fix(providers): update MicrosoftCompilerErrorsProvider for Learn site restructure

- Replace TOC JSON parsing with hardcoded category page paths
- Add trailing slash to DocumentationLink for correct Uri construction
- Parse CS codes from category pages instead of individual error pages
- Handle new documentation structure where errors are grouped by category

Author: Per Kops

src/Atc.CodingRules.AnalyzerProviders/Providers/MicrosoftCompilerErrorsProvider.cs

@@ -13,126 +13,138 @@ public MicrosoftCompilerErrorsProvider(
 
     public static string Name => "Microsoft.CompilerErrors";
 
-    public override Uri? DocumentationLink { get; set; } = new("https://learn.microsoft.com/en-us/dotnet/csharp/language-reference", UriKind.Absolute);
+    public override Uri? DocumentationLink { get; set; } = new("https://learn.microsoft.com/en-us/dotnet/csharp/language-reference/", UriKind.Absolute);
 
     protected override AnalyzerProviderBaseRuleData CreateData()
         => new(Name);
 
-    [SuppressMessage("Design", "MA0051:Method is too long", Justification = "OK.")]
     protected override async Task ReCollect(AnalyzerProviderBaseRuleData data)
     {
         ArgumentNullException.ThrowIfNull(data);
 
-        var web = new HtmlWeb();
-        var htmlDoc = await web
-            .LoadFromWebAsync(DocumentationLink!.AbsoluteUri + "/toc.json")
-            .ConfigureAwait(false);
-
-        if (htmlDoc.DocumentNode.HasTitleWithAccessDenied())
+        var categoryPages = new[]
+        {
+            "compiler-messages/preprocessor-errors",
+            "compiler-messages/attribute-usage-errors",
+            "compiler-messages/feature-version-errors",
+            "compiler-messages/assembly-references",
+            "compiler-messages/constructor-errors",
+            "compiler-messages/overloaded-operator-errors",
+            "compiler-messages/parameter-argument-mismatch",
+            "compiler-messages/generic-type-parameters-errors",
+            "compiler-messages/async-await-errors",
+            "compiler-messages/interface-implementation-errors",
+            "compiler-messages/ref-modifiers-errors",
+            "compiler-messages/ref-safety-errors",
+            "compiler-messages/ref-struct-errors",
+            "compiler-messages/iterator-yield",
+            "compiler-messages/extension-declarations",
+            "compiler-messages/partial-declarations",
+            "compiler-messages/params-arrays",
+            "compiler-messages/nullable-warnings",
+            "compiler-messages/pattern-matching-warnings",
+            "compiler-messages/string-literal",
+            "compiler-messages/array-declaration-errors",
+            "compiler-messages/inline-array-errors",
+            "compiler-messages/lambda-expression-errors",
+            "compiler-messages/overload-resolution",
+            "compiler-messages/expression-tree-restrictions",
+            "compiler-messages/using-directive-errors",
+            "compiler-messages/using-statement-declaration-errors",
+            "compiler-messages/source-generator-errors",
+            "compiler-messages/static-abstract-interfaces",
+            "compiler-messages/lock-semantics",
+            "compiler-messages/dynamic-type-and-binding-errors",
+            "compiler-messages/unsafe-code-errors",
+            "compiler-messages/warning-waves",
+        };
+
+        foreach (var categoryPath in categoryPages)
         {
-            data.ExceptionMessage = "Access Denied";
-            return;
+            var categoryUri = new Uri(DocumentationLink!, categoryPath);
+            var rules = await GetRulesFromCategoryPage(categoryUri.AbsoluteUri);
+            foreach (var rule in rules)
+            {
+                data.Rules.Add(rule);
+            }
         }
+    }
 
-        var jsonDoc = JsonDocument.Parse(htmlDoc.DocumentNode.InnerText);
-        var jsonDocItems = jsonDoc.RootElement
-            .GetProperty("items")
-            .EnumerateArray();
+    [SuppressMessage("Design", "MA0051:Method is too long", Justification = "OK.")]
+    [SuppressMessage("Design", "CA1031:Do not catch general exception types", Justification = "OK - graceful degradation for individual category pages.")]
+    private static async Task<List<Rule>> GetRulesFromCategoryPage(
+        string categoryUrl)
+    {
+        var rules = new List<Rule>();
+        var web = new HtmlWeb();
 
-        while (jsonDocItems.MoveNext())
+        try
         {
-            var jsonElement = jsonDocItems.Current;
+            var htmlDoc = await web
+                .LoadFromWebAsync(categoryUrl)
+                .ConfigureAwait(false);
 
-            if (!jsonElement.GetRawText().Contains("children", StringComparison.Ordinal))
+            if (htmlDoc.DocumentNode.HasTitleWithAccessDenied())
             {
-                continue;
+                return rules;
             }
 
-            var tocTitle = jsonElement
-                .GetProperty("toc_title")
-                .ToString();
-
-            if (!tocTitle.Equals("C# compiler messages", StringComparison.Ordinal))
+            var mainNode = htmlDoc.DocumentNode.SelectSingleNode("//main[@id='main']");
+            if (mainNode is null)
             {
-                continue;
+                return rules;
             }
 
-            var jsonChildItems = jsonElement
-                .GetProperty("children")
-                .EnumerateArray();
+            // Find all strong elements that contain CS error codes
+            var strongNodes = mainNode.SelectNodes(".//strong");
+            if (strongNodes is null)
+            {
+                return rules;
+            }
 
-            while (jsonChildItems.MoveNext())
+            foreach (var strongNode in strongNodes)
             {
-                var jsonChildElement = jsonChildItems.Current;
-                if (jsonChildElement.ValueKind != JsonValueKind.Object ||
-                    !jsonChildElement.TryGetProperty("children", out var jsonChildElement2))
+                var text = strongNode.InnerText.Trim();
+                if (!text.StartsWith("CS", StringComparison.Ordinal))
                 {
                     continue;
                 }
 
-                foreach (var element in jsonChildElement2.EnumerateArray())
+                // Extract just the CS code (e.g., "CS1024" from "CS1024:")
+                var codeTrimmed = text.TrimEnd(':');
+                var code = codeTrimmed.Trim();
+                if (code.Length is < 5 or > 7)
                 {
-                    var hrefPart = element
-                        .GetProperty("href")
-                        .ToString();
-
-                    var code = element
-                        .GetProperty("toc_title")
-                        .ToString();
-
-                    var link = hrefPart.StartsWith("../misc/", StringComparison.Ordinal)
-                        ? "https://docs.microsoft.com/en-us/dotnet/csharp/" + hrefPart.Replace("../", string.Empty, StringComparison.Ordinal)
-                        : DocumentationLink.AbsoluteUri + "/" + hrefPart;
-
-                    var rule = await GetRuleByCode(code, link);
-                    if (rule is not null)
-                    {
-                        data.Rules.Add(rule);
-                    }
+                    continue;
                 }
-            }
-        }
-    }
 
-    private static async Task<Rule?> GetRuleByCode(
-        string code,
-        string link)
-    {
-        var web = new HtmlWeb();
-        var htmlDoc = await web
-            .LoadFromWebAsync(link)
-            .ConfigureAwait(false);
+                // Get the description from the following text
+                var description = string.Empty;
+                var nextSibling = strongNode.NextSibling;
+                if (nextSibling is not null)
+                {
+                    var deEntitized = HtmlEntity.DeEntitize(nextSibling.InnerText);
+                    var trimmed = deEntitized.Trim();
+                    var withoutColon = trimmed.TrimStart(':');
+                    description = withoutColon.Trim();
+                }
 
-        if (htmlDoc.DocumentNode.HasTitleWithAccessDenied())
-        {
-            return null;
-        }
+                var link = categoryUrl + "#" + code.ToLowerInvariant();
 
-        var mainNode = htmlDoc.DocumentNode.SelectSingleNode("//main[@id='main']");
-        if (mainNode is null)
-        {
-            return null;
+                rules.Add(
+                    new Rule(
+                        code,
+                        code,
+                        link,
+                        category: null,
+                        description));
+            }
         }
-
-        var header = mainNode.SelectSingleNode(".//h1");
-
-        var paragraphs = mainNode
-            .SelectNodes(".//p")
-            .ToList();
-
-        if (paragraphs.Count < 2)
+        catch
         {
-            return null;
+            // Ignore errors for individual category pages
         }
 
-        var title = header.InnerText;
-        var description = paragraphs[1].InnerText;
-
-        return new Rule(
-            code,
-            title,
-            link,
-            category: null,
-            description);
+        return rules;
     }
 }