public JsonGrammar()
{
//Terminals
var jstring = new StringLiteral("string", "\"");
var jnumber = new NumberLiteral("number");
var comma = Symbol(",");
//Nonterminals
var jobject = new NonTerminal("Object");
var jarray = new NonTerminal("Array");
var jvalue = new NonTerminal("Value");
var jprop = new NonTerminal("Property");
var jproplist = new NonTerminal("PropertyList");
var jlist = new NonTerminal("List");
//Rules
jvalue.Rule = jstring | jnumber | jobject | jarray | "true" | "false" | "null";
jobject.Rule = "{" + jproplist + "}";
jproplist.Rule = MakeStarRule(jproplist, comma, jprop);
jprop.Rule = jstring + ":" + jvalue;
jarray.Rule = "[" + jlist + "]";
jlist.Rule = MakeStarRule(jlist, comma, jvalue);
//Set grammar root
this.Root = jvalue;
RegisterPunctuation("{", "}", "[", "]", ":", ",");
this.MarkTransient(jvalue, jlist, jproplist);
this.LanguageFlags = LanguageFlags.None; //.BubbleNodes;
}
//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);
term.DefaultIntTypes = new TypeCode[] { TypeCode.Int32, TypeCode.UInt32, TypeCode.Int64, TypeCode.UInt64 };
term.DefaultFloatType = TypeCode.Double;
term.AddPrefix("0x", NumberFlags.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 ExpressionEvaluatorGrammar()
{
// 1. Terminals
var number = new NumberLiteral("number");
var identifier = new IdentifierTerminal("identifier");
var comment = new CommentTerminal("comment", "#", "\n", "\r");
base.NonGrammarTerminals.Add(comment);
// 2. Non-terminals
var Variable = new NonTerminal("Variable", typeof(VarRefNode));
var Expr = new NonTerminal("Expr");
var Term = new NonTerminal("Term");
var BinExpr = new NonTerminal("BinExpr", typeof(BinExprNode));
var ParExpr = new NonTerminal("ParExpr");
var UnExpr = new NonTerminal("UnExpr", typeof(UnExprNode));
var UnOp = new NonTerminal("UnOp");
var BinOp = new NonTerminal("BinOp");
var AssignmentStmt = new NonTerminal("AssignmentStmt", typeof(AssigmentNode));
var Statement = new NonTerminal("Statement");
var ProgramLine = new NonTerminal("ProgramLine");
var Program = new NonTerminal("Program", typeof(StatementListNode));
// 3. BNF rules
Variable.Rule = identifier;
Expr.Rule = Term | UnExpr | BinExpr;
Term.Rule = number | ParExpr | Variable;
ParExpr.Rule = "(" + Expr + ")";
UnExpr.Rule = UnOp + Term;
UnOp.Rule = Symbol("+") | "-";
BinExpr.Rule = Expr + BinOp + Expr;
BinOp.Rule = Symbol("+") | "-" | "*" | "/" | "**";
AssignmentStmt.Rule = Variable + "=" + Expr;
Statement.Rule = AssignmentStmt | Expr | Empty;
ProgramLine.Rule = Statement + NewLine;
Program.Rule = MakeStarRule(Program, ProgramLine);
this.Root = Program; // Set grammar root
// 4. Operators precedence
RegisterOperators(1, "+", "-");
RegisterOperators(2, "*", "/");
RegisterOperators(3, Associativity.Right, "**");
RegisterPunctuation( "(", ")");
MarkTransient(Term, Expr, Statement);
//automatically add NewLine before EOF so that our BNF rules work correctly when there's no final line break in source
this.LanguageFlags |= LanguageFlags.NewLineBeforeEOF | LanguageFlags.SupportsInterpreter;
}
public GWBasicGrammar()
{
#region Initialisation
// BASIC is not case sensitive...
this.CaseSensitive = false;
// Define the Terminals
var lineNumber = new NumberLiteral("NUMBER", NumberFlags.IntOnly);
var fileNumber = new NumberLiteral("NUMBER", NumberFlags.IntOnly);
var number = new NumberLiteral("NUMBER", NumberFlags.AllowStartEndDot);
var variable = new IdentifierTerminal("Identifier", "$%!", string.Empty);
var stringLiteral = new StringLiteral("STRING", "\"", StringFlags.None);
//Important: do not add comment term to base.NonGrammarTerminals list - we do use this terminal in grammar rules
var userFunctionName = variable;
var comment = new CommentTerminal("Comment", "REM", "\n");
var short_comment = new CommentTerminal("ShortComment", "'", "\n");
var comma = Symbol(",", "comma");
var colon = Symbol(":", "colon");
var comma_opt = new NonTerminal("comma_opt");
comma_opt.Rule = Empty | ",";
var semi_opt = new NonTerminal("semi_opt");
semi_opt.Rule = Empty | ";";
var pound_opt = new NonTerminal("pound_opt");
pound_opt.Rule = Empty | "#";
// Define the non-terminals
var PROGRAM = new NonTerminal("PROGRAM");
var LINE = new NonTerminal("LINE");
var LINE_CONTENT = new NonTerminal("LINE_CONTENT");
var SHORT_COMMENT_OPT = new NonTerminal("SHORT_COMMENT_OPT");
var STATEMENT_LIST = new NonTerminal("STATEMENT_LIST");
var STATEMENT = new NonTerminal("STATEMENT");
var PRINT_STMT = new NonTerminal("PRINT_STMT");
var PRINT_LIST = new NonTerminal("PRINT_LIST");
var PRINT_ARG = new NonTerminal("PRINT_ARG");
var OPEN_STMT = new NonTerminal("OPEN_STMT");
var OPEN_STMT_MODE = new NonTerminal("OPEN_STMT_MODE");
var OPEN_STMT_ACCESS = new NonTerminal("OPEN_STMT_ACCESS");
var CLOSE_STMT = new NonTerminal("CLOSE_STMT");
var INPUT_STMT = new NonTerminal("INPUT_STMT");
var VARIABLES = new NonTerminal("VARIABLES");
var IF_STMT = new NonTerminal("IF_STMT");
var THEN_CLAUSE = new NonTerminal("THEN_CLAUSE");
var ELSE_CLAUSE_OPT = new NonTerminal("ELSE_CLAUSE_OPT"); //, typeof(AstNode));
var EXPR = new NonTerminal("EXPRESSION");
var EXPR_LIST = new NonTerminal("EXPRESSION_LIST");
var BINARY_OP = new NonTerminal("BINARY_OP");
var BINARY_EXPR = new NonTerminal("BINARY_EXPR");
var UNARY_EXPR = new NonTerminal("UNARY_EXPR");
var SIGN = new NonTerminal("SIGN");
var ASSIGN_STMT = new NonTerminal("ASSIGN_STMT");
var FOR_STMT = new NonTerminal("FOR_STMT");
var STEP_OPT = new NonTerminal("STEP_OPT");
var NEXT_STMT = new NonTerminal("NEXT_STMT");
var LOCATE_STMT = new NonTerminal("LOCATE_STMT");
var WHILE_STMT = new NonTerminal("WHILE_STMT");
var WEND_STMT = new NonTerminal("WEND_STMT");
var SWAP_STMT = new NonTerminal("SWAP_STMT");
var FUN_CALL = new NonTerminal("FUN_CALL");
var VARIABLE_OR_FUNCTION_EXPR = new NonTerminal("VARIABLE_OR_FUNCTION_EXPR");
var ARG_LIST = new NonTerminal("ARG_LIST");
var COMMENT_STMT = new NonTerminal("COMMENT_STMT");
var LINE_INPUT_STMT = new NonTerminal("LINE_INPUT_STMT");
var LINE_INPUT_POUND_STMT = new NonTerminal("LINE_INPUT_POUND_STMT");
var END_STMT = new NonTerminal("END_STMT");
var CLS_STMT = new NonTerminal("CLS_STMT", typeof(AstNode));
var CLEAR_STMT = new NonTerminal("CLEAR_STMT");
var DIM_STMT = new NonTerminal("DIM_STMT");
var DEF_FN_STMT = new NonTerminal("DEF_FN_STMT");
var GOTO_STMT = new NonTerminal("GOTO_STMT");
var GOSUB_STMT = new NonTerminal("GOSUB_STMT");
var RETURN_STMT = new NonTerminal("RETURN_STMT");
var ON_STMT = new NonTerminal("ON_STMT");
var LINE_NUMBERS = new NonTerminal("LINE_NUMBERS");
var RANDOMIZE_STMT = new NonTerminal("RANDOMIZE_STMT");
// set the PROGRAM to be the root node of BASIC programs.
this.Root = PROGRAM;
#endregion
#region Grammar declaration
// A program is a bunch of lines
PROGRAM.Rule = MakePlusRule(PROGRAM, 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 | lineNumber + LINE_CONTENT + SHORT_COMMENT_OPT + NewLine;
// A statement list is 1 or more statements separated by the ':' character
LINE_CONTENT.Rule = IF_STMT | COMMENT_STMT | STATEMENT_LIST;
STATEMENT_LIST.Rule = MakePlusRule(STATEMENT_LIST, colon, STATEMENT);
SHORT_COMMENT_OPT.Rule = short_comment | Empty;
// A statement can be one of a number of types
STATEMENT.Rule = ASSIGN_STMT | PRINT_STMT | INPUT_STMT | OPEN_STMT | CLOSE_STMT
| LINE_INPUT_POUND_STMT | LINE_INPUT_STMT
| LOCATE_STMT | CLS_STMT
| END_STMT | CLEAR_STMT | DIM_STMT | DEF_FN_STMT
| SWAP_STMT | RANDOMIZE_STMT
| GOSUB_STMT | RETURN_STMT | GOTO_STMT | ON_STMT
| FOR_STMT | NEXT_STMT | WHILE_STMT | WEND_STMT;
// The different statements are defined here
PRINT_STMT.Rule = "print" + PRINT_LIST;
PRINT_LIST.Rule = MakeStarRule(PRINT_LIST, null, PRINT_ARG);
PRINT_ARG.Rule = EXPR + semi_opt;
INPUT_STMT.Rule = "input" + semi_opt + stringLiteral + ";" + VARIABLES;
OPEN_STMT.Rule = "open" + EXPR + (Empty | "for" + OPEN_STMT_MODE) +
(Empty | "access" + OPEN_STMT_ACCESS) + "as" + pound_opt + fileNumber;
OPEN_STMT_ACCESS.Rule = "read" + (Empty | "write") | "write";
OPEN_STMT_MODE.Rule = Symbol("o") | "i" | "a" | "output" | "input" | "append";
CLOSE_STMT.Rule = "close" + pound_opt + number;
LINE_INPUT_STMT.Rule = Symbol("line") + "input" + semi_opt + stringLiteral + ";" + VARIABLE_OR_FUNCTION_EXPR;
LINE_INPUT_POUND_STMT.Rule = Symbol("line") + "input" + Symbol("#") + fileNumber + comma + VARIABLE_OR_FUNCTION_EXPR;
DIM_STMT.Rule = "dim" + VARIABLES;
DEF_FN_STMT.Rule = "def" + userFunctionName + (Empty | "(" + ARG_LIST + ")") + "=" + EXPR;
VARIABLES.Rule = VARIABLE_OR_FUNCTION_EXPR | VARIABLE_OR_FUNCTION_EXPR + "," + VARIABLES;
IF_STMT.Rule = "if" + EXPR + THEN_CLAUSE + ELSE_CLAUSE_OPT;
THEN_CLAUSE.Rule = "then" + STATEMENT_LIST | "goto" + lineNumber;
//Inject PreferShift hint here to explicitly set shift as preferred action. Suppresses warning message about conflict.
ELSE_CLAUSE_OPT.Rule = Empty | PreferShiftHere() + "else" + STATEMENT_LIST;
GOTO_STMT.Rule = "goto" + lineNumber;
GOSUB_STMT.Rule = "gosub" + lineNumber;
RETURN_STMT.Rule = "return";
ON_STMT.Rule = "on" + EXPR + (Symbol("goto") | "gosub") + LINE_NUMBERS;
LINE_NUMBERS.Rule = lineNumber + (Empty | "," + LINE_NUMBERS);
ASSIGN_STMT.Rule = VARIABLE_OR_FUNCTION_EXPR + "=" + EXPR;
LOCATE_STMT.Rule = "locate" + EXPR + comma + EXPR;
SWAP_STMT.Rule = "swap" + EXPR + comma + EXPR;
END_STMT.Rule = "end";
CLS_STMT.Rule = "cls";
CLEAR_STMT.Rule = Symbol("clear") + comma + (Empty | number) + (Empty | comma + number) | "clear" + number | "clear";
COMMENT_STMT.Rule = comment | short_comment;
RANDOMIZE_STMT.Rule = "randomize" + EXPR;
// An expression is a number, or a variable, a string, or the result of a binary comparison.
EXPR.Rule = number | variable | FUN_CALL | stringLiteral | BINARY_EXPR
| "(" + EXPR + ")" | UNARY_EXPR;
BINARY_EXPR.Rule = EXPR + BINARY_OP + EXPR;
UNARY_EXPR.Rule = SIGN + EXPR;
SIGN.Rule = Symbol("-") | "+";
//Inject PreferShift hint here to explicitly set shift as preferred action. Suppresses warning message about conflict.
FUN_CALL.Rule = variable + PreferShiftHere() + "(" + ARG_LIST + ")";
VARIABLE_OR_FUNCTION_EXPR.Rule = variable | FUN_CALL;
BINARY_OP.Rule = Symbol("+") | "^" | "-" | "*" | "/" | "=" | "<=" | ">=" | "<" | ">" | "<>" | "and" | "or";
//let's do operator precedence right here
RegisterOperators(60, "^");
RegisterOperators(50, "*", "/");
RegisterOperators(40, "+", "-");
RegisterOperators(30, "=", "<=", ">=", "<", ">", "<>");
RegisterOperators(20, "and", "or");
EXPR_LIST.Rule = MakeStarRule(EXPR_LIST, null, EXPR);
FOR_STMT.Rule = "for" + ASSIGN_STMT + "to" + EXPR + STEP_OPT;
STEP_OPT.Rule = Empty | "step" + EXPR;
NEXT_STMT.Rule = "next" + VARIABLES | "next";
WHILE_STMT.Rule = "while" + EXPR;
WEND_STMT.Rule = "wend";
//TODO: check number of arguments for particular function in node constructor
ARG_LIST.Rule = MakePlusRule(ARG_LIST, comma, EXPR);
#endregion
#region Punctuation
RegisterPunctuation("(", ")", ",", ";");
#endregion
}
//http://docs.python.org/ref/numbers.html
public static NumberLiteral CreatePythonNumber(String name)
{
NumberLiteral term = new NumberLiteral(name, NumberFlags.AllowStartEndDot);
//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.AddPrefix("0x", NumberFlags.Hex);
term.AddPrefix("0", NumberFlags.Octal);
term.AddSuffixCodes("L", TypeCode.Int64, NumberLiteral.TypeCodeBigInt);
term.AddSuffixCodes("J", NumberLiteral.TypeCodeImaginary);
return term;
}
//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);
term.DefaultIntTypes = new TypeCode[] { TypeCode.Int32, TypeCode.Int64 };
//term.DefaultFloatType = TypeCode.Double; it is default
term.AddPrefix("&H", NumberFlags.Hex);
term.AddPrefix("&O", NumberFlags.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);
term.DefaultIntTypes = new TypeCode[] { TypeCode.Int32, TypeCode.Int64, NumberLiteral.TypeCodeBigInt };
term.DefaultFloatType = TypeCode.Double; // it is default
term.ExponentSymbols = "sfdl";
term.AddPrefix("#b", NumberFlags.Binary);
term.AddPrefix("#o", NumberFlags.Octal);
term.AddPrefix("#x", NumberFlags.Hex);
term.AddPrefix("#d", NumberFlags.None);
term.AddPrefix("#i", NumberFlags.None); // inexact prefix, has no effect
term.AddPrefix("#e", NumberFlags.None); // exact prefix, has no effect
term.AddSuffixCodes("J", NumberLiteral.TypeCodeImaginary);
return term;
}
public SqlGrammar()
{
//Terminals
var comment = new CommentTerminal("comment", "/*", "*/");
var lineComment = new CommentTerminal("line_comment", "--", "\n", "\r\n");
NonGrammarTerminals.Add(comment);
NonGrammarTerminals.Add(lineComment);
var number = new NumberLiteral("number");
var string_literal = new StringLiteral("string", "'", StringFlags.AllowsDoubledQuote);
var name = new IdentifierTerminal("name");
var name_ext = TerminalFactory.CreateSqlExtIdentifier("name_ext");
var comma = Symbol(",");
var dot = Symbol(".");
var CREATE = Symbol("CREATE");
var NULL = Symbol("NULL");
var NOT = Symbol("NOT");
var UNIQUE = Symbol("UNIQUE");
var WITH = Symbol("WITH");
var TABLE = Symbol("TABLE");
var ALTER = Symbol("ALTER");
var ADD = Symbol("ADD");
var COLUMN = Symbol("COLUMN");
var DROP = Symbol("DROP");
var CONSTRAINT = Symbol("CONSTRAINT");
var INDEX = Symbol("INDEX");
var ON = Symbol("ON");
var KEY = Symbol("KEY");
var PRIMARY = Symbol("PRIMARY");
var INSERT = Symbol("INSERT");
var INTO = Symbol("INTO");
var UPDATE = Symbol("UPDATE");
var SET = Symbol("SET");
var VALUES = Symbol("VALUES");
var DELETE = Symbol("DELETE");
var SELECT = Symbol("SELECT");
var FROM = Symbol("FROM");
var AS = Symbol("AS");
var COUNT = Symbol("COUNT");
var JOIN = Symbol("JOIN");
var BY = Symbol("BY");
//Non-terminals
var Id = new NonTerminal("Id");
var Id_simple = new NonTerminal("id_simple");
var stmt = new NonTerminal("stmt");
var createTableStmt = new NonTerminal("createTableStmt");
var createIndexStmt = new NonTerminal("createIndexStmt");
var alterStmt = new NonTerminal("alterStmt");
var dropTableStmt = new NonTerminal("dropTableStmt");
var dropIndexStmt = new NonTerminal("dropIndexStmt");
var selectStmt = new NonTerminal("selectStmt");
var insertStmt = new NonTerminal("insertStmt");
var updateStmt = new NonTerminal("updateStmt");
var deleteStmt = new NonTerminal("deleteStmt");
var fieldDef = new NonTerminal("fieldDef");
var fieldDefList = new NonTerminal("fieldDefList");
var nullSpecOpt = new NonTerminal("nullSpecOpt");
var typeName = new NonTerminal("typeName");
var typeSpec = new NonTerminal("typeSpec");
var typeParamsOpt = new NonTerminal("typeParams");
var constraintDef = new NonTerminal("constraintDef");
var constraintListOpt = new NonTerminal("constraintListOpt");
var constraintTypeOpt = new NonTerminal("constraintTypeOpt");
var idlist = new NonTerminal("idlist");
var idlistPar = new NonTerminal("idlistPar");
var uniqueOpt = new NonTerminal("uniqueOpt");
var orderList = new NonTerminal("orderList");
var orderMember = new NonTerminal("orderMember");
var orderDirOpt = new NonTerminal("orderDirOpt");
var withClauseOpt = new NonTerminal("withClauseOpt");
var alterCmd = new NonTerminal("alterCmd");
var insertData = new NonTerminal("insertData");
var intoOpt = new NonTerminal("intoOpt");
var assignList = new NonTerminal("assignList");
var whereClauseOpt = new NonTerminal("whereClauseOpt");
var assignment = new NonTerminal("assignment");
var expression = new NonTerminal("expression");
var exprList = new NonTerminal("exprList");
var selRestrOpt = new NonTerminal("selRestrOpt");
var selList = new NonTerminal("selList");
var intoClauseOpt = new NonTerminal("intoClauseOpt");
var fromClauseOpt = new NonTerminal("fromClauseOpt");
var groupClauseOpt = new NonTerminal("groupClauseOpt");
var havingClauseOpt = new NonTerminal("havingClauseOpt");
var orderClauseOpt = new NonTerminal("orderClauseOpt");
var columnItemList = new NonTerminal("columnItemList");
var columnItem = new NonTerminal("columnItem");
var columnSource = new NonTerminal("columnSource");
var asOpt = new NonTerminal("asOpt");
var aliasOpt = new NonTerminal("aliasOpt");
var aggregate = new NonTerminal("aggregate");
var aggregateArg = new NonTerminal("aggregateArg");
var aggregateName = new NonTerminal("aggregateName");
var tuple = new NonTerminal("tuple");
var joinChainOpt = new NonTerminal("joinChainOpt");
var joinKindOpt = new NonTerminal("joinKindOpt");
var term = new NonTerminal("term");
var unExpr = new NonTerminal("unExpr");
var unOp = new NonTerminal("unOp");
var binExpr = new NonTerminal("binExpr");
var binOp = new NonTerminal("binOp");
var betweenExpr = new NonTerminal("betweenExpr");
var inExpr = new NonTerminal("inExpr");
var parSelectStmt = new NonTerminal("parSelectStmt");
var notOpt = new NonTerminal("notOpt");
var funCall = new NonTerminal("funCall");
var stmtLine = new NonTerminal("stmtLine");
var semiOpt = new NonTerminal("semiOpt");
var stmtList = new NonTerminal("stmtList");
var funArgs = new NonTerminal("funArgs");
var inStmt = new NonTerminal("inStmt");
//BNF Rules
this.Root = stmtList;
stmtLine.Rule = stmt + semiOpt;
semiOpt.Rule = Empty | ";";
stmtList.Rule = MakePlusRule(stmtList, stmtLine);
//ID
Id_simple.Rule = name | name_ext;
Id.Rule = MakePlusRule(Id, dot, Id_simple);
stmt.Rule = createTableStmt | createIndexStmt | alterStmt
| dropTableStmt | dropIndexStmt
| selectStmt | insertStmt | updateStmt | deleteStmt
| "GO" ;
//Create table
createTableStmt.Rule = CREATE + TABLE + Id + "(" + fieldDefList + ")" + constraintListOpt;
fieldDefList.Rule = MakePlusRule(fieldDefList, comma, fieldDef);
fieldDef.Rule = Id + typeName + typeParamsOpt + nullSpecOpt;
nullSpecOpt.Rule = NULL | NOT + NULL | Empty;
typeName.Rule = Symbol("BIT") | "DATE" | "TIME" | "TIMESTAMP" | "DECIMAL" | "REAL" | "FLOAT" | "SMALLINT" | "INTEGER"
| "INTERVAL" | "CHARACTER"
// MS SQL types:
| "DATETIME" | "INT" | "DOUBLE" | "CHAR" | "NCHAR" | "VARCHAR" | "NVARCHAR"
| "IMAGE" | "TEXT" | "NTEXT";
typeParamsOpt.Rule = "(" + number + ")" | "(" + number + comma + number + ")" | Empty;
constraintDef.Rule = CONSTRAINT + Id + constraintTypeOpt;
constraintListOpt.Rule = MakeStarRule(constraintListOpt, constraintDef );
constraintTypeOpt.Rule = PRIMARY + KEY + idlistPar | UNIQUE + idlistPar | NOT + NULL + idlistPar
| "Foreign" + KEY + idlistPar + "References" + Id + idlistPar;
idlistPar.Rule = "(" + idlist + ")";
idlist.Rule = MakePlusRule(idlist, comma, Id);
//Create Index
createIndexStmt.Rule = CREATE + uniqueOpt + INDEX + Id + ON + Id + orderList + withClauseOpt;
uniqueOpt.Rule = Empty | UNIQUE;
orderList.Rule = MakePlusRule(orderList, comma, orderMember);
orderMember.Rule = Id + orderDirOpt;
orderDirOpt.Rule = Empty | "ASC" | "DESC";
withClauseOpt.Rule = Empty | WITH + PRIMARY | WITH + "Disallow" + NULL | WITH + "Ignore" + NULL;
//Alter
alterStmt.Rule = ALTER + TABLE + Id + alterCmd;
alterCmd.Rule = ADD + COLUMN + fieldDefList + constraintListOpt
| ADD + constraintDef
| DROP + COLUMN + Id
| DROP + CONSTRAINT + Id;
//Drop stmts
dropTableStmt.Rule = DROP + TABLE + Id;
dropIndexStmt.Rule = DROP + INDEX + Id + ON + Id;
//Insert stmt
insertStmt.Rule = INSERT + intoOpt + Id + idlistPar + insertData;
insertData.Rule = selectStmt | VALUES + "(" + exprList + ")";
intoOpt.Rule = Empty | INTO; //Into is optional in MSSQL
//Update stmt
updateStmt.Rule = UPDATE + Id + SET + assignList + whereClauseOpt;
assignList.Rule = MakePlusRule(assignList, comma, assignment);
assignment.Rule = Id + "=" + expression;
//Delete stmt
deleteStmt.Rule = DELETE + FROM + Id + whereClauseOpt;
//Select stmt
selectStmt.Rule = SELECT + selRestrOpt + selList + intoClauseOpt + fromClauseOpt + whereClauseOpt +
groupClauseOpt + havingClauseOpt + orderClauseOpt;
selRestrOpt.Rule = Empty | "ALL" | "DISTINCT";
selList.Rule = columnItemList | "*";
columnItemList.Rule = MakePlusRule(columnItemList, comma, columnItem);
columnItem.Rule = columnSource + aliasOpt;
aliasOpt.Rule = Empty | asOpt + Id;
asOpt.Rule = Empty | AS;
columnSource.Rule = aggregate | Id;
aggregate.Rule = aggregateName + "(" + aggregateArg + ")";
aggregateArg.Rule = expression | "*";
aggregateName.Rule = COUNT | "Avg" | "Min" | "Max" | "StDev" | "StDevP" | "Sum" | "Var" | "VarP";
intoClauseOpt.Rule = Empty | INTO + Id;
fromClauseOpt.Rule = Empty | FROM + idlist + joinChainOpt;
joinChainOpt.Rule = Empty | joinKindOpt + JOIN + idlist + ON + Id + "=" + Id;
joinKindOpt.Rule = Empty | "INNER" | "LEFT" | "RIGHT";
whereClauseOpt.Rule = Empty | "WHERE" + expression;
groupClauseOpt.Rule = Empty | "GROUP" + BY + idlist;
havingClauseOpt.Rule = Empty | "HAVING" + expression;
orderClauseOpt.Rule = Empty | "ORDER" + BY + orderList;
//Expression
exprList.Rule = MakePlusRule(exprList, comma, expression);
expression.Rule = term | unExpr | binExpr;// | betweenExpr; //-- BETWEEN doesn't work - yet; brings a few parsing conflicts
term.Rule = Id | string_literal | number | funCall | tuple | parSelectStmt;// | inStmt;
tuple.Rule = "(" + exprList + ")";
parSelectStmt.Rule = "(" + selectStmt + ")";
unExpr.Rule = unOp + term;
unOp.Rule = NOT | "+" | "-" | "~";
binExpr.Rule = expression + binOp + expression;
binOp.Rule = Symbol("+") | "-" | "*" | "/" | "%" //arithmetic
| "&" | "|" | "^" //bit
| "=" | ">" | "<" | ">=" | "<=" | "<>" | "!=" | "!<" | "!>"
| "AND" | "OR" | "LIKE" | NOT + "LIKE" | "IN" | NOT + "IN" ;
betweenExpr.Rule = expression + notOpt + "BETWEEN" + expression + "AND" + expression;
notOpt.Rule = Empty | NOT;
//funCall covers some psedo-operators and special forms like ANY(...), SOME(...), ALL(...), EXISTS(...), IN(...)
funCall.Rule = name + "(" + funArgs + ")";
funArgs.Rule = selectStmt | exprList;
inStmt.Rule = expression + "IN" + "(" + exprList + ")";
//Operators
this.CaseSensitive = false;
RegisterOperators(10, "*", "/", "%");
RegisterOperators(9, "+", "-");
RegisterOperators(8, "=" , ">" , "<" , ">=" , "<=" , "<>" , "!=" , "!<" , "!>");
RegisterOperators(7, "^", "&", "|");
RegisterOperators(6, "NOT");
RegisterOperators(5, "AND");
RegisterOperators(4, "OR", "LIKE", "IN");
RegisterPunctuation(",", "(", ")");
RegisterPunctuation(semiOpt);
base.MarkTransient(stmt, Id_simple, asOpt, aliasOpt, stmtLine);
AddKeywords("SELECT", "CREATE", "ALTER", "UPDATE", "INSERT", "DELETE", "FROM", "WHERE", "GROUP", "ORDER", "BY",
"INNER", "LEFT", "RIGHT", "JOIN", "ON");
}
// It is loosely based on R6RS specs.
// See Grammar Errors tab in GrammarExplorer for remaining conflicts.
public SchemeGrammar()
{
#region Terminals
ConstantTerminal Constant = new ConstantTerminal("Constant");
Constant.Add("#T", 1);
Constant.Add("#t", 1);
Constant.Add("#F", null);
Constant.Add("#f", null);
Constant.Add("'()", null);
Constant.Add(@"#\nul", '\u0000');
Constant.Add(@"#\alarm", '\u0007');
Constant.Add(@"#\backspace", '\b');
Constant.Add(@"#\tab", '\t');
Constant.Add(@"#\linefeed", '\n');
Constant.Add(@"#\vtab", '\v');
Constant.Add(@"#\page", '\f');
Constant.Add(@"#\return", '\r');
Constant.Add(@"#\esc", '\u001B');
Constant.Add(@"#\space", ' ');
Constant.Add(@"#\delete", '\u007F');
// TODO: build SchemeCharLiteral
// the following is nonsense, just to put something there
var charLiteral = new StringLiteral("Char", "'", StringFlags.None);
var stringLiteral = new StringLiteral("String", "\"", StringFlags.AllowsAllEscapes);
//Identifiers. Note: added "-", just to allow IDs starting with "->"
var SimpleIdentifier = new IdentifierTerminal("SimpleIdentifier", "_+-*/.@?!<>=", "_+-*/.@?!<>=$%&:^~");
// name extraChars extraFirstChars
var Number = TerminalFactory.CreateSchemeNumber("Number");
var Byte = new NumberLiteral("Byte", NumberFlags.IntOnly);
//Comments
Terminal Comment = new CommentTerminal("Comment", "#|", "|#");
Terminal LineComment = new CommentTerminal("LineComment", ";", "\n");
NonGrammarTerminals.Add(Comment); //add comments explicitly to this list as it is not reachable from Root
NonGrammarTerminals.Add(LineComment);
#endregion
#region NonTerminals
var Module = new NonTerminal("Module");
var Library = new NonTerminal("Library");
var LibraryList = new NonTerminal("Library+");
var Script = new NonTerminal("Script");
var Abbreviation = new NonTerminal("Abbreviation");
var Vector = new NonTerminal("Vector");
var ByteList = new NonTerminal("ByteList");
var ByteVector = new NonTerminal("ByteVector");
var Datum = new NonTerminal("Datum"); //Datum in R6RS terms
var DatumOpt = new NonTerminal("DatumOpt"); //Datum in R6RS terms
var DatumList = new NonTerminal("Datum+", typeof(StatementListNode));
var DatumListOpt = new NonTerminal("Datum*", typeof(StatementListNode));
var Statement = new NonTerminal("Statement");
var Atom = new NonTerminal("Atom");
var CompoundDatum = new NonTerminal("CompoundDatum");
var AbbrevPrefix = new NonTerminal("AbbrevPrefix");
var LibraryName = new NonTerminal("LibraryName");
var LibraryBody = new NonTerminal("LibraryBody");
var ImportSection = new NonTerminal("ImportSection");
var ExportSection = new NonTerminal("ExportSection");
var ImportSpec = new NonTerminal("ImportSpec");
var ImportSpecList = new NonTerminal("ImportSpecList");
var ExportSpec = new NonTerminal("ExportSpec");
var ExportSpecList = new NonTerminal("ExportSpecList");
var LP = new NonTerminal("LP"); //"(" or "["
var RP = new NonTerminal("RP"); // ")" or "]"
var Identifier = new NonTerminal("Identifier", typeof(VarRefNode));
var IdentifierList = new NonTerminal("IdentifierList");
var IdentifierListOpt = new NonTerminal("IdentifierListOpt");
var PeculiarIdentifier = new NonTerminal("PeculiarIdentifier");
var LibraryVersion = new NonTerminal("LibraryVersion");
var VersionListOpt = new NonTerminal("VersionListOpt");
var FunctionCall = new NonTerminal("FunctionCall", CreateFunctionCallNode);
var FunctionRef = new NonTerminal("FunctionRef");
var SpecialForm = new NonTerminal("SpecialForm");
var DefineVarForm = new NonTerminal("DefineVarForm", CreateDefineVarNode);
var DefineFunForm = new NonTerminal("DefineFunForm", CreateDefineFunNode);
var LambdaForm = new NonTerminal("LambdaForm", CreateLambdaNode);
var IfForm = new NonTerminal("IfForm", CreateIfThenElseNode);
var CondForm = new NonTerminal("CondForm", CreateCondFormNode);
var CondClause = new NonTerminal("CondClause", CreateCondClauseNode);
var CondClauseList = new NonTerminal("CondClauseList");
var CondElseOpt = new NonTerminal("CondElseOpt");
var BeginForm = new NonTerminal("BeginForm", CreateBeginNode);
var LetForm = new NonTerminal("LetForm"); //not implemented
var LetRecForm = new NonTerminal("LetRecForm"); //not implemented
var LetPair = new NonTerminal("LetPair");
var LetPairList = new NonTerminal("LetPairList");
#endregion
#region Rules
//
// Using optional elements in Scheme grammar brings some nasty conflicts - by default the parser selects "shift" over reduce
// which leads to failure to parse simple programs without libraries and export/import sections.
// This trouble comes from that the fact that Scheme has soooooo many parenthesis. Therefore, using a single next symbol
// as a lookahead (as it happens in LALR parsing) doesn't help much - the next symbol is almost always a parenthesis,
// not some meaningful symbol. That's why in the following expressions I had to use explicit listing of variants,
// instead of simply marking some elements as optional (see Module, Script, LibraryBody elements) - this clears the conflicts
// but would make node construction more difficult.
base.Root = Module;
LP.Rule = Symbol("(") | "["; //R6RS allows mix & match () and []
RP.Rule = Symbol(")") | "]";
// Module.Rule = LibraryListOpt + Script; -- this brings conflicts
Module.Rule = LibraryList + Script | Script;
LibraryList.Rule = MakePlusRule(LibraryList, Library);
Script.Rule = ImportSection + DatumList | DatumList;
//Library
// the following doesn't work - brings conflicts that incorrectly resolved by default shifting
//Library.Rule = LP + "library" + LibraryName + ExportSectionOpt + ImportSectionOpt + DatumListOpt + RP;
Library.Rule = LP + "library" + LibraryName + LibraryBody + RP;
//Note - we should be using DatumListOpt, but that brings 2 conflicts, so for now it is just DatumList
//Note that the following style of BNF expressions is strongly discouraged - all productions should be of the same length,
// so that the process of mapping child nodes to parent's properties is straightforward.
LibraryBody.Rule = ExportSection + ImportSection + DatumList
| ExportSection + DatumList
| ImportSection + DatumList
| DatumList;
LibraryName.Rule = LP + IdentifierList + LibraryVersion.Q() + RP;
LibraryVersion.Rule = LP + VersionListOpt + RP; //zero or more subversion numbers
VersionListOpt.Rule = MakeStarRule(VersionListOpt, Number);
ExportSection.Rule = LP + "export" + ExportSpecList + RP;
ImportSection.Rule = LP + "import" + ImportSpecList + RP;
ExportSpecList.Rule = MakePlusRule(ExportSpecList, ExportSpec);
ImportSpecList.Rule = MakePlusRule(ImportSpecList, ImportSpec);
ExportSpec.Rule = Identifier | LP + "rename" + LP + Identifier + Identifier + RP + RP;
ImportSpec.Rule = LP + Identifier + RP; // - much more complex in R6RS
//Datum
Datum.Rule = Atom | CompoundDatum;
DatumOpt.Rule = Empty | Datum;
DatumList.Rule = MakePlusRule(DatumList, Datum);
DatumListOpt.Rule = MakeStarRule(DatumListOpt, Datum);
Atom.Rule = Number | Identifier | stringLiteral | Constant | charLiteral | ".";
CompoundDatum.Rule = Statement | Abbreviation | Vector | ByteVector;
Identifier.Rule = SimpleIdentifier | PeculiarIdentifier;
IdentifierList.Rule = MakePlusRule(IdentifierList, Identifier);
IdentifierListOpt.Rule = MakeStarRule(IdentifierListOpt, Identifier);
//TODO: create PeculiarIdentifier custom terminal instead of var
// or just custom SchemeIdentifier terminal
PeculiarIdentifier.Rule = Symbol("+") | "-" | "..."; // |"->" + subsequent; (should be!)
Abbreviation.Rule = AbbrevPrefix + Datum;
AbbrevPrefix.Rule = Symbol("'") | "`" | ",@" | "," | "#'" | "#`" | "#,@" | "#,";
Vector.Rule = "#(" + DatumListOpt + ")";
ByteVector.Rule = "#vu8(" + ByteList + ")";
ByteList.Rule = MakeStarRule(ByteList, Byte);
Statement.Rule = FunctionCall | SpecialForm;
FunctionCall.Rule = LP + FunctionRef + DatumListOpt + RP;
FunctionRef.Rule = Identifier | Statement;
SpecialForm.Rule = DefineVarForm | DefineFunForm | LambdaForm | IfForm | CondForm | BeginForm | LetForm | LetRecForm;
DefineVarForm.Rule = LP + "define" + Identifier + Datum + RP;
DefineFunForm.Rule = LP + "define" + LP + Identifier + IdentifierListOpt + RP + DatumList + RP;
LambdaForm.Rule = LP + "lambda" + LP + IdentifierListOpt + RP + DatumList + RP;
IfForm.Rule = LP + "if" + Datum + Datum + DatumOpt + RP;
CondForm.Rule = LP + "cond" + CondClauseList + CondElseOpt + RP;
CondClauseList.Rule = MakePlusRule(CondClauseList, CondClause);
CondClause.Rule = LP + Datum + DatumList + RP;
CondElseOpt.Rule = Empty | LP + "else" + DatumList + RP;
LetForm.Rule = LP + "let" + LP + LetPairList + RP + DatumList + RP;
LetRecForm.Rule = LP + "letrec" + LP + LetPairList + RP + DatumList + RP;
BeginForm.Rule = LP + "begin" + DatumList + RP;
LetPairList.Rule = MakePlusRule(LetPairList, LetPair);
LetPair.Rule = LP + Identifier + Datum + RP;
#endregion
//Register brace pairs
RegisterBracePair("(", ")");
RegisterBracePair("[", "]");
RegisterPunctuation(LP, RP);
//Filters and other stuff
BraceMatchFilter filter = new BraceMatchFilter();
TokenFilters.Add(filter);
//Scheme is tail-recursive language
base.LanguageFlags = LanguageFlags.BubbleNodes | LanguageFlags.TailRecursive |
LanguageFlags.SupportsInterpreter | LanguageFlags.SupportsConsole;
//keywords - just for colorizer
base.AddKeywords("define", "lambda", "cond", "if", "begin", "let");
}
