fix(siteruntime): injectable ScriptExecutionScheduler + un-leakable expression-fault test

Gitea #18: the process-wide ScriptExecutionScheduler was reached via a static
singleton (Shared) directly inside ScriptActor, ScriptExecutionActor, AlarmActor,
AlarmExecutionActor, and ScriptSchedulerStatsReporter, so it could not be
substituted — a shared mutable global for anything running more than one logical
site in a process, most visibly the test assembly. Add an optional scheduler
injection seam to all five: the Host injects nothing and gets the shared pool
(byte-for-byte unchanged), while a test (or a future multi-site host) hands an
actor its own instance. Spawning actors thread their scheduler down to the
execution children they create. Shared() now also recreates a disposed cached
instance rather than returning it (new IsDisposed guard), so a disposed scheduler
can never silently poison every later script/alarm execution.

Gitea #16: ScriptActorTests now runs on its OWN scheduler (disposed at class
teardown), so the faulted-task test can no longer strand a worker on the
process-wide pool and starve unrelated test classes (one prior run failed 22
tests). EvalGate's wait is bounded to 30 s (was unbounded — the actual leak
mechanism) and reset per test; the teardown releases one permit per worker
instead of Release(Entries), which under-released in exactly the failure case.
Full SiteRuntime suite: 6/6 runs green (524/524); was 3/6 failing on clean main.

Fixes: #18
Fixes: #16
This commit was merged in pull request #21.
This commit is contained in:
Joseph Doherty
2026-07-17 15:36:39 -04:00
parent 3e84eee195
commit b1b4874090
9 changed files with 205 additions and 37 deletions
@@ -32,20 +32,30 @@ public sealed class ScriptExecutionScheduler : TaskScheduler, IDisposable
private static readonly object SharedLock = new();
/// <summary>
/// The process-wide script-execution scheduler. Lazily created on first use with the
/// thread count from <see cref="SiteRuntimeOptions.ScriptExecutionThreadCount"/>; the
/// first caller wins, subsequent calls reuse the existing instance.
/// The process-wide script-execution scheduler, used as the default when no scheduler
/// is injected. Lazily created on first use with the thread count from
/// <see cref="SiteRuntimeOptions.ScriptExecutionThreadCount"/>; the first caller wins,
/// subsequent calls reuse the existing instance.
///
/// If the cached instance has been disposed it is recreated rather than handed back:
/// a disposed scheduler can execute no work, so returning it would silently poison
/// every subsequent script and alarm execution for the rest of the process. Nothing
/// disposes the shared instance today, but the guard keeps that latent failure mode
/// closed regardless of a future shutdown path or a multi-site host.
/// </summary>
/// <param name="options">Site runtime options supplying the thread count for the scheduler.</param>
/// <returns>The process-wide <see cref="ScriptExecutionScheduler"/> instance, creating it on first call.</returns>
/// <returns>A live process-wide <see cref="ScriptExecutionScheduler"/> instance, creating it on first call.</returns>
public static ScriptExecutionScheduler Shared(SiteRuntimeOptions options)
{
if (_shared != null)
return _shared;
var existing = _shared;
if (existing is { IsDisposed: false })
return existing;
lock (SharedLock)
{
return _shared ??= new ScriptExecutionScheduler(options.ScriptExecutionThreadCount);
if (_shared is null || _shared.IsDisposed)
_shared = new ScriptExecutionScheduler(options.ScriptExecutionThreadCount);
return _shared;
}
}
@@ -73,6 +83,10 @@ public sealed class ScriptExecutionScheduler : TaskScheduler, IDisposable
}
}
/// <summary><see langword="true"/> once <see cref="Dispose"/> has run; a disposed scheduler
/// can execute no further work and must not be handed out by <see cref="Shared"/>.</summary>
public bool IsDisposed => Volatile.Read(ref _disposed) != 0;
/// <inheritdoc />
public override int MaximumConcurrencyLevel => _threads.Count;
@@ -24,21 +24,30 @@ public sealed class ScriptSchedulerStatsReporter : BackgroundService
private readonly ILogger<ScriptSchedulerStatsReporter> _logger;
private readonly TimeSpan _pollInterval;
/// <summary>
/// Script-execution scheduler seam (#18): the process-wide scheduler when null, or
/// an injected instance so the reporter samples the same pool the actors run on.
/// </summary>
private readonly ScriptExecutionScheduler? _scheduler;
/// <summary>Initializes a new instance of <see cref="ScriptSchedulerStatsReporter"/>.</summary>
/// <param name="collector">The site health collector that receives the scheduler gauges.</param>
/// <param name="options">Site runtime options supplying the shared scheduler's thread count.</param>
/// <param name="logger">Logger instance.</param>
/// <param name="pollInterval">Poll interval override; defaults to <see cref="DefaultPollInterval"/> (10 s).</param>
/// <param name="scheduler">Optional script-execution scheduler override (#18); null samples the process-wide shared scheduler.</param>
public ScriptSchedulerStatsReporter(
ISiteHealthCollector collector,
SiteRuntimeOptions options,
ILogger<ScriptSchedulerStatsReporter> logger,
TimeSpan? pollInterval = null)
TimeSpan? pollInterval = null,
ScriptExecutionScheduler? scheduler = null)
{
_collector = collector ?? throw new ArgumentNullException(nameof(collector));
_options = options ?? throw new ArgumentNullException(nameof(options));
_logger = logger ?? throw new ArgumentNullException(nameof(logger));
_pollInterval = pollInterval ?? DefaultPollInterval;
_scheduler = scheduler;
}
/// <inheritdoc />
@@ -67,7 +76,7 @@ public sealed class ScriptSchedulerStatsReporter : BackgroundService
{
try
{
var scheduler = ScriptExecutionScheduler.Shared(_options);
var scheduler = _scheduler ?? ScriptExecutionScheduler.Shared(_options);
_collector.SetScriptSchedulerStats(
scheduler.QueueDepth,
scheduler.BusyThreadCount,