, tmp_path):
    """resolve_instances discovers searxng.toml under APPDATA (Windows sim)."""
    import search

    # Simulate Windows: set APPDATA to a temp dir, clear other sources
    appdata = tmp_path / "appdata"
    appdata.mkdir()
    cfg_dir = appdata / "searxng-cli"
    cfg_dir.mkdir()
    (cfg_dir / "searxng.toml").write_text(
        '[searxng]\ninstance = "https://from-appdata.example.com"\n',
        encoding="utf-8",
    )
    monkeypatch.setenv("APPDATA", str(appdata))
    monkeypatch.delenv("SEARXNG_INSTANCE", raising=False)

    # CWD and HOME must not have a config (so we isolate the test)
    monkeypatch.chdir(tmp_path)
    fake_home = tmp_path / "fake-home"
    fake_home.mkdir()
    monkeypatch.setenv("HOME", str(fake_home))
    # On Windows, Path.home() uses USERPROFILE, not HOME
    monkeypatch.setenv("USERPROFILE", str(fake_home))

    result = search.resolve_instances()
    assert result == ["https://from-appdata.example.com"]


def test_resolve_instances_appdata_txt_fallback(monkeypatch, tmp_path):
    """instances.txt under APPDATA is also discovered."""
    import search

    appdata = tmp_path / "appdata"
    appdata.mkdir()
    cfg_dir = appdata / "searxng-cli"
    cfg_dir.mkdir()
    (cfg_dir / "instances.txt").write_text(
        "https://from-txt.example.com\n", encoding="utf-8"
    )
    monkeypatch.setenv("APPDATA", str(appdata))
    monkeypatch.delenv("SEARXNG_INSTANCE", raising=False)
    monkeypatch.chdir(tmp_path)
    fake_home = tmp_path / "fake-home"
    fake_home.mkdir()
    monkeypatch.setenv("HOME", str(fake_home))
    monkeypatch.setenv("USERPROFILE", str(fake_home))

    result = search.resolve_instances()
    assert result == ["https://from-txt.example.com"]


def test_resolve_instances_appdata_empty(monkeypatch, tmp_path):
    """When APPDATA set but no config there, falls through cleanly."""
    import search

    appdata = tmp_path / "appdata"
    appdata.mkdir()
    monkeypatch.setenv("APPDATA", str(appdata))
    monkeypatch.delenv("SEARXNG_INSTANCE", raising=False)
    monkeypatch.chdir(tmp_path)
    fake_home = tmp_path / "fake-home"
    fake_home.mkdir()
    monkeypatch.setenv("HOME", str(fake_home))
    monkeypatch.setenv("USERPROFILE", str(fake_home))

    result = search.resolve_instances()
    assert result == []