Version 1.4 released!

[Curtis Wensley]

Hi All,

Version 1.4 is now released to nuget! It has some cool new features:

    * Optimize character-only alternations (sets, ranges, and characters) to a single CharacterSet

    * Add w3c ebnf style, which adds character set functionality

    * Add ability to combine various styles for the Ebnf grammar parser

    * Added surrogate pair and range terminals to match a double byte character (thanks to tpluscode's contribution!)

    * Add TagParser to tag a parse tree to control child parsers to require or exclude tags

    * Add Parser.PreventRecursion() fluent method

Enjoy!

Curtis.

[Matthew Lock]

Hi,
I've finally upgraded but I can no longer get my grammar to parse. It used to work okay in 1.3.1:

         var ebnf = new EbnfGrammar(EbnfStyle.BracketComments | EbnfStyle.CommaSeparator | EbnfStyle.SemicolonTerminator | EbnfStyle.SquareBracketAsOptional);

         var grammar = ebnf.Build(@"
            (* := is an extension to define a literal with no whitespace between repeats and sequences *)

            grammar := [ sub_unit_number, ""/"" ], street_number, ws, street_name, ws, street_type;

            sub_unit_number := number;

            street_number := number, [ letter ];
           
            street_name := word, { ws, -street_type, word };

            street_type := ""street"" | ""road"" | ""terrace"";

            word := letter, {letter};
            ws := ? Terminals.WhiteSpace ?, { ? Terminals.WhiteSpace ? };
            number := digit, {digit};
            letter := ? Terminals.Letter ?;
            digit := ? Terminals.Digit ?;            

         ", "grammar");

I now get the following error:

System.FormatException: Error parsing ebnf:
Index=357, Context="rd, { ws, >>>-street_ty"

Expected:

optional sequence
repeated sequence
grouped sequence
meta identifier: Sequence
Quoted String
special sequence
primary
factor
term

[Matthew Lock]

I've managed to get it working with numeric cardinality instead. Does that look right?

       var ebnf = new EbnfGrammar(EbnfStyle.BracketComments | EbnfStyle.CommaSeparator | EbnfStyle.SemicolonTerminator | EbnfStyle.SquareBracketAsOptional | EbnfStyle.NumericCardinality);

         var grammar = ebnf.Build(@"

            (* := is an extension to define a literal with no whitespace between repeats and sequences *)

            grammar := [ sub_unit_number, ""/"" ], street_number, ws, street_name, ws, street_type;

            sub_unit_number := number;

            street_number := number, [ letter ];
           

            street_name := word, { ws, 0 * (street_type), word };

, "grammar");

[Chris Anders]

please make a github release