//http://www.ecma-international.org/publications/files/ECMA-ST/Ecma-334.pdf section 9.4.4
public static NumberLiteral CreateCSharpNumber(string name)
{
NumberLiteral term = new NumberLiteral(name, TermOptions.EnableQuickParse | TermOptions.SpecialIgnoreCase);
term.DefaultIntTypes = new TypeCode[] { TypeCode.Int32, TypeCode.UInt32, TypeCode.Int64, TypeCode.UInt64 };
term.DefaultFloatType = TypeCode.Double;
term.AddPrefixFlag("0x", ScanFlags.Hex);
term.AddSuffixCodes("u", TypeCode.UInt32, TypeCode.UInt64);
term.AddSuffixCodes("l", TypeCode.Int64, TypeCode.UInt64);
term.AddSuffixCodes("ul", TypeCode.UInt64);
term.AddSuffixCodes("f", TypeCode.Single);
term.AddSuffixCodes("d", TypeCode.Double);
term.AddSuffixCodes("m", TypeCode.Decimal);
return term;
}
public BasicGrammar()
{
#region Initialisation
// BASIC is not case sensitive...
this.CaseSensitive = false;
// By default, new-line characters are ignored. Because BASIC uses
// line breaks to delimit lines, we need to know where the line breaks
// are. The following line is required for this.
this.TokenFilters.Add(new CodeOutlineFilter(false));
// Define the Terminals
Terminal number = new NumberLiteral("NUMBER");
VariableIdentifierTerminal variable = new VariableIdentifierTerminal();
Terminal stringLiteral = new StringLiteral("STRING", "\"", ScanFlags.None);
//Important: do not add comment term to base.NonGrammarTerminals list - we do use this terminal in grammar rules
Terminal comment = new CommentTerminal("Comment", "REM", "\n");
Terminal comma = Symbol(",", "comma");
// Make sure reserved keywords of the BASIC language aren't mistaken
// for variables. Only the keywords ending with '$' could be mistaken
// for variables.
variable.AddKeywords(
"inkey$", "left$", "right$", "mid$", "chr$",
"space$", "str$", "string$"
);
// Define the non-terminals
NonTerminal PROGRAM = new NonTerminal("PROGRAM", typeof(ProgramNode));
NonTerminal LINE = new NonTerminal("LINE", typeof(LineNode));
NonTerminal STATEMENT_LIST = new NonTerminal("STATEMENT_LIST", typeof(StatementListNode));
NonTerminal STATEMENT = new NonTerminal("STATEMENT", typeof(StatementNode));
NonTerminal COMMAND = new NonTerminal("COMMAND", typeof(StatementNode)); //TODO: create command node
NonTerminal PRINT_STMT = new NonTerminal("PRINT_STMT", typeof(PrintStmtNode));
NonTerminal INPUT_STMT = new NonTerminal("INPUT_STMT", typeof(InputStmtNode));
NonTerminal IF_STMT = new NonTerminal("IF_STMT", typeof(IfElseStmtNode)); //TODO: join IfStmtNode and IfElseStmtNode in one
NonTerminal ELSE_CLAUSE_OPT = new NonTerminal("ELSE_CLAUSE_OPT", typeof(GenericJsBasicNode));
NonTerminal EXPR = new NonTerminal("EXPRESSION", typeof(ExpressionNode));
NonTerminal EXPR_LIST = new NonTerminal("EXPRESSION_LIST", typeof(ExprListNode));
NonTerminal BINARY_OP = new NonTerminal("BINARY_OP", typeof(BinaryOpNode));
NonTerminal BINARY_EXPR = new NonTerminal("BINARY_EXPR", typeof(GenericJsBasicNode)); //TODO: create Binary_expr node
NonTerminal BRANCH_STMT = new NonTerminal("BRANCH_STMT", typeof(BranchStmtNode));
NonTerminal ASSIGN_STMT = new NonTerminal("ASSIGN_STMT", typeof(AssignStmtNode));
NonTerminal FOR_STMT = new NonTerminal("FOR_STMT", typeof(ForStmtNode));
NonTerminal STEP_OPT = new NonTerminal("STEP_OPT", typeof(GenericJsBasicNode)); //TODO: create step specifier node
NonTerminal NEXT_STMT = new NonTerminal("NEXT_STMT", typeof(NextStmtNode));
NonTerminal LOCATE_STMT = new NonTerminal("LOCATE_STMT", typeof(LocateStmtNode));
NonTerminal WHILE_STMT = new NonTerminal("WHILE_STMT", typeof(WhileStmtNode));
NonTerminal WEND_STMT = new NonTerminal("WEND_STMT", typeof(WendStmtNode));
NonTerminal SWAP_STMT = new NonTerminal("SWAP_STMT", typeof(SwapStmtNode));
NonTerminal GLOBAL_FUNCTION_EXPR = new NonTerminal("GLOBAL_FUNCTION_EXPR", typeof(GlobalFunctionExpr));
NonTerminal ARG_LIST = new NonTerminal("ARG_LIST", typeof(GenericJsBasicNode));
NonTerminal FUNC_NAME = new NonTerminal("FUNC_NAME", typeof(GenericJsBasicNode));
NonTerminal COMMENT_STMT = new NonTerminal("COMMENT_STMT", typeof(RemStmtNode));
NonTerminal GLOBAL_VAR_EXPR = new NonTerminal("GLOBAL_VAR_EXPR", typeof(GenericJsBasicNode));
// Set the PROGRAM to be the root node of BASIC programs.
// A program is a bunch of lines
this.Root = PROGRAM;
#endregion
#region Grammar declaration
// A program is a collection of lines
PROGRAM.Rule = MakePlusRule(PROGRAM, null, LINE);
// A line can be an empty line, or it's a number followed by a statement list ended by a new-line.
LINE.Rule = NewLine | number + NewLine | number + STATEMENT_LIST + NewLine;
// A statement list is 1 or more statements separated by the ':' character
STATEMENT_LIST.Rule = MakePlusRule(STATEMENT_LIST, Symbol(":"), STATEMENT);
// A statement can be one of a number of types
STATEMENT.Rule = EXPR | ASSIGN_STMT | PRINT_STMT | INPUT_STMT | IF_STMT | COMMENT_STMT
| BRANCH_STMT | COMMAND | FOR_STMT | NEXT_STMT | LOCATE_STMT | SWAP_STMT
| WHILE_STMT | WEND_STMT;
// The different statements are defined here
PRINT_STMT.Rule = "print" + EXPR_LIST;
INPUT_STMT.Rule = "input" + EXPR_LIST + variable;
IF_STMT.Rule = "if" + EXPR + "then" + STATEMENT_LIST + ELSE_CLAUSE_OPT;
ELSE_CLAUSE_OPT.Rule = Empty | "else" + STATEMENT_LIST;
BRANCH_STMT.Rule = "goto" + number | "gosub" + number | "return";
ASSIGN_STMT.Rule = variable + "=" + EXPR;
LOCATE_STMT.Rule = "locate" + EXPR + comma + EXPR;
SWAP_STMT.Rule = "swap" + EXPR + comma + EXPR;
COMMAND.Rule = Symbol("end") | "cls";
COMMENT_STMT.Rule = comment;
// An expression is a number, or a variable, a string, or the result of a binary comparison.
EXPR.Rule = number | variable | stringLiteral | BINARY_EXPR
| GLOBAL_VAR_EXPR | GLOBAL_FUNCTION_EXPR | "(" + EXPR + ")";
BINARY_EXPR.Rule = EXPR + BINARY_OP + EXPR;
BINARY_OP.Rule = Symbol("+") | "-" | "*" | "/" | "=" | "<=" | ">=" | "<" | ">" | "<>" | "and" | "or";
//let's do operator precedence right here
RegisterOperators(50, "*", "/");
RegisterOperators(40, "+", "-");
RegisterOperators(30, "=", "<=", ">=", "<", ">", "<>");
RegisterOperators(20, "and", "or");
// Used by print and input to allow a bunch of expressions separated by whitespace,
// or be empty, for example:
// print
// print "Hi"
// print "Hi " a$
// All of these match "print" EXPR_LIST
EXPR_LIST.Rule = MakeStarRule(EXPR_LIST, null, EXPR);
FOR_STMT.Rule = "for" + ASSIGN_STMT + "to" + EXPR + STEP_OPT;
STEP_OPT.Rule = Empty | "step" + number;
NEXT_STMT.Rule = "next" + variable;
WHILE_STMT.Rule = "while" + EXPR;
WEND_STMT.Rule = "wend";
//TODO: check number of arguments for particular function in node constructor
GLOBAL_FUNCTION_EXPR.Rule = FUNC_NAME + "(" + ARG_LIST + ")";
FUNC_NAME.Rule = Symbol("len") | "left$" | "mid$" | "right$" | "abs" | "asc" | "chr$" | "csrlin$"
| "cvi" | "cvs" | "cvd" | "exp" | "fix" | "log" | "pos" | "sgn" | "sin" | "cos" | "tan"
| "instr" | "space$" | "spc" | "sqr" | "str$" | "string$" | "val" | "cint";
ARG_LIST.Rule = MakePlusRule(ARG_LIST, comma, EXPR);
GLOBAL_VAR_EXPR.Rule = Symbol("rnd") | "timer" | "inkey$" | "csrlin";
// By registering these strings as "punctuation", we exclude them from
// appearing in as nodes in the compiled node tree.
RegisterPunctuation("(", ")", ",");
#endregion
}
//http://www.microsoft.com/downloads/details.aspx?FamilyId=6D50D709-EAA4-44D7-8AF3-E14280403E6E&displaylang=en section 2
public static NumberLiteral CreateVbNumber(string name)
{
NumberLiteral term = new NumberLiteral(name, TermOptions.EnableQuickParse | TermOptions.SpecialIgnoreCase);
term.DefaultIntTypes = new TypeCode[] { TypeCode.Int32, TypeCode.Int64 };
//term.DefaultFloatType = TypeCode.Double; it is default
term.AddPrefixFlag("&H", ScanFlags.Hex);
term.AddPrefixFlag("&O", ScanFlags.Octal);
term.AddSuffixCodes("S", TypeCode.Int16);
term.AddSuffixCodes("I", TypeCode.Int32);
term.AddSuffixCodes("%", TypeCode.Int32);
term.AddSuffixCodes("L", TypeCode.Int64);
term.AddSuffixCodes("&", TypeCode.Int64);
term.AddSuffixCodes("D", TypeCode.Decimal);
term.AddSuffixCodes("@", TypeCode.Decimal);
term.AddSuffixCodes("F", TypeCode.Single);
term.AddSuffixCodes("!", TypeCode.Single);
term.AddSuffixCodes("R", TypeCode.Double);
term.AddSuffixCodes("#", TypeCode.Double);
term.AddSuffixCodes("US", TypeCode.UInt16);
term.AddSuffixCodes("UI", TypeCode.UInt32);
term.AddSuffixCodes("UL", TypeCode.UInt64);
return term;
}
//Note - this is incomplete implementation; need to add functionality to NumberTerminal class to support type detection based
// on exponent symbol.
// From R6RS:
// ... representations of number objects may be written with an exponent marker that indicates the desired precision
// of the inexact representation. The letters s, f, d, and l specify the use of short, single, double, and long precision, respectively.
public static NumberLiteral CreateSchemeNumber(string name)
{
NumberLiteral term = new NumberLiteral(name, TermOptions.EnableQuickParse | TermOptions.SpecialIgnoreCase);
term.DefaultIntTypes = new TypeCode[] { TypeCode.Int32, TypeCode.Int64, NumberLiteral.TypeCodeBigInt };
term.DefaultFloatType = TypeCode.Double; // it is default
term.ExponentSymbols = "sfdl";
term.AddPrefixFlag("#b", ScanFlags.Binary);
term.AddPrefixFlag("#o", ScanFlags.Octal);
term.AddPrefixFlag("#x", ScanFlags.Hex);
term.AddPrefixFlag("#d", ScanFlags.None);
term.AddPrefixFlag("#i", ScanFlags.None); // inexact prefix, has no effect
term.AddPrefixFlag("#e", ScanFlags.None); // exact prefix, has no effect
term.AddSuffixCodes("J", NumberLiteral.TypeCodeImaginary);
return term;
}
//http://docs.python.org/ref/numbers.html
public static NumberLiteral CreatePythonNumber(string name)
{
NumberLiteral term = new NumberLiteral(name, TermOptions.EnableQuickParse | TermOptions.SpecialIgnoreCase | TermOptions.NumberAllowStartEndDot);
//default int types are Integer (32bit) -> LongInteger (BigInt); Try Int64 before BigInt: Better performance?
term.DefaultIntTypes = new TypeCode[] { TypeCode.Int32, TypeCode.Int64, NumberLiteral.TypeCodeBigInt };
// term.DefaultFloatType = TypeCode.Double; -- it is default
//float type is implementation specific, thus try decimal first (higher precision)
//term.DefaultFloatTypes = new TypeCode[] { TypeCode.Decimal, TypeCode.Double };
term.AddPrefixFlag("0x", ScanFlags.Hex);
term.AddPrefixFlag("0", ScanFlags.Octal);
term.AddSuffixCodes("L", TypeCode.Int64, NumberLiteral.TypeCodeBigInt);
term.AddSuffixCodes("J", NumberLiteral.TypeCodeImaginary);
return term;
}
public FlGrammar()
{
// turn off case sensitivity
CaseSensitive = false;
// define all the non-terminals
var program = new NonTerminal("program", typeof(ProgramNode));
var statementList = new NonTerminal("statementList", typeof(StatementNode));
var statement = new NonTerminal("statement", typeof(SkipNode));
var expression = new NonTerminal("expression", typeof(ExpressionNode));
var binaryOperator = new NonTerminal("binaryOperator", typeof(SkipNode));
var variableDeclaration = new NonTerminal("variableDeclaration", typeof(VariableDeclarationNode));
var variableAssignment = new NonTerminal("variableAssignment", typeof(VariableAssignmentNode));
var ifStatement = new NonTerminal("ifStatement", typeof(IfStatementNode));
var elseStatement = new NonTerminal("elseStatement", typeof(ElseStatementNode));
// define all the terminals
var variable = new IdentifierTerminal("variable");
variable.AddKeywords("set", "var" , "to", "if", "freight", "cost", "is", "loop", "through", "order");
var number = new NumberLiteral("number");
var stringLiteral = new StringLiteral("string", "\"", ScanFlags.None);
RegisterPunctuation(";", "[", "]", "(", ")");
//var lpar = new Terminal("(");
//var rpar = new Terminal(")");
//var lbr = new Terminal("{");
//var rbr = new Terminal("}");
// specify the non-terminal which is the root of the AST
//Root = program;
//binaryOperator.Rule = Symbol("+") | "-" | "*" | "/" | "<" | "==" | "!=" | ">" | "<=" | ">=" | "is";
//program.Rule = statementList;
//statementList.Rule = MakeStarRule(statementList, null, statement);
//statement.Rule = variableDeclaration + ";" | variableAssignment + ";" | expression + ";" | ifStatement;
//variableAssignment.Rule = variable + "=" + expression;
//variableDeclaration.Rule = Symbol("var") + variable;
//ifStatement.Rule = Symbol("if") + "(" + expression + ")"
// + "{" + expression + "}"
// + elseStatement;
//elseStatement.Rule = Empty | "else" + "{" + expression + "}";
//expression.Rule = number | variable | stringLiteral
// | expression + binaryOperator + expression
// | "(" + expression + ")";
Root = program;
binaryOperator.Rule = Symbol("+") | "-" | "*" | "/" | "<" | "==" | "!=" | ">" | "<=" | ">=" | "is";
program.Rule = statementList;
statementList.Rule = MakeStarRule(statementList, null, statement);
statement.Rule = variableDeclaration + ";" | variableAssignment + ";" | expression + ";" | ifStatement;
variableAssignment.Rule = variable + "=" + expression;
variableDeclaration.Rule = Symbol("var") + variable;
ifStatement.Rule = "if" + Symbol("(") + expression + Symbol(")")
+ Symbol("{") + statementList + Symbol("}")
+ elseStatement;
elseStatement.Rule = Empty | "else" + Symbol("{") + statementList + Symbol("}");
expression.Rule = number | variable | stringLiteral
| expression + binaryOperator + expression
| "(" + expression + ")";
}
