[.NET] Avoid allocation and improve parsing time (#344)

* [.NET] Change some types from class to struct * [.NET] GherkinLine: remove trimmedLineText * [.NET] GherkinLine.GetTableCells: avoid extra TrimEnd allocation * [.NET] Optimize AstNode.subitems with simple List and custom Enumerator * [.NET] Replace use of arrays with Enumerable * [.NET] GherkinLine: explicit struct Enumerator
cucumber · Feb 5, 2025 · cf4160f · cf4160f
1 parent 5b54a36
commit cf4160f
Show file tree

Hide file tree

Showing 24 changed files with 391 additions and 302 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -26,6 +26,7 @@ This document is formatted according to the principles of [Keep A CHANGELOG](htt
 - [cpp] namespace was changed to 'cucumber::gherkin' to better reflect project structure and prevent clashing
 - [.NET] Removed dependency on System.Text.Json and related logic in GherkinDialectProvider
 - [Elixir] Updates dependencies, bumps messages to 27.0.2
+- [.NET] Changed some types from class to struct, removed IGherkinLine interface and changes some functions from Array to Enumerable
 
 ### Fixed
 - [c] slight update to existing CMakeFiles.txt to propagate VERSION. Close #320 ([#328](https://github.com/cucumber/gherkin/pull/328))

diff --git a/dotnet/Gherkin.Specs/EventStubs/GherkinEventsProvider.cs b/dotnet/Gherkin.Specs/EventStubs/GherkinEventsProvider.cs
@@ -69,7 +69,7 @@ private void AddParseError(List<Envelope> events, ParserException e, String uri)
                 Message = e.Message,
                 Source = new SourceReference()
                 {
-                    Location = new Location(e.Location.Column, e.Location.Line),
+                    Location = e.Location.HasValue ? new Location(e.Location.GetValueOrDefault().Column, e.Location.GetValueOrDefault().Line) : null,
                     Uri = uri
                 }
             }

diff --git a/dotnet/Gherkin.Specs/Tokens/TestTokenFormatter.cs b/dotnet/Gherkin.Specs/Tokens/TestTokenFormatter.cs
@@ -7,7 +7,8 @@ public string FormatToken(Token token)
         if (token.IsEOF)
             return "EOF";
 
-        string stepTypeText;
+        string stepTypeText = string.Empty;
+        string matchedItemsText = null;
         switch (token.MatchedType)
         {
             case TokenType.FeatureLine:
@@ -22,13 +23,14 @@ public string FormatToken(Token token)
                 var tokenType = token.MatchedGherkinDialect.GetStepKeywordType(token.MatchedKeyword);
                 stepTypeText = $"({tokenType})";
                 break;
-            default:
-                stepTypeText = "";
+            case TokenType.TagLine:
+                matchedItemsText = string.Join(",", token.Line.GetTags().Select(i => i.Column + ":" + i.Text));
+                break;
+            case TokenType.TableRow:
+                matchedItemsText = string.Join(",", token.Line.GetTableCells().Select(i => i.Column + ":" + i.Text));
                 break;
         }
 
-        var matchedItemsText = token.MatchedItems == null ? "" : string.Join(",", token.MatchedItems.Select(i => i.Column + ":" + i.Text));
-
         return $"({token.Location.Line}:{token.Location.Column}){token.MatchedType}:{stepTypeText}{token.MatchedKeyword}/{token.MatchedText}/{matchedItemsText}";
     }
 }
diff --git a/dotnet/Gherkin/Ast/Background.cs b/dotnet/Gherkin/Ast/Background.cs
@@ -1,4 +1,4 @@
 namespace Gherkin.Ast;
 
-public class Background(Location location, string keyword, string name, string description, Step[] steps)
+public class Background(Location location, string keyword, string name, string description, IEnumerable<Step> steps)
     : StepsContainer(location, keyword, name, description, steps);
diff --git a/dotnet/Gherkin/Ast/DataTable.cs b/dotnet/Gherkin/Ast/DataTable.cs
@@ -5,10 +5,10 @@ public class DataTable : StepArgument, IHasRows, IHasLocation
     public Location Location { get; private set; }
     public IEnumerable<TableRow> Rows { get; private set; }
 
-    public DataTable(TableRow[] rows)
+    public DataTable(List<TableRow> rows)
     {
         if (rows == null) throw new ArgumentNullException("rows");
-        if (rows.Length == 0) throw new ArgumentException("DataTable must have at least one row", "rows");
+        if (rows.Count == 0) throw new ArgumentException("DataTable must have at least one row", "rows");
 
         Location = rows[0].Location;
         Rows = rows;

diff --git a/dotnet/Gherkin/Ast/Examples.cs b/dotnet/Gherkin/Ast/Examples.cs
@@ -1,6 +1,6 @@
 namespace Gherkin.Ast;
 
-public class Examples(Tag[] tags, Location location, string keyword, string name, string description, TableRow header, TableRow[] body)
+public class Examples(IEnumerable<Tag> tags, Location location, string keyword, string name, string description, TableRow header, IEnumerable<TableRow> body)
     : IHasLocation, IHasDescription, IHasRows, IHasTags
 {
     public IEnumerable<Tag> Tags { get; } = tags;

diff --git a/dotnet/Gherkin/Ast/Feature.cs b/dotnet/Gherkin/Ast/Feature.cs
@@ -1,6 +1,6 @@
 namespace Gherkin.Ast;
 
-public class Feature(Tag[] tags, Location location, string language, string keyword, string name, string description, IHasLocation[] children)
+public class Feature(IEnumerable<Tag> tags, Location location, string language, string keyword, string name, string description, IEnumerable<IHasLocation> children)
     : IHasLocation, IHasDescription, IHasTags, IHasChildren
 {
     public IEnumerable<Tag> Tags { get; } = tags;

diff --git a/dotnet/Gherkin/Ast/GherkinDocument.cs b/dotnet/Gherkin/Ast/GherkinDocument.cs
@@ -1,6 +1,6 @@
 namespace Gherkin.Ast;
 
-public class GherkinDocument(Feature feature, Comment[] comments)
+public class GherkinDocument(Feature feature, IEnumerable<Comment> comments)
 {
     public Feature Feature { get; } = feature;
     public IEnumerable<Comment> Comments { get; } = comments;

diff --git a/dotnet/Gherkin/Ast/Location.cs b/dotnet/Gherkin/Ast/Location.cs
@@ -1,6 +1,6 @@
 namespace Gherkin.Ast;
 
-public class Location(int line = 0, int column = 0)
+public readonly struct Location(int line = 0, int column = 0)
 {
     public int Line { get; } = line;
     public int Column { get; } = column;

diff --git a/dotnet/Gherkin/Ast/Rule.cs b/dotnet/Gherkin/Ast/Rule.cs
@@ -1,6 +1,6 @@
 namespace Gherkin.Ast;
 
-public class Rule(Tag[] tags, Location location, string keyword, string name, string description, IHasLocation[] children)
+public class Rule(IEnumerable<Tag> tags, Location location, string keyword, string name, string description, IEnumerable<IHasLocation> children)
     : IHasLocation, IHasDescription, IHasChildren, IHasTags
 {
     public Location Location { get; } = location;

diff --git a/dotnet/Gherkin/Ast/Scenario.cs b/dotnet/Gherkin/Ast/Scenario.cs
@@ -1,6 +1,6 @@
 namespace Gherkin.Ast;
 
-public class Scenario(Tag[] tags, Location location, string keyword, string name, string description, Step[] steps, Examples[] examples)
+public class Scenario(IEnumerable<Tag> tags, Location location, string keyword, string name, string description, IEnumerable<Step> steps, IEnumerable<Examples> examples)
     : StepsContainer(location, keyword, name, description, steps), IHasTags
 {
     public IEnumerable<Tag> Tags { get; } = tags;

diff --git a/dotnet/Gherkin/Ast/StepsContainer.cs b/dotnet/Gherkin/Ast/StepsContainer.cs
@@ -1,6 +1,6 @@
 namespace Gherkin.Ast;
 
-public abstract class StepsContainer(Location location, string keyword, string name, string description, Step[] steps)
+public abstract class StepsContainer(Location location, string keyword, string name, string description, IEnumerable<Step> steps)
     : IHasLocation, IHasDescription, IHasSteps
 {
     public Location Location { get; } = location;

diff --git a/dotnet/Gherkin/Ast/TableCell.cs b/dotnet/Gherkin/Ast/TableCell.cs
@@ -1,6 +1,6 @@
 namespace Gherkin.Ast;
 
-public class TableCell(Location location, string value) : IHasLocation
+public readonly struct TableCell(Location location, string value) : IHasLocation
 {
     public Location Location { get; } = location;
     public string Value { get; } = value;

diff --git a/dotnet/Gherkin/Ast/TableRow.cs b/dotnet/Gherkin/Ast/TableRow.cs
@@ -1,6 +1,6 @@
 namespace Gherkin.Ast;
 
-public class TableRow(Location location, TableCell[] cells) : IHasLocation
+public class TableRow(Location location, IEnumerable<TableCell> cells) : IHasLocation
 {
     public Location Location { get; } = location;
     public IEnumerable<TableCell> Cells { get; } = cells;

diff --git a/dotnet/Gherkin/AstBuilder.cs b/dotnet/Gherkin/AstBuilder.cs
@@ -92,7 +92,7 @@ private object GetTransformedNode(AstNode node)
 
                     var description = GetDescription(scenarioNode);
                     var steps = GetSteps(scenarioNode);
-                    var examples = scenarioNode.GetItems<Examples>(RuleType.ExamplesDefinition).ToArray();
+                    List<Examples> examples = [.. scenarioNode.GetItems<Examples>(RuleType.ExamplesDefinition)];
                     return CreateScenario(tags, GetLocation(scenarioLine), scenarioLine.MatchedKeyword, scenarioLine.MatchedText, description, steps, examples, node);
                 }
             case RuleType.ExamplesDefinition:
@@ -102,9 +102,9 @@ private object GetTransformedNode(AstNode node)
                     var examplesLine = examplesNode.GetToken(TokenType.ExamplesLine);
                     var description = GetDescription(examplesNode);
 
-                    var allRows = examplesNode.GetSingle<TableRow[]>(RuleType.ExamplesTable);
-                    var header = allRows != null ? allRows.First() : null;
-                    var rows = allRows != null ? allRows.Skip(1).ToArray() : null;
+                    var allRows = examplesNode.GetSingle<List<TableRow>>(RuleType.ExamplesTable);
+                    var header = allRows != null ? allRows[0] : null;
+                    var rows = allRows != null ? allRows.Skip(1).ToList() : null;
                     return CreateExamples(tags, GetLocation(examplesLine), examplesLine.MatchedKeyword, examplesLine.MatchedText, description, header, rows, node);
                 }
             case RuleType.ExamplesTable:
@@ -113,7 +113,7 @@ private object GetTransformedNode(AstNode node)
                 }
             case RuleType.Description:
                 {
-                    var lineTokens = node.GetTokens(TokenType.Other);
+                    IEnumerable<Token> lineTokens = node.GetTokens(TokenType.Other);
 
                     // Trim trailing empty lines
                     lineTokens = lineTokens.Reverse().SkipWhile(t => string.IsNullOrWhiteSpace(t.MatchedText)).Reverse();
@@ -130,16 +130,16 @@ private object GetTransformedNode(AstNode node)
                     var children = new List<IHasLocation>();
                     var background = node.GetSingle<Background>(RuleType.Background);
                     if (background != null)
-                    {
                         children.Add(background);
-                    }
-                    var childrenEnumerable = children.Concat(node.GetItems<IHasLocation>(RuleType.ScenarioDefinition))
-                                                     .Concat(node.GetItems<IHasLocation>(RuleType.Rule));
+                    foreach (var scenarioDefinition in node.GetItems<IHasLocation>(RuleType.ScenarioDefinition))
+                        children.Add(scenarioDefinition);
+                    foreach (var rule in node.GetItems<IHasLocation>(RuleType.Rule))
+                        children.Add(rule);
                     var description = GetDescription(header);
                     if (featureLine.MatchedGherkinDialect == null) return null;
                     var language = featureLine.MatchedGherkinDialect.Language;
 
-                    return CreateFeature(tags, GetLocation(featureLine), language, featureLine.MatchedKeyword, featureLine.MatchedText, description, childrenEnumerable.ToArray(), node);
+                    return CreateFeature(tags, GetLocation(featureLine), language, featureLine.MatchedKeyword, featureLine.MatchedText, description, children, node);
                 }
             case RuleType.Rule:
                 {
@@ -151,14 +151,13 @@ private object GetTransformedNode(AstNode node)
                     var children = new List<IHasLocation>();
                     var background = node.GetSingle<Background>(RuleType.Background);
                     if (background != null)
-                    {
                         children.Add(background);
-                    }
-                    var childrenEnumerable = children.Concat(node.GetItems<IHasLocation>(RuleType.ScenarioDefinition));
+                    foreach (var scenarioDefinition in node.GetItems<IHasLocation>(RuleType.ScenarioDefinition))
+                        children.Add(scenarioDefinition);
                     var description = GetDescription(header);
                     if (ruleLine.MatchedGherkinDialect == null) return null;
 
-                    return CreateRule(tags, GetLocation(ruleLine), ruleLine.MatchedKeyword, ruleLine.MatchedText, description, childrenEnumerable.ToArray(), node);
+                    return CreateRule(tags, GetLocation(ruleLine), ruleLine.MatchedKeyword, ruleLine.MatchedText, description, children, node);
                 }
             case RuleType.GherkinDocument:
                 {
@@ -179,12 +178,12 @@ protected virtual StepKeywordType GetKeywordType(Token stepLine)
         return stepKeywordType.Value;
     }
 
-    protected virtual Background CreateBackground(Location location, string keyword, string name, string description, Step[] steps, AstNode node)
+    protected virtual Background CreateBackground(Location location, string keyword, string name, string description, IEnumerable<Step> steps, AstNode node)
     {
         return new Background(location, keyword, name, description, steps);
     }
 
-    protected virtual DataTable CreateDataTable(TableRow[] rows, AstNode node)
+    protected virtual DataTable CreateDataTable(List<TableRow> rows, AstNode node)
     {
         return new DataTable(rows);
     }
@@ -194,12 +193,12 @@ protected virtual Comment CreateComment(Location location, string text)
         return new Comment(location, text);
     }
 
-    protected virtual Examples CreateExamples(Tag[] tags, Location location, string keyword, string name, string description, TableRow header, TableRow[] body, AstNode node)
+    protected virtual Examples CreateExamples(IEnumerable<Tag> tags, Location location, string keyword, string name, string description, TableRow header, IEnumerable<TableRow> body, AstNode node)
     {
         return new Examples(tags, location, keyword, name, description, header, body);
     }
 
-    protected virtual Scenario CreateScenario(Tag[] tags, Location location, string keyword, string name, string description, Step[] steps, Examples[] examples, AstNode node)
+    protected virtual Scenario CreateScenario(IEnumerable<Tag> tags, Location location, string keyword, string name, string description, IEnumerable<Step> steps, IEnumerable<Examples> examples, AstNode node)
     {
         return new Scenario(tags, location, keyword, name, description, steps, examples);
     }
@@ -214,17 +213,17 @@ protected virtual Step CreateStep(Location location, string keyword, StepKeyword
         return new Step(location, keyword, keywordType, text, argument);
     }
 
-    protected virtual GherkinDocument CreateGherkinDocument(Feature feature, Comment[] gherkinDocumentComments, AstNode node)
+    protected virtual GherkinDocument CreateGherkinDocument(Feature feature, IEnumerable<Comment> gherkinDocumentComments, AstNode node)
     {
         return new GherkinDocument(feature, gherkinDocumentComments);
     }
 
-    protected virtual Feature CreateFeature(Tag[] tags, Location location, string language, string keyword, string name, string description, IHasLocation[] children, AstNode node)
+    protected virtual Feature CreateFeature(IEnumerable<Tag> tags, Location location, string language, string keyword, string name, string description, IEnumerable<IHasLocation> children, AstNode node)
     {
         return new Feature(tags, location, language, keyword, name, description, children);
     }
 
-    protected virtual Rule CreateRule(Tag[] tags, Location location, string keyword, string name, string description, IHasLocation[] children, AstNode node)
+    protected virtual Rule CreateRule(IEnumerable<Tag> tags, Location location, string keyword, string name, string description, IEnumerable<IHasLocation> children, AstNode node)
     {
         return new Rule(tags, location, keyword, name, description, children);
     }
@@ -234,17 +233,17 @@ protected virtual Tag CreateTag(Location location, string name, AstNode node)
         return new Tag(location, name);
     }
 
-    protected virtual Location CreateLocation(int line, int column)
+    protected Location CreateLocation(int line, int column)
     {
         return new Location(line, column);
     }
 
-    protected virtual TableRow CreateTableRow(Location location, TableCell[] cells, AstNode node)
+    protected virtual TableRow CreateTableRow(Location location, IEnumerable<TableCell> cells, AstNode node)
     {
         return new TableRow(location, cells);
     }
 
-    protected virtual TableCell CreateTableCell(Location location, string value)
+    protected TableCell CreateTableCell(Location location, string value)
     {
         return new TableCell(location, value);
     }
@@ -254,60 +253,54 @@ private Location GetLocation(Token token, int column = 0)
         return column == 0 ? token.Location : CreateLocation(token.Location.Line, column);
     }
 
-    private Tag[] GetTags(AstNode node)
+    private IEnumerable<Tag> GetTags(AstNode node)
     {
         var tagsNode = node.GetSingle<AstNode>(RuleType.Tags);
         if (tagsNode == null)
             return [];
 
-        return tagsNode.GetTokens(TokenType.TagLine)
-            .SelectMany(t => t.MatchedItems, (t, tagItem) =>
-                CreateTag(GetLocation(t, tagItem.Column), tagItem.Text, tagsNode))
-            .ToArray();
-    }
-
-    private TableRow[] GetTableRows(AstNode node)
-    {
-        var rows = node.GetTokens(TokenType.TableRow).Select(token => CreateTableRow(GetLocation(token), GetCells(token), node)).ToArray();
-        CheckCellCountConsistency(rows);
-        return rows;
+        var tags = new List<Tag>();
+        foreach (var line in tagsNode.GetTokens(TokenType.TagLine))
+        {
+            foreach (var matchedItem in line.Line.GetTags())
+                tags.Add(CreateTag(GetLocation(line, matchedItem.Column), matchedItem.Text, tagsNode));
+        }
+        return tags;
     }
 
-    protected virtual void CheckCellCountConsistency(TableRow[] rows)
+    private List<TableRow> GetTableRows(AstNode node)
     {
-        if (rows.Length == 0)
-            return;
-
-        int cellCount = rows[0].Cells.Count();
-        for (int i = 1; i < rows.Length; i++)
+        var rows = new List<TableRow>();
+        int cellCount = 0;
+        bool firstRow = true;
+        foreach (var rowToken in node.GetTokens(TokenType.TableRow))
         {
-            var row = rows[i];
-            if (row.Cells.Count() != cellCount)
+            var rowLocation = GetLocation(rowToken);
+            var cells = new List<TableCell>();
+            foreach (var cellItem in rowToken.Line.GetTableCells())
+                cells.Add(CreateTableCell(GetLocation(rowToken, cellItem.Column), cellItem.Text));
+            if (firstRow)
             {
-                HandleAstError("inconsistent cell count within the table", row.Location);
+                cellCount = cells.Count;
+                firstRow = false;
             }
+            else if (cells.Count != cellCount)
+            {
+                HandleAstError("inconsistent cell count within the table", rowLocation);
+            }
+            rows.Add(CreateTableRow(rowLocation, cells, node));
         }
+        return rows;
     }
 
     protected virtual void HandleAstError(string message, Location location)
     {
         throw new AstBuilderException(message, location);
     }
 
-    private TableCell[] GetCells(Token tableRowToken)
-    {
-        var cells = new TableCell[tableRowToken.MatchedItems.Length];
-        for (int i = 0; i < cells.Length; i++)
-        {
-            var cellItem = tableRowToken.MatchedItems[i];
-            cells[i] = CreateTableCell(GetLocation(tableRowToken, cellItem.Column), cellItem.Text);
-        }
-        return cells;
-    }
-
-    private static Step[] GetSteps(AstNode scenarioDefinitionNode)
+    private static List<Step> GetSteps(AstNode scenarioDefinitionNode)
     {
-        return scenarioDefinitionNode.GetItems<Step>(RuleType.Step).ToArray();
+        return [..scenarioDefinitionNode.GetItems<Step>(RuleType.Step)];
     }
 
     private static string GetDescription(AstNode scenarioDefinitionNode)