wehub-resource-sync
342 lines
13 KiB
Python
Executable File
342 lines
13 KiB
Python
Executable File
#!/usr/bin/env python3
|
|
"""Run SwiftPM tests in suite shards so CI cannot hang inside one aggregate run."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import argparse
|
|
import os
|
|
import re
|
|
import signal
|
|
import subprocess
|
|
import sys
|
|
import time
|
|
from collections.abc import Iterable
|
|
from dataclasses import dataclass
|
|
|
|
|
|
@dataclass(frozen=True)
|
|
class TestSelection:
|
|
name: str
|
|
filter_pattern: str
|
|
suite_name: str | None = None
|
|
|
|
|
|
@dataclass
|
|
class RunStats:
|
|
discovered_selections: int = 0
|
|
selected_selections: int = 0
|
|
selected_groups: int = 0
|
|
group_size: int = 0
|
|
shard_index: int | None = None
|
|
shard_count: int | None = None
|
|
discovery_seconds: float = 0
|
|
execution_seconds: float = 0
|
|
total_seconds: float = 0
|
|
first_pass_successful_groups: int = 0
|
|
first_pass_failed_groups: int = 0
|
|
full_group_retries: int = 0
|
|
timed_out_groups: int = 0
|
|
recovered_groups: int = 0
|
|
isolated_selection_retries: int = 0
|
|
|
|
def summary_rows(self) -> list[tuple[str, str]]:
|
|
shard = "none"
|
|
if self.shard_index is not None and self.shard_count is not None:
|
|
shard = f"{self.shard_index + 1}/{self.shard_count}"
|
|
return [
|
|
("Shard", shard),
|
|
("Group size", str(self.group_size)),
|
|
("Discovered selections", str(self.discovered_selections)),
|
|
("Selected selections", str(self.selected_selections)),
|
|
("Selected groups", str(self.selected_groups)),
|
|
("First-pass successful groups", str(self.first_pass_successful_groups)),
|
|
("First-pass failed groups", str(self.first_pass_failed_groups)),
|
|
("Full-group retries", str(self.full_group_retries)),
|
|
("Recovered groups", str(self.recovered_groups)),
|
|
("Timed out groups", str(self.timed_out_groups)),
|
|
("Isolated selection retries", str(self.isolated_selection_retries)),
|
|
("Discovery seconds", f"{self.discovery_seconds:.1f}"),
|
|
("Execution seconds", f"{self.execution_seconds:.1f}"),
|
|
("Total seconds", f"{self.total_seconds:.1f}"),
|
|
]
|
|
|
|
|
|
def parse_args() -> argparse.Namespace:
|
|
parser = argparse.ArgumentParser()
|
|
parser.add_argument("--group-size", type=int, default=12)
|
|
parser.add_argument("--timeout", type=int, default=180)
|
|
parser.add_argument("--limit-groups", type=int)
|
|
parser.add_argument("--shard-index", type=int)
|
|
parser.add_argument("--shard-count", type=int)
|
|
parser.add_argument(
|
|
"--no-retry-non-timeout-failures",
|
|
action="store_false",
|
|
dest="retry_non_timeout_failures",
|
|
help="fail immediately when a group exits without timing out",
|
|
)
|
|
parser.add_argument("--list-only", action="store_true")
|
|
parser.add_argument("--swift-command", default="swift")
|
|
parser.add_argument("--swift-command-arg", action="append", default=[])
|
|
return parser.parse_args()
|
|
|
|
|
|
def run_command(command: list[str], timeout: int | None = None) -> int:
|
|
print(f"+ {' '.join(command)}", flush=True)
|
|
process = subprocess.Popen(command, start_new_session=True)
|
|
try:
|
|
return process.wait(timeout=timeout)
|
|
except subprocess.TimeoutExpired:
|
|
print(f"::warning::Command timed out after {timeout}s: {' '.join(command)}", flush=True)
|
|
os.killpg(process.pid, signal.SIGTERM)
|
|
try:
|
|
process.wait(timeout=10)
|
|
except subprocess.TimeoutExpired:
|
|
os.killpg(process.pid, signal.SIGKILL)
|
|
process.wait()
|
|
return 124
|
|
|
|
|
|
def swift_test_list(swift_command: list[str]) -> list[TestSelection]:
|
|
command = [*swift_command, "test", "list"]
|
|
try:
|
|
result = subprocess.run(command, check=True, capture_output=True, text=True)
|
|
except subprocess.CalledProcessError as error:
|
|
print(f"+ {swift_command[0]} test list", flush=True)
|
|
if error.stdout:
|
|
print(error.stdout, end="" if error.stdout.endswith("\n") else "\n", flush=True)
|
|
if error.stderr:
|
|
print(error.stderr, end="" if error.stderr.endswith("\n") else "\n", file=sys.stderr, flush=True)
|
|
raise
|
|
selections: set[TestSelection] = set()
|
|
unknown: list[str] = []
|
|
for line in result.stdout.splitlines():
|
|
top_level = re.fullmatch(r"(?P<module>[^.]+)\.(?:`(?P<display>.+)`|(?P<function>[^()/]+))\(\)", line)
|
|
if top_level is not None:
|
|
module = top_level.group("module")
|
|
test_name = top_level.group("display") or top_level.group("function")
|
|
selections.add(
|
|
TestSelection(
|
|
name=line,
|
|
# SwiftPM matches top-level Swift Testing functions by their display name,
|
|
# not the backtick-wrapped identifier printed by `swift test list`.
|
|
filter_pattern=rf"{re.escape(module)}\..*{re.escape(test_name)}",
|
|
)
|
|
)
|
|
continue
|
|
|
|
if "/" in line:
|
|
suite = line.split("/", 1)[0]
|
|
if "." in suite:
|
|
selections.add(
|
|
TestSelection(
|
|
name=suite,
|
|
filter_pattern=rf"^{re.escape(suite)}/",
|
|
suite_name=suite,
|
|
)
|
|
)
|
|
continue
|
|
|
|
unknown.append(line)
|
|
|
|
if unknown:
|
|
rendered = "\n".join(f"- {line}" for line in unknown)
|
|
raise RuntimeError(f"Unrecognized `swift test list` output:\n{rendered}")
|
|
return sorted(selections, key=lambda selection: selection.name)
|
|
|
|
|
|
def append_github_summary(stats: RunStats) -> None:
|
|
summary_path = os.environ.get("GITHUB_STEP_SUMMARY")
|
|
if not summary_path:
|
|
return
|
|
|
|
with open(summary_path, "a", encoding="utf-8") as summary:
|
|
summary.write("### macOS Swift test timing\n\n")
|
|
summary.write("| Field | Value |\n")
|
|
summary.write("| --- | --- |\n")
|
|
for field, value in stats.summary_rows():
|
|
safe_value = value.replace("|", "\\|")
|
|
summary.write(f"| {field} | `{safe_value}` |\n")
|
|
summary.write("\n")
|
|
|
|
|
|
def print_timing_summary(stats: RunStats) -> None:
|
|
print("Swift test timing summary:", flush=True)
|
|
for field, value in stats.summary_rows():
|
|
print(f"- {field}: {value}", flush=True)
|
|
|
|
|
|
def chunks(items: list[TestSelection], size: int) -> Iterable[list[TestSelection]]:
|
|
for index in range(0, len(items), size):
|
|
yield items[index : index + size]
|
|
|
|
|
|
def shard_groups(groups: list[list[TestSelection]], shard_index: int | None, shard_count: int | None) -> list[list[TestSelection]]:
|
|
if shard_index is None and shard_count is None:
|
|
return groups
|
|
if shard_index is None or shard_count is None:
|
|
raise ValueError("--shard-index and --shard-count must be passed together")
|
|
if shard_count < 1:
|
|
raise ValueError("--shard-count must be positive")
|
|
if shard_index < 0 or shard_index >= shard_count:
|
|
raise ValueError("--shard-index must be in the range [0, --shard-count)")
|
|
return [group for index, group in enumerate(groups) if index % shard_count == shard_index]
|
|
|
|
|
|
def prioritized_suites(suites: list[TestSelection]) -> list[TestSelection]:
|
|
priority = ["CodexBarTests.CLIEntryTests"]
|
|
ordered = [suite for name in priority for suite in suites if suite.suite_name == name]
|
|
ordered.extend(suite for suite in suites if suite.suite_name not in priority)
|
|
return ordered
|
|
|
|
|
|
def filtered_suites_for_environment(suites: list[TestSelection]) -> list[TestSelection]:
|
|
if os.environ.get("GITHUB_ACTIONS") != "true" or sys.platform != "darwin":
|
|
return suites
|
|
|
|
# SwiftPM hangs before suite output for this executable-target suite on the Intel macOS runner.
|
|
# Linux CI still runs it in the full Swift test lane, and local macOS runs it directly.
|
|
skipped = {"CodexBarTests.CLIEntryTests"}
|
|
filtered = [suite for suite in suites if suite.suite_name not in skipped]
|
|
if len(filtered) != len(suites):
|
|
print(f"Skipping macOS CI-only suites: {', '.join(sorted(skipped))}", flush=True)
|
|
return filtered
|
|
|
|
|
|
def filter_for(suites: list[TestSelection]) -> str:
|
|
return rf"({'|'.join(suite.filter_pattern for suite in suites)})"
|
|
|
|
|
|
def run_group(suites: list[TestSelection], timeout: int, swift_command: list[str]) -> int:
|
|
return run_command(
|
|
[*swift_command, "test", "--skip-build", "--no-parallel", "--filter", filter_for(suites)],
|
|
timeout=timeout,
|
|
)
|
|
|
|
|
|
def retry_selections_individually(
|
|
suites: list[TestSelection],
|
|
timeout: int,
|
|
swift_command: list[str],
|
|
stats: RunStats,
|
|
) -> int:
|
|
for suite in suites:
|
|
stats.isolated_selection_retries += 1
|
|
print(f"::group::Swift test retry {suite.name}", flush=True)
|
|
retry_result = run_group([suite], timeout, swift_command)
|
|
print("::endgroup::", flush=True)
|
|
if retry_result != 0:
|
|
return retry_result
|
|
return 0
|
|
|
|
|
|
def main() -> int:
|
|
total_started = time.monotonic()
|
|
args = parse_args()
|
|
stats = RunStats(
|
|
group_size=args.group_size,
|
|
shard_index=args.shard_index,
|
|
shard_count=args.shard_count,
|
|
)
|
|
if args.group_size < 1:
|
|
print("--group-size must be positive", file=sys.stderr)
|
|
return 2
|
|
|
|
swift_command = [args.swift_command, *args.swift_command_arg]
|
|
result = 0
|
|
try:
|
|
discovery_started = time.monotonic()
|
|
try:
|
|
suites = prioritized_suites(filtered_suites_for_environment(swift_test_list(swift_command)))
|
|
finally:
|
|
stats.discovery_seconds = time.monotonic() - discovery_started
|
|
stats.discovered_selections = len(suites)
|
|
|
|
suite_groups = list(chunks(suites, args.group_size))
|
|
try:
|
|
suite_groups = shard_groups(suite_groups, args.shard_index, args.shard_count)
|
|
except ValueError as error:
|
|
print(str(error), file=sys.stderr)
|
|
result = 2
|
|
return result
|
|
if args.limit_groups is not None:
|
|
suite_groups = suite_groups[: args.limit_groups]
|
|
stats.selected_selections = sum(len(group) for group in suite_groups)
|
|
stats.selected_groups = len(suite_groups)
|
|
|
|
shard_suffix = ""
|
|
if args.shard_index is not None and args.shard_count is not None:
|
|
shard_suffix = f" in shard {args.shard_index + 1}/{args.shard_count}"
|
|
print(
|
|
f"Discovered {len(suites)} test selections; running {stats.selected_selections} selections "
|
|
f"in {len(suite_groups)} groups{shard_suffix}",
|
|
flush=True,
|
|
)
|
|
if args.list_only:
|
|
for group in suite_groups:
|
|
for suite in group:
|
|
print(suite.name)
|
|
return 0
|
|
|
|
if not suite_groups:
|
|
print("No test groups selected.", flush=True)
|
|
return 0
|
|
|
|
execution_started = time.monotonic()
|
|
for group_index, group in enumerate(suite_groups, start=1):
|
|
print(
|
|
f"::group::Swift test group {group_index}/{len(suite_groups)} "
|
|
f"({len(group)} selections)",
|
|
flush=True,
|
|
)
|
|
group_result = run_group(group, args.timeout, swift_command)
|
|
print("::endgroup::", flush=True)
|
|
if group_result == 0:
|
|
stats.first_pass_successful_groups += 1
|
|
continue
|
|
|
|
stats.first_pass_failed_groups += 1
|
|
group_timed_out = group_result == 124
|
|
if group_timed_out:
|
|
stats.timed_out_groups += 1
|
|
if len(group) == 1:
|
|
result = group_result
|
|
return result
|
|
|
|
if group_result != 124:
|
|
if not args.retry_non_timeout_failures:
|
|
result = group_result
|
|
return result
|
|
|
|
stats.full_group_retries += 1
|
|
print(f"Group {group_index} failed with exit code {group_result}; retrying group once", flush=True)
|
|
retry_result = run_group(group, args.timeout, swift_command)
|
|
if retry_result == 0:
|
|
stats.recovered_groups += 1
|
|
continue
|
|
if retry_result != 124:
|
|
result = retry_result
|
|
return result
|
|
group_timed_out = True
|
|
stats.timed_out_groups += 1
|
|
|
|
print(f"Group {group_index} timed out; retrying selections one at a time", flush=True)
|
|
retry_result = retry_selections_individually(group, args.timeout, swift_command, stats)
|
|
if retry_result != 0:
|
|
result = retry_result
|
|
return result
|
|
if group_timed_out:
|
|
stats.recovered_groups += 1
|
|
|
|
return result
|
|
finally:
|
|
stats.total_seconds = time.monotonic() - total_started
|
|
if "execution_started" in locals():
|
|
stats.execution_seconds = time.monotonic() - execution_started
|
|
if not args.list_only:
|
|
print_timing_summary(stats)
|
|
append_github_summary(stats)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
raise SystemExit(main())
|