camoufox/pythonlib/tests/test_identity_salt.py

"""Per-identity draws: unrelated launches must not share them, a pinned identity must.

identity_seed() used to hash only the UA, platform, screen size and core count.
Those take a handful of values per OS, so over 500 launches the seed took 12-30
distinct values, and every install drew its fonts, voices, GPU, media devices and
canvas/audio noise seeds from that same short list.
"""

from contextlib import contextmanager
from unittest import mock

import orjson
import pytest

from camoufox import cpu_affinity, utils
from camoufox import fingerprints as fp


@contextmanager
def host():
    with mock.patch.object(utils, "get_screen_cons", lambda headless: None), (
        mock.patch.object(utils, "has_display", lambda env: False)
    ), mock.patch.object(utils, "installed_verstr", lambda: "150.0.2"), (
        mock.patch.object(utils, "launch_path", lambda **kwargs: "/nonexistent/camoufox")
    ):
        yield


def config_of(options):
    env = options["env"]
    chunks = sorted(
        (int(k.rsplit("_", 1)[1]), v) for k, v in env.items() if k.startswith("CAMOU_CONFIG_")
    )
    return orjson.loads("".join(chunk for _, chunk in chunks))


def launch(**kwargs):
    kwargs.setdefault("os", "linux")
    kwargs.setdefault("headless", True)
    kwargs.setdefault("i_know_what_im_doing", True)
    with host():
        return config_of(utils.launch_options(**kwargs))


DRAWN = ("canvas:seed", "audio:seed", "fonts", "voices", "webGl:renderer")


def drawn(config):
    return {k: orjson.dumps(config.get(k)) for k in DRAWN}


class TestUnpinnedLaunchesAreDistinct:
    def test_noise_seeds_do_not_collide(self):
        seeds = [launch()["canvas:seed"] for _ in range(40)]
        # 40 draws from 2**32: any collision means the seed space collapsed.
        assert len(set(seeds)) == len(seeds)

    def test_same_presented_values_still_differ(self):
        # The same UA/platform/screen/cores, i.e. what two users on the same
        # common machine present, must not yield the same noise seeds.
        config = {"navigator.userAgent": "x", "navigator.platform": "Win32",
                  "screen.width": 1920, "screen.height": 1080, "navigator.hardwareConcurrency": 8}
        seeds = {fp.identity_seed(config, fp.identity_salt()) for _ in range(200)}
        assert len(seeds) == 200


class TestPinnedIdentityIsStable:
    def test_fixed_fingerprint_reproduces_every_draw(self):
        fingerprint = fp.generate_fingerprint(os="linux")
        first = launch(fingerprint=fingerprint)
        second = launch(fingerprint=fingerprint)
        assert drawn(first) == drawn(second)

    def test_fixed_preset_reproduces_noise_seeds(self):
        preset = fp.get_random_preset(os="windows", ff_version="150")
        if not preset:
            pytest.skip("no presets bundled")
        first = launch(os="windows", fingerprint_preset=preset)
        second = launch(os="windows", fingerprint_preset=preset)
        assert (first["canvas:seed"], first["audio:seed"]) == (second["canvas:seed"], second["audio:seed"])
        assert first["fonts"] == second["fonts"]

    def test_caller_seeds_are_kept(self):
        config = launch(config={"canvas:seed": 7, "audio:seed": 9})
        assert (config["canvas:seed"], config["audio:seed"]) == (7, 9)

    def test_salt_of_equal_objects_is_equal(self):
        a = fp.generate_fingerprint(os="windows")
        assert fp.identity_salt(a) == fp.identity_salt(a)
        assert fp.identity_salt({"a": 1, "b": 2}) == fp.identity_salt({"b": 2, "a": 1})


class TestVoicesFollowLocale:
    def test_windows_fr_identity_has_french_voices(self, monkeypatch):
        for _ in range(5):
            config = launch(os="windows", locale="fr-FR")
            langs = {v["lang"] for v in config["voices"]}
            assert "fr-FR" in langs, langs


class TestCoreCountFloor:
    def test_small_pinnable_host_reports_table_floor(self, monkeypatch):
        # A 1-3 core host reports 2, the table's floor and the lowest count the
        # corpus records. It used to report 4, which no 2-core machine can back
        # up: 4 cannot be pinned on a 3-core host, so the page measured 3 while
        # being told 4. At 2 the pin succeeds on a 2- or 3-core host, and the
        # 1-core tail (told 2, measures 1) is closer than 4 was.
        monkeypatch.setattr(cpu_affinity, "supported", lambda: True)
        for host_cores in (1, 2, 3):
            monkeypatch.setattr(fp, "host_cpu_count", lambda n=host_cores: n)
            for drawn_cores in (1, 2, 3, 8):
                c = {"navigator.hardwareConcurrency": drawn_cores}
                fp.fix_hardware_concurrency(c)
                assert c["navigator.hardwareConcurrency"] == 2, (host_cores, drawn_cores)

    def test_unpinned_launch_reports_host(self, monkeypatch):
        monkeypatch.setattr(cpu_affinity, "supported", lambda: True)
        monkeypatch.setattr(fp, "host_cpu_count", lambda: 16)
        c = {"navigator.hardwareConcurrency": 8}
        fp.fix_hardware_concurrency(c, can_pin=False)
        assert c["navigator.hardwareConcurrency"] == 16

    def test_recorded_counts_are_in_the_table(self):
        for n in (18, 22, 28, 32):
            assert n in fp.PLAUSIBLE_CORE_COUNTS


class TestAffinityPick:
    def test_adjacent_cores_from_a_random_start(self):
        cores = list(range(16))
        starts = set()
        for _ in range(200):
            picked = cpu_affinity._pick(cores, 4)
            assert len(picked) == 4 and set(picked) <= set(cores)
            ring = sorted(picked)
            # adjacent modulo 16
            assert any(all((s + i) % 16 in picked for i in range(4)) for s in ring)
            starts.add(tuple(picked))
        assert len(starts) > 4


class TestPrefsEnvIsAscii:
    def test_non_ascii_pref_round_trips(self):
        prefs = {"font.name.serif.ja": "游明朝", "intl.accept_languages": "fr-FR, fr"}
        env = utils.get_pref_env_vars(prefs)
        joined = "".join(env[f"CAMOU_PREFS_{i}"] for i in range(1, len(env) + 1))
        assert joined.isascii()
        assert orjson.loads(joined) == prefs