Clean up test and benchmark infrastructure

benchmarks/AngleSharp.ReadOnlyDom.Benchmarks/Support/BenchmarkCorpus.cs

@@ -51,7 +51,13 @@ private static string FindCorpusDirectory()
         var directory = new DirectoryInfo(AppContext.BaseDirectory);
         while (directory is not null)
         {
-            var candidate = Path.Combine(directory.FullName, "tests", "AngleSharp.ReadOnlyDom.Tests", "temp");
+            var candidate = Path.Combine(
+                directory.FullName,
+                "tests",
+                "AngleSharp.ReadOnlyDom.Tests",
+                "TestData",
+                "corpus"
+            );
             if (Directory.Exists(candidate))
             {
                 return candidate;
@@ -61,7 +67,7 @@ private static string FindCorpusDirectory()
         }
 
         throw new DirectoryNotFoundException(
-            "Could not locate the checked-in tests/AngleSharp.ReadOnlyDom.Tests/temp corpus."
+            "Could not locate the checked-in tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus directory."
         );
     }
 }

benchmarks/AngleSharp.ReadOnlyDom.Benchmarks/Support/Utf8TokenSmoke.cs

@@ -16,7 +16,8 @@ public static int Run()
         Directory.CreateDirectory(directory);
         var failures = 0;
         long totalBytes = 0;
-        foreach (var document in BenchmarkCorpus.Load("full"))
+        var documents = BenchmarkCorpus.Load("full");
+        foreach (var document in documents)
         {
             var utf8 = Encoding.UTF8.GetBytes(document.Html);
             totalBytes += utf8.Length;
@@ -48,7 +49,7 @@ public static int Run()
                 Console.WriteLine($"FAIL {document.Name}: {difference}");
             }
         }
-        Console.WriteLine($"Checked 47 documents / {totalBytes:N0} UTF-8 bytes; failures: {failures}.");
+        Console.WriteLine($"Checked {documents.Count} documents / {totalBytes:N0} UTF-8 bytes; failures: {failures}.");
         Console.WriteLine($"Traces: {directory}");
         return failures == 0 ? 0 : 1;
     }

docs/BENCHMARKING.md

@@ -126,7 +126,7 @@ and be removed after the decision.
 
 ## Workload notes
 
-`CorpusBenchmark` uses checked-in snapshots under `tests/AngleSharp.ReadOnlyDom.Tests/temp`. `small` selects five
+`CorpusBenchmark` uses checked-in snapshots under `tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus`. `small` selects five
 representative pages; `full` runs the complete corpus.
 
 `LongSyntheticConstructionBenchmark` generates a deterministic 1.96 MB page with 5,000 irrelevant attribute-heavy

scripts/bench-ab-corpus.ps1

@@ -15,7 +15,7 @@ $root = Split-Path -Parent $PSScriptRoot
 if (-not $CandidateDll) {
     $CandidateDll = Join-Path $root "benchmarks/ProductComparison/AngleSharp.NativeConsole/bin/Release/net10.0/AngleSharp.NativeConsole.dll"
 }
-$corpDir = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/temp"
+$corpDir = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus"
 if (-not $ResultsFile) {
     $timestamp = Get-Date -Format "yyyyMMdd-HHmmss"
     $ResultsFile = Join-Path $root "artifacts/benchmarks/$timestamp-ab-corpus/results.jsonl"

scripts/bench-ab.ps1

@@ -48,7 +48,7 @@ $ErrorActionPreference = "Stop"
 $culture = [Globalization.CultureInfo]::InvariantCulture
 $root = Split-Path -Parent $PSScriptRoot
 $angleProject = Join-Path $root "benchmarks/ProductComparison/AngleSharp.NativeConsole/AngleSharp.NativeConsole.csproj"
-$corpusPath = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/temp/$Corpus"
+$corpusPath = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus/$Corpus"
 if (-not (Test-Path $corpusPath)) { throw "Missing corpus: $corpusPath" }
 if (-not (Test-Path $BaselineDll)) { throw "Missing baseline console: $BaselineDll" }
 

scripts/bench-cross-engine-corpus.ps1

@@ -10,7 +10,7 @@
       codeproject) each contain one <a href> inside a <noscript> element, which this engine
       extracts and lol-html does not: <noscript> is raw text only with scripting enabled, and
       lol-html hardcodes that while this engine follows the scripting-disabled default. Over
-      47 documents a throwing guard just aborts the run, so those rows are marked instead and
+      a full-corpus run a throwing guard just aborts the run, so those rows are marked instead and
       left out of every aggregate.
     - It runs several independent passes and reports run-to-run drift, because a single pass
       cannot distinguish a structural delta from machine state.
@@ -56,7 +56,7 @@ $ErrorActionPreference = "Stop"
 $culture = [Globalization.CultureInfo]::InvariantCulture
 $root = Split-Path -Parent $PSScriptRoot
 $angleProject = Join-Path $root "benchmarks/ProductComparison/AngleSharp.NativeConsole/AngleSharp.NativeConsole.csproj"
-$corpusRoot = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/temp"
+$corpusRoot = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus"
 $extension = if ($IsWindows) { ".exe" } else { "" }
 $timestamp = Get-Date -Format "yyyyMMdd-HHmmss"
 $outputDirectory = Join-Path $root "artifacts/benchmarks/$timestamp-cross-engine-corpus"

scripts/bench-native-console.ps1

@@ -20,7 +20,7 @@ $ErrorActionPreference = "Stop"
 $root = Split-Path -Parent $PSScriptRoot
 $rustManifest = Join-Path $root "benchmarks/ProductComparison/lol-html-server/Cargo.toml"
 $angleProject = Join-Path $root "benchmarks/ProductComparison/AngleSharp.NativeConsole/AngleSharp.NativeConsole.csproj"
-$corpus = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/temp/qq.html"
+$corpus = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus/qq.html"
 $timestamp = Get-Date -Format "yyyyMMdd-HHmmss"
 $reportPath = Join-Path $root "artifacts/benchmarks/$timestamp-native-console/report.md"
 

scripts/bench-product.ps1

@@ -12,7 +12,7 @@ $root = Split-Path -Parent $PSScriptRoot
 $rustManifest = Join-Path $root "benchmarks/ProductComparison/lol-html-server/Cargo.toml"
 $angleProject = Join-Path $root "benchmarks/ProductComparison/AngleSharp.NativeServer/AngleSharp.NativeServer.csproj"
 $runnerProject = Join-Path $root "benchmarks/ProductComparison/LoadRunner/LoadRunner.csproj"
-$corpus = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/temp/qq.html"
+$corpus = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus/qq.html"
 $timestamp = Get-Date -Format "yyyyMMdd-HHmmss"
 $output = Join-Path $root "artifacts/benchmarks/$timestamp-product-comparison/report.md"
 

scripts/bench-sweep.ps1

@@ -49,7 +49,7 @@ $ErrorActionPreference = "Stop"
 $culture = [Globalization.CultureInfo]::InvariantCulture
 $root = Split-Path -Parent $PSScriptRoot
 $angleProject = Join-Path $root "benchmarks/ProductComparison/AngleSharp.NativeConsole/AngleSharp.NativeConsole.csproj"
-$corpusRoot = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/temp"
+$corpusRoot = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus"
 $timestamp = Get-Date -Format "yyyyMMdd-HHmmss"
 $reportPath = Join-Path $root "artifacts/benchmarks/$timestamp-sweep/report.md"
 

scripts/bench.ps1

@@ -24,7 +24,7 @@ $metadata = @(
     "- Runtime: ``$(dotnet --version)``"
     "- GC: Server GC (enforced by the benchmark executable and BenchmarkDotNet job)"
     "- Job: BenchmarkDotNet Default, LaunchCount=1, out-of-process (no in-process emit toolchain)"
-    "- Corpus: checked-in snapshots under tests/AngleSharp.ReadOnlyDom.Tests/temp"
+    "- Corpus: checked-in snapshots under tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus"
     "- Hardware counters: $hardwareCounterNote"
     "- Note: a single launch cannot separate per-process variance from a real effect; allocation results remain the primary micro gate."
 )
@@ -52,6 +52,7 @@ function Invoke-Benchmark([string] $filter, [string] $name, [string] $corpusTier
     }
 }
 
+if ($HardwareCounters) { $env:AS_BENCH_HARDWARE_COUNTERS = "1" }
 if ($Tier -in @("compact", "all")) {
     Invoke-Benchmark "*CompactBuildBenchmark*" "compact"
 }
@@ -67,8 +68,6 @@ if ($Tier -in @("long-streaming", "extraction", "scraping", "all")) {
 if ($Tier -in @("scraping", "extraction", "all")) {
     Invoke-Benchmark "*QqArticleScraperBenchmark*" "qq-scraper"
 }
-if ($HardwareCounters) { $env:AS_BENCH_HARDWARE_COUNTERS = "1" }
-
 if ($Tier -in @("utf8-baseline", "utf8", "all")) {
     Invoke-Benchmark "*Utf8TokenizerBaselineBenchmark*" "utf8-baseline"
     dotnet run --project $project -c Release -f net10.0 --no-build -- `

scripts/collect-pgo.ps1

@@ -57,7 +57,7 @@ $env:DOTNET_TC_QuickJitForLoops = "1"
 $env:DOTNET_JitCollect64BitCounts = "1"
 try {
     foreach ($corpus in $Corpora) {
-        $corpusPath = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/temp/$corpus"
+        $corpusPath = Join-Path $root "tests/AngleSharp.ReadOnlyDom.Tests/TestData/corpus/$corpus"
         if (-not (Test-Path $corpusPath)) { throw "Missing corpus: $corpusPath" }
         foreach ($workload in $Workloads) {
             $stem = "$($corpus -replace '\.html$', '')-$workload"

tests/AngleSharp.ReadOnlyDom.Tests/AngleSharp.ReadOnlyDom.Tests.csproj

@@ -26,7 +26,6 @@
   </ItemGroup>
 
   <ItemGroup>
-    <None Update="TestData\html5lib-tokenizer\**\*" CopyToOutputDirectory="PreserveNewest" />
-    <None Update="temp\*.html" CopyToOutputDirectory="PreserveNewest" />
+    <None Update="TestData\**\*" CopyToOutputDirectory="PreserveNewest" />
   </ItemGroup>
 </Project>

tests/AngleSharp.ReadOnlyDom.Tests/BackpressuredQueryTests.cs

@@ -126,7 +126,6 @@ public async Task SlowOutputStopsInputDrainAndResumesWithoutChangingBytes(bool e
 
         var firstRead = await output.Reader.ReadAsync();
         await Assert.That(execution.IsCompleted).IsFalse();
-        await Task.Delay(10);
 
         var received = new ArrayBufferWriter<byte>();
         Copy(firstRead.Buffer, received);

tests/AngleSharp.ReadOnlyDom.Tests/TestReporting.cs

@@ -13,7 +13,7 @@ internal static void SuppressFileReportsByDefault()
 
     private static void DefaultToDisabled(string variable)
     {
-        if (Environment.GetEnvironmentVariable(variable).IsNullOrWhiteSpace())
+        if (String.IsNullOrWhiteSpace(Environment.GetEnvironmentVariable(variable)))
         {
             Environment.SetEnvironmentVariable(variable, "true");
         }

tests/AngleSharp.ReadOnlyDom.Tests/TopLevelArenaSmoke.cs

@@ -10,7 +10,7 @@ namespace AngleSharp.Readonly.Tests;
 
 internal class TopLevelArenaSmoke
 {
-    private static readonly string BaseDir = Path.Combine(AppContext.BaseDirectory, "temp");
+    private static readonly string BaseDir = Path.Combine(AppContext.BaseDirectory, "TestData", "corpus");
 
     private static readonly ConcurrentDictionary<string, string> FileContents = new();
     private static readonly ConcurrentDictionary<string, IHtmlDocument> ParsedMutableDocs = new();

tests/AngleSharp.ReadOnlyDom.Tests/TopLevelSmoke.cs

@@ -15,7 +15,7 @@ public class TopLevelSmoke
 {
     const int MaxSize = (512 + 128) * 1024;
 
-    private static readonly string BaseDir = Path.Combine(AppContext.BaseDirectory, "temp");
+    private static readonly string BaseDir = Path.Combine(AppContext.BaseDirectory, "TestData", "corpus");
 
     private static readonly ConcurrentDictionary<string, string> FileContents = new();
     private static readonly ConcurrentDictionary<string, IHtmlDocument> ParsedMutableDocs = new();
@@ -303,7 +303,7 @@ private static IEnumerable<SelectorTestCase> GetTestCases(string file, string ta
         };
 
         id =
-            id.IsNullOrWhiteSpace()
+            String.IsNullOrWhiteSpace(id)
             || id is "19ee99feeb254bf99a88146643d1afa2" or "19ee99feeb254bf99a88146643d1afa3"
             || HasBadChar(id.AsSpan())
                 ? null
@@ -468,7 +468,7 @@ private static IEnumerable<SelectorTestCase> DistinctSelectors(IEnumerable<Selec
                 var doc = ParsedMutableDocs.GetOrAdd(fileName, k => parser().ParseDocument(html));
                 return doc
                     .All.Select(it => it.Id)
-                    .Where(id => !id.IsNullOrWhiteSpace() && !HasBadChar(id.AsSpan()))
+                    .Where(id => !String.IsNullOrWhiteSpace(id) && !HasBadChar(id.AsSpan()))
                     .Distinct()
                     .Take(75)
                     .Select(id => (fileName, id!));