Files
libretro/tests/test_install_experience.py
T
Abdessamad Derraz 392aacb207 fix: hash a large download through its own handle
On Windows the partial copy of a file of 16 MB or more is locked with msvcrt.locking, a LockFile region: the locking process cannot access that region through a second handle (Microsoft LockFile remarks). Hashing reopened the path, the read failed with a sharing violation, and every large file was reported as 'folder not writable' (issue 82). Both digests are now read in one pass through the handle that holds the lock. A read-only file already at the destination had the same misleading report: on Windows its attribute is now cleared before the verified copy replaces it, and a refusal that remains is named for what it is.
2026-10-21 03:46:14 +02:00

438 lines
18 KiB
Python

"""What the one-liner shows and how it recovers.
A live progress line with speed and time left, a plan before a large
download, a dry run, a quiet mode, a version, a pause between retries that
honours Retry-After, and large files that resume where they stopped.
"""
from __future__ import annotations
import contextlib
import hashlib
import http.server
import io
import json
import os
import subprocess
import sys
import tempfile
import threading
import unittest
from pathlib import Path
from unittest import mock
REPO_ROOT = Path(__file__).resolve().parent.parent
sys.path.insert(0, str(REPO_ROOT))
sys.path.insert(0, str(REPO_ROOT / "tests"))
import install # noqa: E402
from test_install_selection import FakeRepository # noqa: E402
PAYLOAD = bytes(range(256)) * 400 # 102400 bytes
class _RangeHandler(http.server.BaseHTTPRequestHandler):
"""Serves PAYLOAD with Range support and fails first when asked to."""
ranges: list[str] = []
fail_first: list[int] = []
def log_message(self, *args) -> None: # noqa: D102
pass
def do_GET(self) -> None: # noqa: N802, D102
if self.fail_first:
code = self.fail_first.pop(0)
self.send_response(code)
self.send_header("Retry-After", "0")
self.end_headers()
return
header = self.headers.get("Range")
self.ranges.append(header or "")
start = 0
if header and header.startswith("bytes="):
start = int(header[6:].split("-")[0])
self.send_response(206)
self.send_header("Content-Range", f"bytes {start}-{len(PAYLOAD) - 1}/{len(PAYLOAD)}")
else:
self.send_response(200)
body = PAYLOAD[start:]
self.send_header("Content-Length", str(len(body)))
self.end_headers()
self.wfile.write(body)
class _Server:
def __enter__(self):
_RangeHandler.ranges = []
_RangeHandler.fail_first = []
self.httpd = http.server.ThreadingHTTPServer(("127.0.0.1", 0), _RangeHandler)
threading.Thread(target=self.httpd.serve_forever, daemon=True).start()
port = self.httpd.server_address[1]
self.patch = mock.patch.object(install, "RAW_FILE_URL", f"http://127.0.0.1:{port}/{{path}}")
self.patch.start()
return self
def __exit__(self, *exc):
self.patch.stop()
self.httpd.shutdown()
self.httpd.server_close()
ENTRY = {
"dest": "big.bin", "repo_path": "bios/big.bin", "size": len(PAYLOAD),
"sha256": hashlib.sha256(PAYLOAD).hexdigest(),
}
class LargeFilesResume(unittest.TestCase):
def test_a_partial_copy_is_completed_with_a_range_request(self):
with _Server(), tempfile.TemporaryDirectory() as tmp, \
mock.patch.object(install, "RESUME_FROM_SIZE", 1024):
part = Path(tmp) / ".big.bin.part"
part.write_bytes(PAYLOAD[:40000])
dest, ok, reason = install._download_one(dict(ENTRY), Path(tmp))
self.assertEqual((dest, ok, reason), ("big.bin", True, ""))
self.assertEqual((Path(tmp) / "big.bin").read_bytes(), PAYLOAD)
self.assertEqual(_RangeHandler.ranges, ["bytes=40000-"])
self.assertFalse(part.exists())
def test_a_small_file_starts_over(self):
with _Server(), tempfile.TemporaryDirectory() as tmp:
dest, ok, _ = install._download_one(dict(ENTRY), Path(tmp))
self.assertTrue(ok)
self.assertEqual(_RangeHandler.ranges, [""])
@unittest.skipUnless(hasattr(os, "symlink"), "needs symlinks")
def test_a_link_at_the_partial_name_is_never_written_through(self):
with _Server(), tempfile.TemporaryDirectory() as tmp, \
mock.patch.object(install, "RESUME_FROM_SIZE", 1024):
victim = Path(tmp) / "victim.txt"
victim.write_bytes(b"keep me")
part = Path(tmp) / "bios" / ".big.bin.part"
part.parent.mkdir()
part.symlink_to(victim)
dest, ok, _ = install._download_one(dict(ENTRY), Path(tmp) / "bios")
self.assertTrue(ok)
self.assertEqual(victim.read_bytes(), b"keep me")
final = Path(tmp) / "bios" / "big.bin"
self.assertFalse(final.is_symlink())
self.assertEqual(final.read_bytes(), PAYLOAD)
self.assertEqual(_RangeHandler.ranges, [""])
class TwoRunsOneFolder(unittest.TestCase):
"""Two installs into one BIOS folder appended to the same partial copy:
each hashed the other's bytes and one renamed a file still growing."""
@unittest.skipIf(os.name == "nt", "flock semantics")
def test_a_partial_held_by_another_run_is_left_alone(self):
with _Server(), tempfile.TemporaryDirectory() as tmp, \
mock.patch.object(install, "RESUME_FROM_SIZE", 1024):
part = Path(tmp) / ".big.bin.part"
part.write_bytes(PAYLOAD[:40000])
other = install._claim_partial(part)
self.assertIsNotNone(other)
try:
dest, ok, reason = install._download_one(dict(ENTRY), Path(tmp))
self.assertEqual((ok, reason), (True, ""))
self.assertEqual((Path(tmp) / "big.bin").read_bytes(), PAYLOAD)
self.assertEqual(part.read_bytes(), PAYLOAD[:40000])
self.assertEqual(_RangeHandler.ranges, [""])
self.assertIsNone(install._claim_partial(part))
finally:
other.close()
self.assertIsNotNone(install._claim_partial(part))
class APartialRenamedBeforeTheLock(unittest.TestCase):
"""Open and lock are two steps. The run holding the partial could finish
in between, rename it into place and release it: the second run then
locked the installed BIOS, took it for a finished partial and truncated it."""
@unittest.skipIf(os.name == "nt", "flock semantics")
def test_the_installed_file_is_not_claimed(self):
import fcntl
with tempfile.TemporaryDirectory() as tmp:
part = Path(tmp) / ".big.bin.part"
final = Path(tmp) / "big.bin"
part.write_bytes(PAYLOAD)
real_flock = fcntl.flock
def finish_first(fd, op):
os.replace(part, final)
return real_flock(fd, op)
with mock.patch.object(fcntl, "flock", finish_first):
claimed = install._claim_partial(part)
self.assertIsNone(claimed)
self.assertEqual(final.read_bytes(), PAYLOAD)
class InstalledFileModes(unittest.TestCase):
"""A file came out 0600, the mode of the temporary it was written to: run
with sudo for a root-owned BIOS folder, the emulator's own user could not
read it, and RetroDECK's pico8 and vox launchers were never executable."""
def _mode(self, entry: dict) -> int:
with _Server(), tempfile.TemporaryDirectory() as tmp:
dest, ok, _ = install._download_one(dict(entry), Path(tmp))
self.assertTrue(ok)
return (Path(tmp) / dest).stat().st_mode & 0o777
def test_the_umask_decides_as_for_any_new_file(self):
umask = os.umask(0o022)
try:
self.assertEqual(self._mode(ENTRY), 0o644)
self.assertEqual(self._mode({**ENTRY, "executable": True}), 0o755)
finally:
os.umask(umask)
class ProgramsAreMarked(unittest.TestCase):
"""The manifest says which files are programs, read from their header:
the git mode cannot, hundreds of MAME archives carrying an execute bit."""
def test_header_decides(self):
sys.path.insert(0, str(REPO_ROOT / "scripts"))
import generate_pack
with tempfile.TemporaryDirectory() as tmp:
for data, expected in ((b"\x7fELF\x02", True), (b"#!/bin/sh\n", True),
(b"PK\x03\x04", False), (b"", False)):
path = Path(tmp) / "f"
path.write_bytes(data)
with self.subTest(head=data):
self.assertEqual(generate_pack._is_program(str(path)), expected)
def test_retrodeck_launchers_are_executable(self):
manifest = json.loads((REPO_ROOT / "install" / "retrodeck.json").read_text())
launchers = [f for f in manifest["files"]
if f["dest"].endswith(("pico-8/pico8", "voxatron/vox"))]
if not launchers:
self.skipTest("no launcher in the RetroDECK manifest")
for entry in launchers:
with self.subTest(dest=entry["dest"]):
self.assertIs(entry.get("executable"), True)
class ManifestHashesMatchTheScan(unittest.TestCase):
"""A manifest entry's sha1 is checked against the database scan of the
same path: under load this host hashed a 252 MB DSi NAND to a sha1 no
file holds, the same wrong value in two runs."""
def test_a_contradicted_hash_is_refused(self):
sys.path.insert(0, str(REPO_ROOT / "scripts"))
import generate_pack
with tempfile.TemporaryDirectory() as tmp:
path = Path(tmp) / "nand.bin"
path.write_bytes(b"real")
real = hashlib.sha1(b"real").hexdigest()
good = {"files": {real: {"path": str(path)}}}
self.assertEqual(generate_pack._stable_hashes(str(path), good)["sha1"], real)
bad = {"files": {"0" * 40: {"path": str(path)}}}
with self.assertRaises(RuntimeError):
generate_pack._stable_hashes(str(path), bad)
self.assertEqual(generate_pack._stable_hashes(str(path), {"files": {}})["sha1"], real)
def test_the_cache_is_keyed_by_what_was_checked(self):
"""Keyed by id(db), a database freed and replaced at the same address
got hashes verified against the first one back, unchecked."""
source = (REPO_ROOT / "scripts" / "generate_pack.py").read_text(encoding="utf-8")
body = source[source.index("def _stable_hashes("):]
body = body[: body.index("\ndef ")]
key_line = next(line for line in body.splitlines() if line.strip().startswith("key ="))
self.assertNotIn("id(", key_line)
self.assertIn("expected", key_line)
class RetriesWait(unittest.TestCase):
def test_a_server_error_is_retried_after_the_delay_it_asks(self):
with _Server(), tempfile.TemporaryDirectory() as tmp:
_RangeHandler.fail_first = [503]
dest, ok, _ = install._download_one(dict(ENTRY), Path(tmp))
self.assertTrue(ok)
def test_the_delay_doubles_and_follows_retry_after(self):
self.assertEqual([install._retry_delay(n, None) for n in (1, 2, 3, 4, 5)],
[1.0, 2.0, 4.0, 8.0, 8.0])
self.assertEqual(install._retry_delay(1, "5"), 5.0)
self.assertEqual(install._retry_delay(1, "999"), 60.0)
def test_a_refusal_names_the_rate_limit(self):
with _Server(), tempfile.TemporaryDirectory() as tmp, \
mock.patch.object(install, "MAX_RETRIES", 1):
_RangeHandler.fail_first = [403]
_dest, ok, reason = install._download_one(dict(ENTRY), Path(tmp))
self.assertFalse(ok)
self.assertIn("rate limit", reason)
class ProgressOutput(unittest.TestCase):
def _term(self, quiet=False):
stream = io.StringIO()
term = install.Terminal(stream, quiet=quiet)
return stream, term
def test_a_log_gets_one_line_per_file_and_failures_say_why(self):
stream, term = self._term()
progress = install.Progress(2, 10, term)
progress.done("a.bin", True)
progress.done("b.bin", False, "HTTP 404")
out = stream.getvalue()
self.assertIn("[1/2] a.bin ok", out)
self.assertIn("FAILED b.bin (HTTP 404)", out)
self.assertNotIn("\033", out)
def test_quiet_leaves_failures_to_the_summary(self):
stream, term = self._term(quiet=True)
progress = install.Progress(2, 10, term)
progress.done("a.bin", True)
progress.done("b.bin", False, "HTTP 404")
self.assertEqual(stream.getvalue(), "")
def test_durations_read_like_a_clock(self):
self.assertEqual([install.format_duration(s) for s in (0, 42, 185, 3729)],
["0:00", "0:42", "3:05", "1:02:09"])
def test_the_summary_names_each_failure(self):
buf = io.StringIO()
with contextlib.redirect_stdout(buf), self.assertRaises(SystemExit):
install._report_outcome(
[Path("/b")], 1, 0, 1, 0, False,
failures=[("x.bin", "HTTP 404")], transferred=2048, seconds=2.0,
)
out = buf.getvalue()
self.assertIn("failed: x.bin (HTTP 404)", out)
self.assertIn("2.0 KB downloaded in 0:02", out)
class CommandLine(unittest.TestCase):
def test_a_dry_run_writes_nothing(self):
with FakeRepository() as repo:
proc = repo.run("--dry-run")
self.assertEqual(proc.returncode, 0, proc.stderr)
self.assertIn("Would download", proc.stdout)
self.assertIn("Nothing was written", proc.stdout)
self.assertEqual(repo.installed(), [])
def test_quiet_prints_the_summary_only(self):
with FakeRepository() as repo:
proc = repo.run("--quiet")
self.assertEqual(proc.returncode, 0, proc.stderr)
self.assertNotIn("Checking existing files", proc.stdout)
self.assertIn("Done.", proc.stdout)
self.assertTrue(repo.installed())
def test_version_names_the_file_the_bootstraps_pin(self):
proc = subprocess.run(
[sys.executable, str(REPO_ROOT / "install.py"), "--version"],
capture_output=True, text=True, timeout=60, check=True,
)
digest = hashlib.sha256((REPO_ROOT / "install.py").read_bytes()).hexdigest()
self.assertIn(digest[:12], proc.stdout)
class ReviewCases(unittest.TestCase):
"""Saying no to the large-download question ended on Done, --quiet hid
fatal messages and swallowed --check, and a resumed copy inflated the speed."""
def test_declining_is_not_done(self):
buf = io.StringIO()
with contextlib.redirect_stdout(buf):
install._report_outcome([Path("/b")], 0, 3, 0, 0, False, declined=12)
out = buf.getvalue()
self.assertNotIn("Done.", out)
self.assertIn("12 files not downloaded", out)
self.assertIn("3 files already up to date", out)
def test_quiet_refuses_modes_it_would_silence(self):
for mode in ("--check", "--dry-run", "--list-platforms", "--list-systems"):
proc = subprocess.run(
[sys.executable, str(REPO_ROOT / "install.py"), "--quiet", mode],
capture_output=True, text=True, timeout=60, check=False,
)
with self.subTest(mode=mode):
self.assertEqual(proc.returncode, 2)
self.assertIn("--quiet applies to an install", proc.stderr)
def test_quiet_shows_what_a_failed_run_printed(self):
with tempfile.TemporaryDirectory() as home:
env = dict(os.environ, HOME=home, USERPROFILE=home, RETROBIOS_OS="linux",
XDG_CONFIG_HOME=home)
proc = subprocess.run(
[sys.executable, str(REPO_ROOT / "install.py"), "--quiet"],
capture_output=True, text=True, timeout=60, check=False,
env=env, stdin=subprocess.DEVNULL,
)
self.assertEqual(proc.returncode, 1)
self.assertIn("No supported platform detected", proc.stderr)
def test_a_resumed_copy_is_not_counted_as_fetched(self):
progress = install.Progress(1, 100, install.Terminal(io.StringIO()))
progress.add(60, resumed=True)
progress.add(40)
self.assertEqual(progress.transferred(), 40)
class WindowsMandatoryLock(unittest.TestCase):
"""Issue 82: on Windows the partial copy's lock is mandatory, so hashing
it by reopening its path failed with a sharing violation, and every file
of 16 MB or more was reported as "folder not writable"."""
def test_the_locked_partial_is_hashed_through_its_handle(self):
real_open = open
def locked_open(path, *args, **kwargs):
if str(path).endswith(".part"):
raise PermissionError(33, "The process cannot access the file")
return real_open(path, *args, **kwargs)
with _Server(), tempfile.TemporaryDirectory() as tmp, \
mock.patch.object(install, "RESUME_FROM_SIZE", 1024), \
mock.patch("builtins.open", locked_open):
dest, ok, reason = install._download_one(dict(ENTRY), Path(tmp))
self.assertEqual((dest, ok, reason), ("big.bin", True, ""))
def test_a_read_only_file_in_place_is_named_as_such(self):
with _Server(), tempfile.TemporaryDirectory() as tmp, \
mock.patch.object(install.os, "replace",
side_effect=PermissionError(5, "Access is denied")), \
mock.patch.object(install, "_read_only_file", return_value=True):
_dest, ok, reason = install._download_one(dict(ENTRY), Path(tmp))
self.assertFalse(ok)
self.assertEqual(reason, install.READ_ONLY)
def test_a_read_only_file_in_place_is_made_writable_on_windows(self):
with tempfile.TemporaryDirectory() as tmp:
dest = Path(tmp) / "big.bin"
dest.write_bytes(b"old")
os.chmod(dest, 0o444)
with mock.patch.object(install.os, "name", "nt"):
install._clear_read_only(dest)
self.assertTrue(os.access(dest, os.W_OK))
def test_elsewhere_the_attribute_is_left_alone(self):
with tempfile.TemporaryDirectory() as tmp:
dest = Path(tmp) / "big.bin"
dest.write_bytes(b"old")
os.chmod(dest, 0o444)
with mock.patch.object(install.os, "name", "posix"):
install._clear_read_only(dest)
self.assertFalse(os.access(dest, os.W_OK))
os.chmod(dest, 0o644)
def test_a_refused_folder_keeps_its_own_reason(self):
with _Server(), tempfile.TemporaryDirectory() as tmp, \
mock.patch.object(install.os, "replace",
side_effect=PermissionError(13, "Permission denied")), \
mock.patch.object(install, "_read_only_file", return_value=False):
_dest, ok, reason = install._download_one(dict(ENTRY), Path(tmp))
self.assertEqual((ok, reason), (False, install.NOT_WRITABLE))
if __name__ == "__main__":
unittest.main()