cucumber · mpkorstanje · Mar 30, 2025 · Mar 30, 2025 · Mar 30, 2025 · Mar 30, 2025
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -8,6 +8,7 @@ This document is formatted according to the principles of [Keep A CHANGELOG](htt
 
 ## [Unreleased]
 ### Changed
+- [Java] Optimize Location performance ([#385](https://github.com/cucumber/gherkin/pull/385))
 - [Java] Optimize AstNode performance ([#383](https://github.com/cucumber/gherkin/pull/383))
 - [Java] Optimize EncodingParser performance ([#382](https://github.com/cucumber/gherkin/pull/382))
 - [Java] Optimize GherkinDialect performance ([#380](https://github.com/cucumber/gherkin/pull/380))

diff --git a/java/gherkin-java.razor b/java/gherkin-java.razor
@@ -128,7 +128,7 @@ class @Model.ParserClassName<T> {
 
         endRule(context, [email protected]);
 
-        if (context.errors.size() > 0) {
+        if (!context.errors.isEmpty()) {
             throw new ParserException.CompositeParserException(context.errors);
         }
 

diff --git a/java/src/main/java/io/cucumber/gherkin/AstNode.java b/java/src/main/java/io/cucumber/gherkin/AstNode.java
@@ -8,6 +8,7 @@
 
 import static io.cucumber.gherkin.Parser.RuleType;
 import static io.cucumber.gherkin.Parser.TokenType;
+import static java.util.Objects.requireNonNull;
 
 class AstNode {
     // subItems is relatively sparse, so pre-initializing all values with empty
@@ -28,6 +29,7 @@ void add(RuleType ruleType, Object obj) {
         items.add(obj);
     }
 
+    @SuppressWarnings("unchecked")
     <T> T getSingle(RuleType ruleType, T defaultResult) {
         // if not null, then at least one item is present because
         // the list was created in add(), so no need to check isEmpty()
@@ -45,7 +47,7 @@ <T> List<T> getItems(RuleType ruleType) {
     }
 
     Token getToken(TokenType tokenType) {
-        return getSingle(tokenType.ruleType, new Token(null, null));
+        return requireNonNull(getSingle(tokenType.ruleType, null));
     }
 
     List<Token> getTokens(TokenType tokenType) {

diff --git a/java/src/main/java/io/cucumber/gherkin/GherkinDocumentBuilder.java b/java/src/main/java/io/cucumber/gherkin/GherkinDocumentBuilder.java
@@ -24,6 +24,7 @@
 import java.util.List;
 import java.util.stream.Collectors;
 
+import static io.cucumber.gherkin.Locations.atColumn;
 import static io.cucumber.gherkin.Parser.Builder;
 import static io.cucumber.gherkin.Parser.RuleType;
 import static io.cucumber.gherkin.Parser.TokenType;
@@ -54,7 +55,7 @@ private AstNode currentNode() {
     @Override
     public void build(Token token) {
         if (token.matchedType == TokenType.Comment) {
-            Comment comment = new Comment(getLocation(token, 0), token.matchedText);
+            Comment comment = new Comment(token.location, token.matchedText);
             comments.add(comment);
         } else {
             currentNode().add(token.matchedType.ruleType, token);
@@ -78,7 +79,7 @@ private Object getTransformedNode(AstNode node) {
             case Step: {
                 Token stepLine = node.getToken(TokenType.StepLine);
                 return new Step(
-                        getLocation(stepLine, 0),
+                        stepLine.location,
                         stepLine.matchedKeyword,
                         stepLine.keywordType,
                         stepLine.matchedText,
@@ -100,7 +101,7 @@ private Object getTransformedNode(AstNode node) {
                 }
 
                 return new DocString(
-                        getLocation(separatorToken, 0),
+                        separatorToken.location,
                         mediaType,
                         content.toString(),
                         separatorToken.matchedKeyword
@@ -113,7 +114,7 @@ private Object getTransformedNode(AstNode node) {
             case Background: {
                 Token backgroundLine = node.getToken(TokenType.BackgroundLine);
                 return new Background(
-                        getLocation(backgroundLine, 0),
+                        backgroundLine.location,
                         backgroundLine.matchedKeyword,
                         backgroundLine.matchedText,
                         getDescription(node),
@@ -126,7 +127,7 @@ private Object getTransformedNode(AstNode node) {
                 Token scenarioLine = scenarioNode.getToken(TokenType.ScenarioLine);
 
                 return new Scenario(
-                        getLocation(scenarioLine, 0),
+                        scenarioLine.location,
                         getTags(node),
                         scenarioLine.matchedKeyword,
                         scenarioLine.matchedText,
@@ -144,7 +145,7 @@ private Object getTransformedNode(AstNode node) {
                 List<TableRow> tableBody = rows != null && !rows.isEmpty() ? rows.subList(1, rows.size()) : Collections.emptyList();
 
                 return new Examples(
-                        getLocation(examplesLine, 0),
+                        examplesLine.location,
                         getTags(node),
                         examplesLine.matchedKeyword,
                         examplesLine.matchedText,
@@ -195,7 +196,7 @@ private Object getTransformedNode(AstNode node) {
                 String language = featureLine.matchedGherkinDialect.getLanguage();
 
                 return new Feature(
-                        getLocation(featureLine, 0),
+                        featureLine.location,
                         tags,
                         language,
                         featureLine.matchedKeyword,
@@ -223,7 +224,7 @@ private Object getTransformedNode(AstNode node) {
                 }
 
                 return new Rule(
-                        getLocation(ruleLine, 0),
+                        ruleLine.location,
                         tags,
                         ruleLine.matchedKeyword,
                         ruleLine.matchedText,
@@ -247,7 +248,7 @@ private List<TableRow> getTableRows(AstNode node) {
         List<TableRow> rows = new ArrayList<>(tokens.size());
 
         for (Token token : tokens) {
-            rows.add(new TableRow(getLocation(token, 0), getCells(token), idGenerator.newId()));
+            rows.add(new TableRow(token.location, getCells(token), idGenerator.newId()));
         }
         ensureCellCount(rows);
         return rows;
@@ -260,21 +261,16 @@ private void ensureCellCount(List<TableRow> rows) {
         int cellCount = rows.get(0).getCells().size();
         for (TableRow row : rows) {
             if (row.getCells().size() != cellCount) {
-                io.cucumber.messages.types.Location rowLocation = row.getLocation();
-                io.cucumber.gherkin.Location location = new io.cucumber.gherkin.Location(
-                        rowLocation.getLine().intValue(),
-                        rowLocation.getColumn().orElse(0L).intValue()
-                );
-                throw new ParserException.AstBuilderException("inconsistent cell count within the table", location);
+                throw new ParserException.AstBuilderException("inconsistent cell count within the table", row.getLocation());
             }
         }
     }
 
     private List<TableCell> getCells(Token token) {
-        List<TableCell> cells = new ArrayList<>();
+        List<TableCell> cells = new ArrayList<>(token.matchedItems.size());
         for (GherkinLineSpan cellItem : token.matchedItems) {
             TableCell tableCell = new TableCell(
-                    getLocation(token, cellItem.column),
+                    atColumn(token.location, cellItem.column),
                     cellItem.text
             );
             cells.add(tableCell);
@@ -286,25 +282,20 @@ private List<Step> getSteps(AstNode node) {
         return node.getItems(RuleType.Step);
     }
 
-    private io.cucumber.messages.types.Location getLocation(Token token, int column) {
-        column = column == 0 ? token.location.getColumn() : column;
-        return new io.cucumber.messages.types.Location((long) token.location.getLine(), (long) column);
-    }
-
     private String getDescription(AstNode node) {
         return node.getSingle(RuleType.Description, "");
     }
 
     private List<Tag> getTags(AstNode node) {
         AstNode tagsNode = node.getSingle(RuleType.Tags, new AstNode(RuleType.None));
         if (tagsNode == null)
-            return new ArrayList<>();
+            return Collections.emptyList();
 
         List<Token> tokens = tagsNode.getTokens(TokenType.TagLine);
         List<Tag> tags = new ArrayList<>();
         for (Token token : tokens) {
             for (GherkinLineSpan tagItem : token.matchedItems) {
-                tags.add(new Tag(getLocation(token, tagItem.column), tagItem.text, idGenerator.newId()));
+                tags.add(new Tag(atColumn(token.location, tagItem.column), tagItem.text, idGenerator.newId()));
             }
         }
         return tags;

diff --git a/java/src/main/java/io/cucumber/gherkin/GherkinLine.java b/java/src/main/java/io/cucumber/gherkin/GherkinLine.java
@@ -1,5 +1,7 @@
 package io.cucumber.gherkin;
 
+import io.cucumber.messages.types.Location;
+
 import java.util.ArrayList;
 import java.util.List;
 import java.util.PrimitiveIterator;
@@ -12,19 +14,20 @@
 import static io.cucumber.gherkin.StringUtils.rtrimKeepNewLines;
 import static io.cucumber.gherkin.StringUtils.symbolCount;
 import static io.cucumber.gherkin.StringUtils.trim;
+import static java.util.Objects.requireNonNull;
 
 class GherkinLine implements IGherkinLine {
     // TODO: set this to 0 when/if we change to 0-indexed columns
     private static final int OFFSET = 1;
     private final String lineText;
     private final String trimmedLineText;
     private final int indent;
-    private final int line;
+    private final Location line;
 
-    public GherkinLine(String lineText, int line) {
-        this.lineText = lineText;
+    public GherkinLine(String lineText, Location line) {
+        this.lineText = requireNonNull(lineText);
         this.trimmedLineText = trim(lineText);
-        this.line = line;
+        this.line = requireNonNull(line);
         indent = symbolCount(lineText) - symbolCount(ltrim(lineText));
     }
 
@@ -47,7 +50,7 @@ public String getLineText(int indentToRemove) {
 
     @Override
     public boolean isEmpty() {
-        return trimmedLineText.length() == 0;
+        return trimmedLineText.isEmpty();
     }
 
     @Override
@@ -76,7 +79,7 @@ public List<GherkinLineSpan> getTags() {
             int symbolLength = uncommentedLine.codePointCount(0, indexInUncommentedLine);
             int column = indent() + symbolLength + 1;
             if (!token.matches("^\\S+$")) {
-                throw new ParserException("A tag may not contain whitespace", new Location(line, column));
+                throw new ParserException("A tag may not contain whitespace", Locations.atColumn(line, column));
             }
             tags.add(new GherkinLineSpan(column, TAG_PREFIX + token));
             indexInUncommentedLine += element.length() + 1;
@@ -142,8 +145,7 @@ public boolean startsWithTitleKeyword(String text) {
         int textLength = text.length();
         return trimmedLineText.length() > textLength &&
                 trimmedLineText.startsWith(text) &&
-                trimmedLineText.substring(textLength, textLength + GherkinLanguageConstants.TITLE_KEYWORD_SEPARATOR.length())
-                        .equals(GherkinLanguageConstants.TITLE_KEYWORD_SEPARATOR);
+                trimmedLineText.startsWith(GherkinLanguageConstants.TITLE_KEYWORD_SEPARATOR, textLength);
     }
 
 }
diff --git a/java/src/main/java/io/cucumber/gherkin/GherkinParser.java b/java/src/main/java/io/cucumber/gherkin/GherkinParser.java
@@ -149,19 +149,12 @@ private List<Envelope> parse(String uri, String data) {
     }
 
     private Envelope createParseError(ParserException e, String uri) {
-        long line = e.location.getLine();
-        long column = e.location.getColumn();
         return Envelope.of(new ParseError(
                 new SourceReference(
                         uri,
                         null,
                         null,
-                        // We want 0 values not to be serialised, which is why we set them to null
-                        // This is a legacy requirement brought over from old protobuf behaviour
-                        new io.cucumber.messages.types.Location(
-                                line,
-                                column == 0 ? null : column
-                        )
+                        e.location
                 ),
                 e.getMessage()
         ));

diff --git a/java/src/main/java/io/cucumber/gherkin/Location.java b/java/src/main/java/io/cucumber/gherkin/Location.java
diff --git a/java/src/main/java/io/cucumber/gherkin/Locations.java b/java/src/main/java/io/cucumber/gherkin/Locations.java
@@ -0,0 +1,39 @@
+package io.cucumber.gherkin;
+
+import io.cucumber.messages.types.Location;
+
+import static java.util.Objects.requireNonNull;
+
+class Locations {
+
+    /**
+     * Cache of Long objects for the range 0-3999. This is used
+     * to avoid creating a huge amount of Long objects in getLocation().
+     * We can't use Long.valueOf() because it caches only the first 128
+     * values, and typical feature files have much more lines.
+     */
+    private final static Long[] longs = new Long[4000];
+    static {
+        for (int i = 0; i < longs.length; i++) {
+            longs[i] = (long) i;
+        }
+    }
+
+    private static Long getLong(int i) {
+        if (i>=longs.length) return (long) i;
+        return longs[i];
+    }
+
+    static Location atColumn(Location location, int column) {
+        requireNonNull(location);
+        if (column <= 0) {
+            throw new IllegalArgumentException("Columns are index-1 based");
+        }
+        return new Location(location.getLine(), getLong(column));
+    }
+
+    static Location atLine(int line) {
+        return new Location(getLong(line), null);
+    }
+
+}
diff --git a/java/src/main/java/io/cucumber/gherkin/Parser.java b/java/src/main/java/io/cucumber/gherkin/Parser.java
@@ -141,7 +141,7 @@ T parse(ITokenScanner tokenScanner, ITokenMatcher tokenMatcher, String uri) {
 
         endRule(context, RuleType.GherkinDocument);
 
-        if (context.errors.size() > 0) {
+        if (!context.errors.isEmpty()) {
             throw new ParserException.CompositeParserException(context.errors);
         }