'{ MATCHES(@"\\n", LS(1)) }? S' will generate '[self matchS:NO];' instead of '[self matchWhitespace:NO];'

[joemcbride]

This sample grammar demonstrates the issue where ParseGenApp creates invalid code. My eol rule is an attempt to get PEGKit to match newline characters. Maybe there's an easier way?

  program
  @before {
    PKTokenizer *t = self.tokenizer;

    // whitespace
    self.silentlyConsumesWhitespace = NO;
    t.whitespaceState.reportsWhitespaceTokens = YES;
  //  self.assembly.preservesWhitespaceTokens = YES;

    [t.symbolState add:@"\\n"];

    // setup comments
    t.commentState.reportsCommentTokens = YES;
    [t.commentState addSingleLineStartMarker:@"//"];
    [t.commentState addMultiLineStartMarker:@"/*" endMarker:@"*/"];
  }
    = eol;

  eol
    = { MATCHES(@"\\n", LS(1)) }? S
    ;

[itod]

Thx!

Fixed in c5ff94b7d194bd90636fd525744025a5067447e4

Here's my test case grammar, it should be instructive:

@before {
    PKTokenizer *t = self.tokenizer;

    // whitespace
    self.silentlyConsumesWhitespace = NO;
    t.whitespaceState.reportsWhitespaceTokens = YES;

    // NOTE: mated `S` (i.e. whitespace) tokens will never be preserved by this parser's assembly, unless you turn on the `preservesWhitespaceTokens` below
    // So by default, it is as if all `S` references were actually defined as `S!`. Not sure I still like this default, but that's how it is for now.
    //self.assembly.preservesWhitespaceTokens = YES;
}

lines = line+;
line  = ~eol* eol; // note the `~` Not unary operator. this means "zero or more NON eol tokens, followed by a single eol token"
eol   = { EQ(@"\\n", LS(1)) }? S;

[joemcbride]

Thank you!