add #line preprocessor directive

Author: fglock

MILESTONES.md

@@ -235,6 +235,7 @@ The following areas are currently under active development to enhance the functi
 ## Upcoming Milestones
 
 - **v3.0.1**: Next minor version
+  - Added `# line` preprocessor directive.
   - Planned release date: 2025-05-10
 
 ### v4.0.0 Milestone (Planned Release Date: 2026-05-10)

docs/FEATURE_MATRIX.md

@@ -293,7 +293,7 @@ This list is under review.
 - ✅  **`UNITCHECK`**: special block is implemented.
 - ✅  **Labels**: Labels are implemented.
 - ✅  **Here-docs**: Here-docs for multiline string literals are implemented.
-- ❌  **Preprocessor**: `# line` directive is not yet implemented.
+- ✅  **Preprocessor**: `# line` directive is implemented.
 - ✅  **`glob`**: `glob` operator is implemented.
 - ✅  **`<>`**: `<>` operator is implemented.
 - ✅  **`<$fh>`**: `<$fh>` and `<STDIN>` operators are implemented.

src/main/java/org/perlonjava/parser/CoreOperatorResolver.java

@@ -37,7 +37,7 @@ public static Node parseCoreOperator(Parser parser, LexerToken token, int startI
                 return new NumberNode(Integer.toString(parser.ctx.errorUtil.getLineNumber(parser.tokenIndex)), parser.tokenIndex);
             case "__FILE__":
                 // Returns the current file name as a StringNode
-                return new StringNode(parser.ctx.compilerOptions.fileName, parser.tokenIndex);
+                return new StringNode(parser.ctx.errorUtil.getFileName(), parser.tokenIndex);
             case "__PACKAGE__":
                 // Returns the current package name as a StringNode
                 return new StringNode(parser.ctx.symbolTable.getCurrentPackage(), parser.tokenIndex);

src/main/java/org/perlonjava/parser/Whitespace.java

@@ -65,14 +65,17 @@ public static int skipWhitespace(Parser parser, int tokenIndex, List<LexerToken>
 
                 case OPERATOR:
                     if (token.text.equals("#")) {
+                        // # line directive must appear at the beginning of the line
+                        boolean maybeLineDirective = tokenIndex == 0 || tokens.get(tokenIndex - 1).type == LexerTokenType.NEWLINE;
+
                         // Skip optional whitespace after '#'
                         tokenIndex++;
                         while (tokenIndex < tokens.size() && tokens.get(tokenIndex).type == LexerTokenType.WHITESPACE) {
                             tokenIndex++;
                         }
                         // Check if it's a "# line" directive
-                        if (tokenIndex < tokens.size() && tokens.get(tokenIndex).text.equals("line")) {
-                            tokenIndex = parseLineDirective(tokenIndex, tokens);
+                        if (maybeLineDirective && tokenIndex < tokens.size() && tokens.get(tokenIndex).text.equals("line")) {
+                            tokenIndex = parseLineDirective(parser, tokenIndex, tokens);
                         }
                         // Skip comment until end of line
                         while (tokenIndex < tokens.size() && tokens.get(tokenIndex).type != LexerTokenType.NEWLINE) {
@@ -103,7 +106,7 @@ public static int skipWhitespace(Parser parser, int tokenIndex, List<LexerToken>
         return tokenIndex;
     }
 
-    private static int parseLineDirective(int tokenIndex, List<LexerToken> tokens) {
+    private static int parseLineDirective(Parser parser, int tokenIndex, List<LexerToken> tokens) {
         tokenIndex++; // Skip 'line'
         // Skip optional whitespace after 'line'
         while (tokenIndex < tokens.size() && tokens.get(tokenIndex).type == LexerTokenType.WHITESPACE) {
@@ -115,6 +118,11 @@ private static int parseLineDirective(int tokenIndex, List<LexerToken> tokens) {
             try {
                 int lineNumber = Integer.parseInt(lineNumberStr);
                 tokenIndex++;
+
+                // Update the context (ErrorMessageUtil instance) with the new line number
+                parser.ctx.errorUtil.setLineNumber(lineNumber - 1);
+                parser.ctx.errorUtil.setTokenIndex(tokenIndex - 1);
+
                 // Skip optional whitespace before filename
                 while (tokenIndex < tokens.size() && tokens.get(tokenIndex).type == LexerTokenType.WHITESPACE) {
                     tokenIndex++;
@@ -130,8 +138,9 @@ private static int parseLineDirective(int tokenIndex, List<LexerToken> tokens) {
                     if (tokenIndex < tokens.size() && tokens.get(tokenIndex).type == LexerTokenType.OPERATOR && tokens.get(tokenIndex).text.equals("\"")) {
                         tokenIndex++; // Skip closing quote
                         String filename = filenameBuilder.toString();
-                        // Handle the line number and filename as needed
-                        // For example, update a state or context object
+
+                        // Update the context (ErrorMessageUtil instance) with the new file name
+                        parser.ctx.errorUtil.setFileName(filename);
                     }
                 }
             } catch (NumberFormatException e) {

src/main/java/org/perlonjava/runtime/ErrorMessageUtil.java

@@ -13,7 +13,7 @@
  * Utility class for generating error messages with context from a list of tokens.
  */
 public class ErrorMessageUtil {
-    private final String fileName;
+    private String fileName;
     private final List<LexerToken> tokens;
     private int tokenIndex;
     private int lastLineNumber;
@@ -31,6 +31,22 @@ public ErrorMessageUtil(String fileName, List<LexerToken> tokens) {
         this.lastLineNumber = 1;
     }
 
+    public String getFileName() {
+        return fileName;
+    }
+
+    public void setFileName(String fileName) {
+        this.fileName = fileName;
+    }
+
+    public void setLineNumber(int lineNumber) {
+        this.lastLineNumber = lineNumber;
+    }
+
+    public void setTokenIndex(int index) {
+        this.tokenIndex = index;
+    }
+
     /**
      * Quotes the specified string for inclusion in an error message.
      * Escapes special characters such as newlines, tabs, and backslashes.

src/test/resources/line.t

@@ -0,0 +1,42 @@
+use 5.38.0;
+use strict;
+use warnings;
+use Test::More;
+
+# Test #line directive
+
+my $original_file = __FILE__;
+
+my $line;
+my $file;
+
+# line 50
+$line = __LINE__;
+$file = __FILE__;
+is( $line, 50,             "Line number is $line" );
+is( $file, $original_file, "File is original file" );
+
+{
+    # This is an invalid directive because it is indented
+    # line 200
+    ;
+}
+$line = __LINE__;
+$file = __FILE__;
+isnt( $line, 200, "This is a comment; Line number is $line" );
+is( $file, $original_file, "File is original file" );
+
+# line 100 "testfile.pl"
+$line = __LINE__;
+$file = __FILE__;
+is( $line, 100,           "Line number is $line" );
+is( $file, 'testfile.pl', "File is '$file'" );
+
+## # line 150 "testfile2.pl" other comments
+## $line = __LINE__;
+## $file = __FILE__;
+## is($line, 150, "Line number is $line");
+## is($file, 'testfile2.pl', "File is '$file'");
+
+done_testing();
+