basic: parser now process (:) operator

This commit is contained in:
minjaesong
2020-12-12 16:24:14 +09:00
parent 19f323b90e
commit 81e10c54de
2 changed files with 733 additions and 334 deletions

View File

@@ -44,6 +44,13 @@ function isNumable(s) {
return s !== undefined && (typeof s.trim == "function" && s.trim() !== "" || s.trim == undefined) && !isNaN(s); return s !== undefined && (typeof s.trim == "function" && s.trim() !== "" || s.trim == undefined) && !isNaN(s);
} }
class ParserError extends Error {
constructor(...args) {
super(...args);
Error.captureStackTrace(this, ParserError);
}
}
let lang = {}; let lang = {};
lang.badNumberFormat = Error("Illegal number format"); lang.badNumberFormat = Error("Illegal number format");
lang.badOperatorFormat = Error("Illegal operator format"); lang.badOperatorFormat = Error("Illegal operator format");
@@ -202,28 +209,29 @@ let BasicVar = function(literal, type) {
} }
// Abstract Syntax Tree // Abstract Syntax Tree
// creates empty tree node // creates empty tree node
let astToString = function(ast) { let astToString = function(ast, depth) {
let l__ = String.fromCharCode(0x2502,32);
let recDepth = depth || 0;
if (ast === undefined || ast.astType === undefined) return ""; if (ast === undefined || ast.astType === undefined) return "";
var sb = ""; var sb = "";
var marker = ("lit" == ast.astType) ? "i" : var marker = ("lit" == ast.astType) ? "i" :
("op" == ast.astType) ? String.fromCharCode(177) : ("op" == ast.astType) ? String.fromCharCode(0xB1) :
("string" == ast.astType) ? String.fromCharCode(182) : ("string" == ast.astType) ? String.fromCharCode(0xB6) :
("num" == ast.astType) ? String.fromCharCode(162) : ("num" == ast.astType) ? String.fromCharCode(0xA2) :
("array" == ast.astType) ? "[" : String.fromCharCode(163); ("array" == ast.astType) ? "[" : String.fromCharCode(0x192);
sb += "| ".repeat(ast.astDepth) + marker+" Line "+ast.astLnum+" ("+ast.astType+")\n"; sb += l__.repeat(recDepth) + marker+" Line "+ast.astLnum+" ("+ast.astType+")\n";
sb += "| ".repeat(ast.astDepth+1) + "leaves: "+(ast.astLeaves.length)+"\n"; sb += l__.repeat(recDepth+1) + "leaves: "+(ast.astLeaves.length)+"\n";
sb += "| ".repeat(ast.astDepth+1) + "value: "+ast.astValue+" (type: "+typeof ast.astValue+")\n"; sb += l__.repeat(recDepth+1) + "value: "+ast.astValue+" (type: "+typeof ast.astValue+")\n";
for (var k = 0; k < ast.astLeaves.length; k++) { for (var k = 0; k < ast.astLeaves.length; k++) {
if (k > 0) if (k > 0)
sb += "| ".repeat(ast.astDepth+1) + " " + ast.astSeps[k - 1] + "\n"; sb += l__.repeat(recDepth+1) + " " + ast.astSeps[k - 1] + "\n";
sb += astToString(ast.astLeaves[k]); sb += astToString(ast.astLeaves[k], recDepth + 1);
} }
sb += "| ".repeat(ast.astDepth) + "`-----------------\n"; sb += l__.repeat(recDepth)+String.fromCharCode(0x2570)+String.fromCharCode(0x2500).repeat(13)+'\n';
return sb; return sb;
} }
let BasicAST = function() { let BasicAST = function() {
this.astLnum = 0; this.astLnum = 0;
this.astDepth = 0;
this.astLeaves = []; this.astLeaves = [];
this.astSeps = []; this.astSeps = [];
this.astValue = undefined; this.astValue = undefined;
@@ -1148,7 +1156,7 @@ DEFUN'd functions must be treated as if their args is "vararg"
}; };
Object.freeze(bStatus.builtin); Object.freeze(bStatus.builtin);
let bF = {}; let bF = {};
bF._1os = {"!":1,"~":1,"#":1,"<":1,"=":1,">":1,"*":1,"+":1,"-":1,"/":1,"^":1}; bF._1os = {"!":1,"~":1,"#":1,"<":1,"=":1,">":1,"*":1,"+":1,"-":1,"/":1,"^":1,":":1};
bF._2os = {"<":1,"=":1,">":1}; bF._2os = {"<":1,"=":1,">":1};
bF._uos = {"+":1,"-":1}; bF._uos = {"+":1,"-":1};
bF._isNum = function(code) { bF._isNum = function(code) {
@@ -1582,6 +1590,8 @@ bF._parserElaboration = function(lnum, tokens, states) {
states[k] = "op"; states[k] = "op";
else if (tokens[k].toUpperCase() == "TRUE" || tokens[k].toUpperCase() == "FALSE") else if (tokens[k].toUpperCase() == "TRUE" || tokens[k].toUpperCase() == "FALSE")
states[k] = "bool"; states[k] = "bool";
else if (tokens[k] == ":" && states[k] == "op")
states[k] = "seq";
// decimalise hex/bin numbers (because Nashorn does not support binary literal) // decimalise hex/bin numbers (because Nashorn does not support binary literal)
if (states[k] == "num") { if (states[k] == "num") {
@@ -1602,38 +1612,40 @@ bF._recurseApplyAST = function(tree, action) {
} }
} }
/** EBNF notation: /** EBNF notation:
(* quick reference to EBNF *)
(* { word } = word is repeated 0 or more times *)
(* [ word ] = word is optional (repeated 0 or 1 times) *)
line = linenumber , stmt , {":" , stmt} ; line = linenumber , stmt , {":" , stmt} ;
linenumber = digits ; linenumber = digits ;
stmt = stmt =
"IF" , if_equation , "THEN" , stmt , ["ELSE" , stmt] "IF" , expr_sans_asgn , "THEN" , stmt , ["ELSE" , stmt]
| "DEFUN" , [ident] , "(" , [ident , {" , " , ident}] , ")" , "=" , stmt | "DEFUN" , [ident] , "(" , [ident , {" , " , ident}] , ")" , "=" , expr
| "ON" , if_equation , ident , if_equation , {"," , if_equation} | "ON" , expr_sans_asgn , ("GOTO" | "GOSUB") , expr_sans_asgn , {"," , expr_sans_asgn}
| "(" , stmt , ")" | "(" , stmt , ")"
| expr ;
expr = (* this basically blocks some funny attemps such as using DEFUN as anon function because everything is global in BASIC *)
lit
| "(" , expr , ")"
| "IF" , expr_sans_asgn , "THEN" , expr , ["ELSE" , expr]
| expr , op , expr
| op_uni , expr
| function_call ; | function_call ;
expr_sans_asgn = ? identical to expr except errors out whenever "=" is found ? ;
function_call = function_call =
equation ident , "(" , [expr , {argsep , expr} , [argsep]] , ")"
| ident , "(" , [function_call , {argsep , function_call} , [argsep]] , ")" | ident , expr , {argsep , expr} , [argsep] ;
| ident , function_call , {argsep , function_call} , [argsep] ;
equation =
lit
| "(" , equation , ")"
| equation , op , equation
| op_uni , equation ;
if_equation = if_equation , op - ("=") , if_equation
| op_uni , if_equation
| lit
| "(" , if_equation , ")" ;
(* don't bother looking at these, because you already know the stuff *) (* don't bother looking at these, because you already know the stuff *)
function = lit ; argsep = "," | ";" ;
argsep = ","|";" ; ident = alph , [digits] ; (* variable and function names *)
ident = alph , [digits] ; lit = alph , [digits] | num | string ; (* ident + numbers and string literals *)
lit = alph , [digits] | num | string ; (* example: "MyVar_2" *)
op = "^" | "*" | "/" | "MOD" | "+" | "-" | "<<" | ">>" | "<" | ">" | "<=" op = "^" | "*" | "/" | "MOD" | "+" | "-" | "<<" | ">>" | "<" | ">" | "<="
| "=<" | ">=" | "=>" | "==" | "<>" | "><" | "BAND" | "BXOR" | "BOR" | "=<" | ">=" | "=>" | "==" | "<>" | "><" | "BAND" | "BXOR" | "BOR"
| "AND" | "OR" | "TO" | "STEP" | "!" | "~" | "#" | "=" ; | "AND" | "OR" | "TO" | "STEP" | "!" | "~" | "#" | "=" ;
@@ -1678,8 +1690,8 @@ DEFUN (type: function, value: DEFUN)
2. stmt 2. stmt
ON (type: function, value: ON) ON (type: function, value: ON)
1. varname 1. testvalue
2. functionname 2. functionname (type: lit)
3. arg0 3. arg0
[4. arg1] [4. arg1]
[5. argN...] [5. argN...]
@@ -1689,148 +1701,406 @@ FUNCTION_CALL (type: function, value: PRINT or something)
2. arg1 2. arg1
[3. argN...] [3. argN...]
*/ */
// @return BasicAST
bF._parseEquation = functoin(lnum, tokens, states, recDepth) {
}
// @returns BasicAST // @returns BasicAST
bF._parseTokens = function(lnum, tokens, states, recDepth) { bF._EquationIllegalTokens = ["IF","THEN","ELSE","DEFUN","ON"];
bF.isSemanticLiteral = function(token, state) {
function isSemanticLiteral(token, state) {
return "]" == token || ")" == token || return "]" == token || ")" == token ||
"qot" == state || "num" == state || "bool" == state || "lit" == state; "qot" == state || "num" == state || "bool" == state || "lit" == state;
}
bF.parserDoDebugPrint = false;
bF.parserPrintdbg = any => if (parserDoDebugPrint) serial.println(any);
bF.parserPrintdbg2 = function(icon, lnum, tokens, states, recDepth) {
if (parserDoDebugPrint) {
let treeHead = String.fromCharCode(0x2502,32).repeat(recDepth);
bF.parserPrintdbg(`${icon}${lnum} ${treeHead}${tokens.join(' ')}`);
bF.parserPrintdbg(`${icon}${lnum} ${treeHead}${states.join(' ')}`);
}
}
bF.parserPrintdbgline = function(icon, msg, lnum, recDepth) {
if (parserDoDebugPrint) {
let treeHead = String.fromCharCode(0x2502,32).repeat(recDepth);
bF.parserPrintdbg(`${icon}${lnum} ${treeHead}${msg}`);
}
}
/**
* The starting point to parse those tokens
* @return: BasicAST
*/
bF._parseTokens = function(lnum, tokens, states) {
bF.parserPrintdbg2('Line ', lnum, tokens, states, 0);
/*************************************************************************/
let parenDepth = 0;
let parenStart = -1;
let parenEnd = -1;
let seps = [];
// scan for parens and (:)s
for (let k = 0; k < tokens.length; k++) {
// increase paren depth and mark paren start position
if (tokens[k] == "(" && states[k] != "qot") {
parenDepth += 1;
if (parenStart == -1 && parenDepth == 1) parenStart = k;
}
// decrease paren depth
else if (tokens[k] == ")" && states[k] != "qot") {
if (parenEnd == -1 && parenDepth == 1) parenEnd = k;
parenDepth -= 1;
} }
var _debugSyntaxAnalysis = false; if (parenDepth == 0 && tokens[k] == ":" && states[k] == "seq")
seps.push(k);
if (_debugSyntaxAnalysis) serial.println("@@ SYNTAX ANALYSIS @@");
if (_debugSyntaxAnalysis) {
serial.println("Parser Ln "+lnum+", Rec "+recDepth);
serial.println("Tokens: "+tokens);
serial.println("States: "+states);
} }
if (tokens.length != states.length) throw Error("BasicIntpError: size of tokens and states does not match (line: "+lnum+", recursion depth: "+recDepth+")"); let startPos = [0].concat(seps.map(k => k+1));
if (tokens.length == 0) { let stmtPos = startPos.map((s,i) => {return{start:s, end:(seps[i] || tokens.length)}}); // use end of token position as separator position
if (_debugSyntaxAnalysis) serial.println("*empty tokens*");
var retTreeHead = new BasicAST();
retTreeHead.depth = recDepth;
retTreeHead.lnum = lnum;
return retTreeHead;
}
var k; return stmtPos.map((x,i) => {
var headWord = tokens[0].toLowerCase(); if (stmtPos.length > 1)
var treeHead = new BasicAST(); bF.parserPrintdbgline('Line ', 'Statement #'+(i+1), lnum, 0);
treeHead.astDepth = recDepth;
// check for empty tokens
if (x.end - x.start <= 0) throw new ParserError("Malformed Line");
return bF._parseStmt(lnum,
tokens.slice(x.start, x.end),
states.slice(x.start, x.end),
1
);
});
}
/** Parses following EBNF rule:
stmt =
"IF" , expr_sans_asgn , "THEN" , stmt , ["ELSE" , stmt]
| "DEFUN" , [ident] , "(" , [ident , {" , " , ident}] , ")" , "=" , expr
| "ON" , expr_sans_asgn , ident , expr_sans_asgn , {"," , expr_sans_asgn}
| "(" , stmt , ")"
| expr ;
* @return: BasicAST
*/
bF._parseStmt = function(lnum, tokens, states, recDepth) {
bF.parserPrintdbg2('$', lnum, tokens, states, recDepth);
/*************************************************************************/
let headTkn = tokens[0].toUpperCase();
let headSta = states[0];
let treeHead = new BasicAST();
treeHead.astLnum = lnum; treeHead.astLnum = lnum;
// LITERAL let parenDepth = 0;
if (tokens.length == 1 && (isSemanticLiteral(tokens[0], states[0]))) { let parenStart = -1;
// special case where there were only one word let parenEnd = -1;
if (recDepth == 0) { let onGoPos = -1;
// if that word is literal (e.g. "10 CLEAR"), interpret it as a function let sepsZero = [];
if (states[0] == "lit") { let sepsOne = [];
treeHead.astValue = tokens[0];
// scan for parens that will be used for several rules
// also find nearest THEN and ELSE but also take parens into account
for (let k = 0; k < tokens.length; k++) {
// increase paren depth and mark paren start position
if (tokens[k] == "(" && states[k] != "qot") {
parenDepth += 1;
if (parenStart == -1 && parenDepth == 1) parenStart = k;
}
// decrease paren depth
else if (tokens[k] == ")" && states[k] != "qot") {
if (parenEnd == -1 && parenDepth == 1) parenEnd = k;
parenDepth -= 1;
}
if (parenDepth == 0 && states[k] == "sep")
sepsZero.push(k);
if (parenDepth == 1 && states[k] == "sep")
sepsOne.push(k);
if (parenDepth == 0) {
let tok = tokens[k].toUpperCase();
if (-1 == onGoPos && ("GOTO" == tok || "GOSUB" == tok) && "lit" == states[k])
onGoPos = k;
}
}
// unmatched brackets, duh!
if (parenDepth != 0) throw lang.syntaxfehler(lnum, lang.unmatchedBrackets);
/*************************************************************************/
// ## case for:
// "IF" , expr_sans_asgn , "THEN" , stmt , ["ELSE" , stmt]
try {
bF.parserPrintdbgline('$', "Trying IF Statement...", lnum, recDepth);
return bF._parseIfMode(lnum, tokens, states, recDepth + 1, false);
}
// if ParserError is raised, continue to apply other rules
catch (e) {
bF.parserPrintdbgline('$', 'It was NOT!', lnum, recDepth);
if (!(e instanceof ParserError)) throw e;
}
/*************************************************************************/
// ## case for:
// | "DEFUN" , [ident] , "(" , [ident , {" , " , ident}] , ")" , "=" , expr
if ("DEFUN" == headTkn && "lit" == headSta &&
parenStart == 2 && tokens[parenEnd + 1] == "=" && states[parenEnd + 1] == "op"
) {
bF.parserPrintdbgline('$', 'DEFUN Stmt', lnum, recDepth);
treeHead.astValue = "DEFUN";
treeHead.astType = "function"; treeHead.astType = "function";
// parse function name
if (tokens[1] == "(") {
// anonymous function
treeHead.astLeaves[0] = BasicAST();
treeHead.astLeaves[0].astLnum = lnum;
treeHead.astLeaves[0].astType = "lit";
}
else {
treeHead.astLeaves[0] = bF._parseIdent(lnum, [tokens[1]], [states[1]], recDepth + 1);
}
// parse function arguments
treeHead.astLeaves[0].astLeaves = sepsOne.map(i=>i-1).concat([parenEnd - 1])
.map(i=>bF._parseIdent(lnum, [tokens[i]], [states[i]], recDepth + 2));
// parse function body
treeHead.astLeaves[1] = bF._parseExpr(lnum,
tokens.slice(parenEnd + 2, tokens.length),
states.slice(parenEnd + 2, states.length),
recDepth + 1
);
return treeHead; return treeHead;
} }
// else, screw it
else {
throw lang.syntaxfehler(lnum, "TRAP_LITERALLY_LITERAL");
}
}
if (_debugSyntaxAnalysis) serial.println("literal/number: "+tokens[0]); /*************************************************************************/
treeHead.astValue = ("qot" == states[0]) ? tokens[0] : tokens[0].toUpperCase();
treeHead.astType = ("qot" == states[0]) ? "string" : ("num" == states[0]) ? "num" : "lit";
}
else if (tokens[0].toUpperCase() == "IF" && states[0] != "qot") {
// find ELSE and THEN
var indexElse = undefined;
var indexThen = undefined;
for (k = tokens.length - 1; k >= 1; k--) {
if (indexElse === undefined && tokens[k].toUpperCase() == "ELSE" && states[k] != "qot") {
indexElse = k;
}
else if (indexThen === undefined && tokens[k].toUpperCase() == "THEN" && states[k] != "qot") {
indexThen = k;
}
}
// find GOTO and use it as THEN
var useGoto = false;
if (indexThen === undefined) {
for (k = (indexElse !== undefined) ? indexElse - 1 : tokens.length - 1; k >= 1; k--) {
if (indexThen == undefined && tokens[k].toUpperCase() == "GOTO" && states[k] != "qot") {
useGoto = true;
indexThen = k;
break;
}
}
}
// generate tree // ## case for:
if (indexThen === undefined) throw lang.syntaxfehler(lnum, "IF without THEN"); // | "ON" , if_equation , ident , if_equation , {"," , if_equation}
if ("ON" == headTkn && "lit" == headSta) {
bF.parserPrintdbgline('$', 'ON Stmt', lnum, recDepth);
treeHead.astValue = "if"; if (onGoPos == -1) throw ParserError("Malformed ON Statement");
treeHead.astValue = "ON";
treeHead.astType = "function"; treeHead.astType = "function";
treeHead.astLeaves[0] = bF._parseTokens(
lnum, // parse testvalue
tokens.slice(1, indexThen), let testvalue = bF._parseExpr(lnum,
states.slice(1, indexThen), tokens.slice(1, onGoPos),
recDepth + 1 states.slice(1, onGoPos),
recDepth + 1,
true
); );
if (!useGoto)
treeHead.astLeaves[1] = bF._parseTokens( // parse functionname
lnum, let functionname = bF._parseExpr(lnum,
tokens.slice(indexThen + 1, (indexElse !== undefined) ? indexElse : tokens.length), [tokens[onGoPos]],
states.slice(indexThen + 1, (indexElse !== undefined) ? indexElse : tokens.length), [states[onGoPos]],
recDepth + 1 recDepth + 1,
true
); );
else
treeHead.astLeaves[1] = bF._parseTokens( // parse arguments
lnum, // get list of comma but filter ones appear before GOTO/GOSUB
[].concat("goto", tokens.slice(indexThen + 1, (indexElse !== undefined) ? indexElse : tokens.length)), let onArgSeps = sepsZero.filter(i => (i > onGoPos));
[].concat("lit", states.slice(indexThen + 1, (indexElse !== undefined) ? indexElse : tokens.length)), let onArgStartPos = [onGoPos + 1].concat(onArgSeps.map(k => k + 1));
recDepth + 1 let onArgPos = onArgStartPos.map((s,i) => {return{start:s, end: (onArgSeps[i] || tokens.length)}}); // use end of token position as separator position
// recursively parse expressions
treeHead.astLeaves = [testvalue, functionname].concat(onArgPos.map((x,i) => {
bF.parserPrintdbgline('$', 'ON GOTO/GOSUB Arguments #'+(i+1), lnum, recDepth);
// check for empty tokens
if (x.end - x.start <= 0) throw new ParserError("Malformed ON arguments");
return bF._parseExpr(lnum,
tokens.slice(x.start, x.end),
states.slice(x.start, x.end),
recDepth + 1,
true
); );
if (indexElse !== undefined) { }));
treeHead.astLeaves[2] = bF._parseTokens(
lnum, return treeHead;
tokens.slice(indexElse + 1, tokens.length), }
states.slice(indexElse + 1, tokens.length),
/*************************************************************************/
// ## case for:
// | "(" , stmt , ")"
if (parenStart == 0 && parenEnd == tokens.length - 1) {
bF.parserPrintdbgline('$', '( Stmt )', lnum, recDepth);
return bF._parseStmt(lnum,
tokens.slice(parenStart + 1, parenEnd),
states.slice(parenStart + 1, parenEnd),
recDepth + 1 recDepth + 1
); );
} }
/*************************************************************************/
// ## case for:
// | expr ;
try {
bF.parserPrintdbgline('$', 'Trying Expression Call...', lnum, recDepth);
return bF._parseExpr(lnum, tokens, states, recDepth + 1);
} }
else { catch (e) {
bF.parserPrintdbgline('$', 'Error!', lnum, recDepth);
throw new ParserError("Statement cannot be parsed: "+e.stack);
}
/*************************************************************************/
throw new ParserError("Statement cannot be parsed: "+e.stack);
} // END of STMT
/** Parses following EBNF rule:
"IF" , expr_sans_asgn , "THEN" , stmt , ["ELSE" , stmt]
| "IF" , expr_sans_asgn , "THEN" , expr , ["ELSE" , expr]
if exprMode is true, only the latter will be used; former otherwise
* @return: BasicAST
*/
bF._parseIfMode = function(lnum, tokens, states, recDepth, exprMode) {
bF.parserPrintdbg2('/', lnum, tokens, states, recDepth);
/*************************************************************************/
let headTkn = tokens[0].toUpperCase();
let headSta = states[0];
let parseFunction = (exprMode) ? bF._parseExpr : bF._parseStmt
let thenPos = -1;
let elsePos = -1;
let parenDepth = 0;
let parenStart = -1;
let parenEnd = -1;
// scan for parens that will be used for several rules
// also find nearest THEN and ELSE but also take parens into account
for (let k = 0; k < tokens.length; k++) {
// increase paren depth and mark paren start position
if (tokens[k] == "(" && states[k] != "qot") {
parenDepth += 1;
if (parenStart == -1 && parenDepth == 1) parenStart = k;
}
// decrease paren depth
else if (tokens[k] == ")" && states[k] != "qot") {
if (parenEnd == -1 && parenDepth == 1) parenEnd = k;
parenDepth -= 1;
}
if (parenDepth == 0) {
if (-1 == thenPos && "THEN" == tokens[k].toUpperCase() && "lit" == states[k])
thenPos = k;
else if (-1 == elsePos && "ELSE" == tokens[k].toUpperCase() && "lit" == states[k])
elsePos = k;
}
}
// unmatched brackets, duh!
if (parenDepth != 0) throw lang.syntaxfehler(lnum, lang.unmatchedBrackets);
let treeHead = new BasicAST();
treeHead.astLnum = lnum;
// ## case for:
// "IF" , expr_sans_asgn , "THEN" , stmt , ["ELSE" , stmt]
if ("IF" == headTkn && "lit" == headSta) {
// "THEN" not found, raise error!
if (thenPos == -1) throw new ParserError("IF without THEN in " + lnum);
treeHead.astValue = "IF";
treeHead.astType = "function";
treeHead.astLeaves[0] = bF._parseExpr(lnum,
tokens.slice(1, thenPos),
states.slice(1, thenPos),
recDepth + 1,
true // if_equation mode
);
treeHead.astLeaves[1] = parseFunction(lnum,
tokens.slice(thenPos + 1, (elsePos != -1) ? elsePos : tokens.length),
states.slice(thenPos + 1, (elsePos != -1) ? elsePos : tokens.length),
recDepth + 1
);
if (elsePos != -1)
treeHead.astLeaves[2] = parseFunction(lnum,
tokens.slice(elsePos + 1, tokens.length),
states.slice(elsePos + 1, tokens.length),
recDepth + 1
);
return treeHead;
}
throw new ParserError("not an IF "+(exprMode) ? "expression" : "statement");
} // END of IF
/** Parses following EBNF rule:
expr = (* this basically blocks some funny attemps such as using DEFUN as anon function because everything is global in BASIC *)
lit
| "(" , expr , ")"
| "IF" , expr_sans_asgn , "THEN" , expr , ["ELSE" , expr]
| expr , op , expr
| op_uni , expr
| function_call ; //TODO
* @return: BasicAST
*/
bF._parseExpr = function(lnum, tokens, states, recDepth, ifMode) {
bF.parserPrintdbg2('E', lnum, tokens, states, recDepth);
/*************************************************************************/
// ## case for:
// lit
let headTkn = tokens[0].toUpperCase();
if (!bF._EquationIllegalTokens.includes(headTkn) && tokens.length == 1) {
bF.parserPrintdbgline('E', 'Literal Call', lnum, recDepth);
return bF._parseLit(lnum, tokens, states, recDepth + 1);
}
/*************************************************************************/
// scan for operators with highest precedence, use rightmost one if multiple were found // scan for operators with highest precedence, use rightmost one if multiple were found
var topmostOp; let topmostOp;
var topmostOpPrc = 0; let topmostOpPrc = 0;
var operatorPos = -1; let operatorPos = -1;
// find and mark position of separators and parentheses // find and mark position of parentheses
// properly deal with the nested function calls // properly deal with the nested function calls
var parenDepth = 0; let parenDepth = 0;
var parenStart = -1; let parenStart = -1;
var parenEnd = -1; let parenEnd = -1;
var separators = [];
// initial scan for adding omitted parens // Scan for unmatched parens and mark off the right operator we must deal with
for (k = 0; k < tokens.length; k++) { // every function_call need to re-scan because it is recursively called
for (let k = 0; k < tokens.length; k++) {
// increase paren depth and mark paren start position
if (tokens[k] == "(" && states[k] != "qot") { if (tokens[k] == "(" && states[k] != "qot") {
parenDepth += 1; parenDepth += 1;
if (parenStart == -1 && parenDepth == 1) parenStart = k; if (parenStart == -1 && parenDepth == 1) parenStart = k;
} }
// decrease paren depth
else if (tokens[k] == ")" && states[k] != "qot") { else if (tokens[k] == ")" && states[k] != "qot") {
if (parenEnd == -1 && parenDepth == 1) parenEnd = k; if (parenEnd == -1 && parenDepth == 1) parenEnd = k;
parenDepth -= 1; parenDepth -= 1;
} }
// determine the right operator to deal with
if (parenDepth == 0) { if (parenDepth == 0) {
if (states[k] == "op" && isSemanticLiteral(tokens[k-1], states[k-1]) && if (states[k] == "op" && bF.isSemanticLiteral(tokens[k-1], states[k-1]) &&
((bF._opPrc[tokens[k].toUpperCase()] > topmostOpPrc) || ((bF._opPrc[tokens[k].toUpperCase()] > topmostOpPrc) ||
(!bF._opRh[tokens[k].toUpperCase()] && bF._opPrc[tokens[k].toUpperCase()] == topmostOpPrc)) (!bF._opRh[tokens[k].toUpperCase()] && bF._opPrc[tokens[k].toUpperCase()] == topmostOpPrc))
) { ) {
@@ -1841,142 +2111,218 @@ bF._parseTokens = function(lnum, tokens, states, recDepth) {
} }
} }
// == AUTOPAREN == // unmatched brackets, duh!
// TODO do it properly by counting number of arguments and whatnot
if (parenDepth != 0) throw lang.syntaxfehler(lnum, lang.unmatchedBrackets); if (parenDepth != 0) throw lang.syntaxfehler(lnum, lang.unmatchedBrackets);
if (_debugSyntaxAnalysis) serial.println("Paren position: "+parenStart+", "+parenEnd); if (_debugSyntaxAnalysis) serial.println("Equation NEW Paren position: "+parenStart+", "+parenEnd);
// if there is no paren or paren does NOT start index 1 /*************************************************************************/
// e.g. negative three should NOT require to be written as "-(3)"
if ((parenStart > 1 || parenStart == -1) && (operatorPos != 1 && operatorPos != 0) && states[0] == "lit" && states[1] != "op") {
// make a paren!
tokens = [].concat(tokens[0], "(", tokens.slice(1, tokens.length), ")");
states = [].concat(states[0], "paren", states.slice(1, states.length), "paren");
if (_debugSyntaxAnalysis) serial.println("inserting paren at right place"); // ## case for:
if (_debugSyntaxAnalysis) serial.println(tokens.join(",")); // | "(" , expr , ")"
if (parenStart == 0 && parenEnd == tokens.length - 1) {
bF.parserPrintdbgline('E', '( Expr )', lnum, recDepth);
return bF._parseTokens(lnum, tokens, states, recDepth); return bF._parseEquation(lnum,
tokens.slice(parenStart + 1, parenEnd),
states.slice(parenStart + 1, parenEnd),
recDepth + 1
);
} }
// get the position of parens and separators /*************************************************************************/
parenStart = -1; parenEnd = -1; parenDepth = 0;
topmostOpPrc = 0; operatorPos = -1; // ## case for:
// running again but now with newly added parens // | "IF" , expr_sans_asgn , "THEN" , expr , ["ELSE" , expr]
for (k = 0; k < tokens.length; k++) { try {
if (tokens[k] == "(" && states[k] != "qot") { bF.parserPrintdbgline('E', "Trying IF Expression...", lnum, recDepth);
parenDepth += 1; return bF._parseIfMode(lnum, tokens, states, recDepth + 1, false);
if (parenStart == -1 && parenDepth == 1) parenStart = k;
} }
else if (tokens[k] == ")" && states[k] != "qot") { // if ParserError is raised, continue to apply other rules
if (parenEnd == -1 && parenDepth == 1) parenEnd = k; catch (e) {
parenDepth -= 1; bF.parserPrintdbgline('E', 'It was NOT!', lnum, recDepth);
if (!(e instanceof ParserError)) throw e;
} }
if (parenDepth == 1 && states[k] == "sep") { /*************************************************************************/
separators.push(k);
}
if (parenDepth == 0) {
if (states[k] == "op" && isSemanticLiteral(tokens[k-1], states[k-1]) &&
((bF._opPrc[tokens[k].toUpperCase()] > topmostOpPrc) ||
(!bF._opRh[tokens[k].toUpperCase()] && bF._opPrc[tokens[k].toUpperCase()] == topmostOpPrc))
) {
topmostOp = tokens[k].toUpperCase();
topmostOpPrc = bF._opPrc[tokens[k].toUpperCase()];
operatorPos = k;
}
}
}
if (parenDepth != 0) throw lang.syntaxfehler(lnum, lang.unmatchedBrackets); // ## case for:
if (_debugSyntaxAnalysis) serial.println("NEW Paren position: "+parenStart+", "+parenEnd); // | expr , op, expr
// | op_uni , expr
// BINARY_OP/UNARY_OP // if operator is found, split by the operator and recursively parse the LH and RH
if (topmostOp !== undefined) { if (topmostOp !== undefined) {
bF.parserPrintdbgline('E', 'Operators', lnum, recDepth);
if (_debugSyntaxAnalysis) serial.println("operator: "+topmostOp+", pos: "+operatorPos); if (_debugSyntaxAnalysis) serial.println("operator: "+topmostOp+", pos: "+operatorPos);
if (ifMode && topmostOp == "=") throw lang.syntaxfehler(lnum, "'=' used on IF, did you mean '=='?");
if (ifMode && topmostOp == ":") throw lang.syntaxfehler(lnum, "':' used on IF");
// this is the AST we're going to build up and return
// (other IF clauses don't use this)
let treeHead = new BasicAST();
treeHead.astLnum = lnum;
treeHead.astValue = topmostOp;
treeHead.astType = "op";
// BINARY_OP? // BINARY_OP?
if (operatorPos > 0) { if (operatorPos > 0) {
var subtknL = tokens.slice(0, operatorPos); let subtknL = tokens.slice(0, operatorPos);
var subtknR = tokens.slice(operatorPos + 1, tokens.length); let substaL = states.slice(0, operatorPos);
var substaL = states.slice(0, operatorPos); let subtknR = tokens.slice(operatorPos + 1, tokens.length);
var substaR = states.slice(operatorPos + 1, tokens.length); let substaR = states.slice(operatorPos + 1, tokens.length);
treeHead.astValue = topmostOp; treeHead.astLeaves[0] = bF._parseExpr(lnum, subtknL, substaL, recDepth + 1);
treeHead.astType = "op"; treeHead.astLeaves[1] = bF._parseExpr(lnum, subtknR, substaR, recDepth + 1);
treeHead.astLeaves[0] = bF._parseTokens(lnum, subtknL, substaL, recDepth + 1);
treeHead.astLeaves[1] = bF._parseTokens(lnum, subtknR, substaR, recDepth + 1);
} }
else { else {
if (_debugSyntaxAnalysis) serial.println("re-parenthesising unary op"); treeHead.astValue = (topmostOp === "-") ? "UNARYMINUS" : "UNARYPLUS";
treeHead.astLeaves[0] = bF._parseExpr(lnum,
// parenthesize the unary op tokens.slice(operatorPos + 1, tokens.length),
var unaryParenEnd = 1; states.slice(operatorPos + 1, states.length),
while (unaryParenEnd < tokens.length) { recDepth + 1
if (states[unaryParenEnd] == "op" && bF._opPrc[tokens[unaryParenEnd]] > 1) );
break;
unaryParenEnd += 1;
} }
var newTokens = [].concat("(", tokens.slice(0, unaryParenEnd), ")", tokens.slice(unaryParenEnd, tokens.length));
var newStates = [].concat("paren", states.slice(0, unaryParenEnd), "paren", states.slice(unaryParenEnd, tokens.length));
return bF._parseTokens(lnum, newTokens, newStates, recDepth + 1);
}
}
// FUNCTION CALL
else {
if (_debugSyntaxAnalysis) serial.println("function call");
var currentFunction = (states[0] == "paren") ? undefined : tokens[0];
treeHead.astValue = ("-" == currentFunction) ? "UNARYMINUS" : ("+" == currentFunction) ? "UNARYPLUS" : currentFunction;
treeHead.astType = (currentFunction === undefined) ? "null" : "function";
if (_debugSyntaxAnalysis) serial.println("function name: "+treeHead.astValue);
var leaves = [];
var seps = [];
// if there is no paren (this part deals with unary ops ONLY!)
if (parenStart == -1 && parenEnd == -1) {
var subtkn = tokens.slice(1, tokens.length);
var substa = states.slice(1, tokens.length);
if (_debugSyntaxAnalysis) serial.println("subtokenA: "+subtkn.join("/"));
leaves.push(bF._parseTokens(lnum, subtkn, substa, recDepth + 1))
}
else if (parenEnd > parenStart) {
separators = [parenStart].concat(separators, [parenEnd]);
if (_debugSyntaxAnalysis) serial.println("separators: "+separators.join(","));
// recursively parse comma-separated arguments
// print ( plus ( 3 , 2 ) , times ( 8 , 7 ) )
// s ^ e
// separators = [1,8,15]
// plus ( 3 , 2 ) / times ( 8 , 7 )
// s ^ e s ^ e
// separators = [1,5] ; [1,5]
// 3 / 2 / 8 / 7
for (k = 1; k < separators.length; k++) {
var subtkn = tokens.slice(separators[k - 1] + 1, separators[k]);
var substa = states.slice(separators[k - 1] + 1, separators[k]);
if (_debugSyntaxAnalysis) serial.println("subtokenB: "+subtkn.join("/"));
leaves.push(bF._parseTokens(lnum, subtkn, substa, recDepth + 1));
}
separators.slice(1, separators.length - 1).forEach((v) => { if (v !== undefined) seps.push(tokens[v]) });
}
else throw lang.badFunctionCallFormat();
treeHead.astLeaves = leaves;//.filter(function(__v) { return __v !== undefined; });
treeHead.astSeps = seps;
}
}
return treeHead; return treeHead;
}
/*************************************************************************/
// ## case for:
// | function_call ;
try {
bF.parserPrintdbgline('E', "Trying Function Call...", lnum, recDepth);
return bF._parseFunctionCall(lnum, tokens, states, recDepth + 1);
}
catch (e) {
throw new ParserError("Expression cannot be parsed: "+e.stack);
}
/*************************************************************************/
throw new ParserError("Expression cannot be parsed: "+e.stack);
} // END of EXPR
/** Parses following EBNF rule:
function_call =
ident , "(" , [expr , {argsep , expr} , [argsep]] , ")"
| ident , expr , {argsep , expr} , [argsep] ;
* @return: BasicAST
*/
bF._parseFunctionCall = function(lnum, tokens, states, recDepth) {
bF.parserPrintdbg2(String.fromCharCode(0x192), lnum, tokens, states, recDepth);
/*************************************************************************/
let parenDepth = 0;
let parenStart = -1;
let parenEnd = -1;
let _argsepsOnLevelZero = []; // argseps collected when parenDepth == 0
let _argsepsOnLevelOne = []; // argseps collected when parenDepth == 1
// Scan for unmatched parens and mark off the right operator we must deal with
// every function_call need to re-scan because it is recursively called
for (let k = 0; k < tokens.length; k++) {
// increase paren depth and mark paren start position
if (tokens[k] == "(" && states[k] != "qot") {
parenDepth += 1;
if (parenStart == -1 && parenDepth == 1) parenStart = k;
}
// decrease paren depth
else if (tokens[k] == ")" && states[k] != "qot") {
if (parenEnd == -1 && parenDepth == 1) parenEnd = k;
parenDepth -= 1;
}
if (parenDepth == 0 && states[k] == "sep")
_argsepsOnLevelZero.push(k);
if (parenDepth == 1 && states[k] == "sep")
_argsepsOnLevelOne.push(k);
}
// unmatched brackets, duh!
if (parenDepth != 0) throw lang.syntaxfehler(lnum, lang.unmatchedBrackets);
let parenUsed = (parenStart == 1 && parenEnd == states.length - 1);
/*************************************************************************/
// ## case for:
// ident , "(" , [expr , {argsep , expr} , [argsep]] , ")"
// | ident , expr , {argsep , expr} , [argsep] ;
bF.parserPrintdbgline(String.fromCharCode(0x192), 'Function Call', lnum, recDepth);
let treeHead = new BasicAST();
treeHead.astLnum = lnum;
// set function name and also check for syntax by deliberately parsing the word
treeHead.astValue = bF._parseIdent(lnum, [tokens[0]], [states[0]], recDepth + 1).astValue; // always UPPERCASE
// 5 8 11 [end]
let argSeps = parenUsed ? _argsepsOnLevelOne : _argsepsOnLevelZero; // choose which "sep tray" to use
// 1 6 9 12
let argStartPos = [1 + (parenUsed)].concat(argSeps.map(k => k+1));
// [1,5) [6,8) [9,11) [12,end)
let argPos = argStartPos.map((s,i) => {return{start:s, end:(argSeps[i] || tokens.length - (parenUsed))}}); // use end of token position as separator position
// check for trailing separator
let hasTrailingSep = (states[states.length - 1 - (parenUsed)] == "sep");
// exclude last separator from recursion if input tokens has trailing separator
if (hasTrailingSep) argPos.pop();
// recursively parse function arguments
treeHead.astLeaves = argPos.map((x,i) => {
bF.parserPrintdbgline(String.fromCharCode(0x192), 'Function Arguments #'+(i+1), lnum, recDepth);
// check for empty tokens
if (x.end - x.start <= 0) throw new ParserError("not a function call because it's malformed");
return bF._parseExpr(lnum,
tokens.slice(x.start, x.end),
states.slice(x.start, x.end),
recDepth + 1
)}
);
treeHead.astType = "function";
treeHead.astSeps = argSeps.map(i => tokens[i]);
return treeHead;
}
bF._parseIdent = function(lnum, tokens, states, recDepth) {
bF.parserPrintdbg2('i', lnum, tokens, states, recDepth);
if (!Array.isArray(tokens) && !Array.isArray(states)) throw new ParserError("Tokens and states are not array");
if (tokens.length != 1 || states[0] != "lit") throw new ParserError(`illegal tokens '${tokens}' with states '${states}' in ${lnum}`);
let treeHead = new BasicAST();
treeHead.astLnum = lnum;
treeHead.astValue = tokens[0].toUpperCase();
treeHead.astType = "lit";
return treeHead;
}
/**
* @return: BasicAST
*/
bF._parseLit = function(lnum, tokens, states, recDepth) {
bF.parserPrintdbg2(String.fromCharCode(0xA2), lnum, tokens, states, recDepth);
if (!Array.isArray(tokens) && !Array.isArray(states)) throw new ParserError("Tokens and states are not array");
if (tokens.length != 1) throw new ParserError("parseLit 1");
let treeHead = new BasicAST();
treeHead.astLnum = lnum;
if (_debugSyntaxAnalysis) serial.println("literal/number: "+tokens[0]);
treeHead.astValue = ("qot" == states[0]) ? tokens[0] : tokens[0].toUpperCase();
treeHead.astType = ("qot" == states[0]) ? "string" : ("num" == states[0]) ? "num" : "lit";
return treeHead;
}
};
// @return is defined in BasicAST // @return is defined in BasicAST
let JStoBASICtype = function(object) { let JStoBASICtype = function(object) {
if (typeof object === "boolean") return "bool"; if (typeof object === "boolean") return "bool";

View File

@@ -5,17 +5,64 @@ class ParserError extends Error {
} }
} }
let bF = {}; let bF = {};
let printdbg = any => serial.println(any); bF.parserPrintdbg = any => serial.println(any);
let printdbg2 = function(icon, lnum, tokens, states, recDepth) { bF.parserPrintdbg2 = function(icon, lnum, tokens, states, recDepth) {
let treeHead = String.fromCharCode(0x2502,32).repeat(recDepth); let treeHead = String.fromCharCode(0x2502,32).repeat(recDepth);
printdbg(`${icon}${lnum} ${treeHead}${tokens.join(' ')}`); bF.parserPrintdbg(`${icon}${lnum} ${treeHead}${tokens.join(' ')}`);
printdbg(`${icon}${lnum} ${treeHead}${states.join(' ')}`); bF.parserPrintdbg(`${icon}${lnum} ${treeHead}${states.join(' ')}`);
} }
let printdbgline = function(icon, msg, lnum, recDepth) { bF.parserPrintdbgline = function(icon, msg, lnum, recDepth) {
let treeHead = String.fromCharCode(0x2502,32).repeat(recDepth); let treeHead = String.fromCharCode(0x2502,32).repeat(recDepth);
printdbg(`${icon}${lnum} ${treeHead}${msg}`); bF.parserPrintdbg(`${icon}${lnum} ${treeHead}${msg}`);
} }
/**
* @return ARRAY of BasicAST
*/
bF._parseTokens = function(lnum, tokens, states) {
bF.parserPrintdbg2('Line ', lnum, tokens, states, 0);
/*************************************************************************/
let parenDepth = 0;
let parenStart = -1;
let parenEnd = -1;
let seps = [];
// scan for parens and (:)s
for (let k = 0; k < tokens.length; k++) {
// increase paren depth and mark paren start position
if (tokens[k] == "(" && states[k] != "qot") {
parenDepth += 1;
if (parenStart == -1 && parenDepth == 1) parenStart = k;
}
// decrease paren depth
else if (tokens[k] == ")" && states[k] != "qot") {
if (parenEnd == -1 && parenDepth == 1) parenEnd = k;
parenDepth -= 1;
}
if (parenDepth == 0 && tokens[k] == ":" && states[k] == "seq")
seps.push(k);
}
let startPos = [0].concat(seps.map(k => k+1));
let stmtPos = startPos.map((s,i) => {return{start:s, end:(seps[i] || tokens.length)}}); // use end of token position as separator position
return stmtPos.map((x,i) => {
if (stmtPos.length > 1)
bF.parserPrintdbgline('Line ', 'Statement #'+(i+1), lnum, 0);
// check for empty tokens
if (x.end - x.start <= 0) throw new ParserError("Malformed Line");
return bF._parseStmt(lnum,
tokens.slice(x.start, x.end),
states.slice(x.start, x.end),
1
);
});
}
/** Parses following EBNF rule: /** Parses following EBNF rule:
stmt = stmt =
@@ -27,7 +74,7 @@ stmt =
* @return: BasicAST * @return: BasicAST
*/ */
bF._parseStmt = function(lnum, tokens, states, recDepth) { bF._parseStmt = function(lnum, tokens, states, recDepth) {
printdbg2('$', lnum, tokens, states, recDepth); bF.parserPrintdbg2('$', lnum, tokens, states, recDepth);
/*************************************************************************/ /*************************************************************************/
@@ -78,12 +125,12 @@ bF._parseStmt = function(lnum, tokens, states, recDepth) {
// ## case for: // ## case for:
// "IF" , expr_sans_asgn , "THEN" , stmt , ["ELSE" , stmt] // "IF" , expr_sans_asgn , "THEN" , stmt , ["ELSE" , stmt]
try { try {
printdbgline('$', "Trying IF Statement...", lnum, recDepth); bF.parserPrintdbgline('$', "Trying IF Statement...", lnum, recDepth);
return bF._parseIfMode(lnum, tokens, states, recDepth + 1, false); return bF._parseIfMode(lnum, tokens, states, recDepth + 1, false);
} }
// if ParserError is raised, continue to apply other rules // if ParserError is raised, continue to apply other rules
catch (e) { catch (e) {
printdbgline('$', 'It was NOT!', lnum, recDepth); bF.parserPrintdbgline('$', 'It was NOT!', lnum, recDepth);
if (!(e instanceof ParserError)) throw e; if (!(e instanceof ParserError)) throw e;
} }
@@ -94,7 +141,7 @@ bF._parseStmt = function(lnum, tokens, states, recDepth) {
if ("DEFUN" == headTkn && "lit" == headSta && if ("DEFUN" == headTkn && "lit" == headSta &&
parenStart == 2 && tokens[parenEnd + 1] == "=" && states[parenEnd + 1] == "op" parenStart == 2 && tokens[parenEnd + 1] == "=" && states[parenEnd + 1] == "op"
) { ) {
printdbgline('$', 'DEFUN Stmt', lnum, recDepth); bF.parserPrintdbgline('$', 'DEFUN Stmt', lnum, recDepth);
treeHead.astValue = "DEFUN"; treeHead.astValue = "DEFUN";
treeHead.astType = "function"; treeHead.astType = "function";
@@ -129,7 +176,7 @@ bF._parseStmt = function(lnum, tokens, states, recDepth) {
// ## case for: // ## case for:
// | "ON" , if_equation , ident , if_equation , {"," , if_equation} // | "ON" , if_equation , ident , if_equation , {"," , if_equation}
if ("ON" == headTkn && "lit" == headSta) { if ("ON" == headTkn && "lit" == headSta) {
printdbgline('$', 'ON Stmt', lnum, recDepth); bF.parserPrintdbgline('$', 'ON Stmt', lnum, recDepth);
if (onGoPos == -1) throw ParserError("Malformed ON Statement"); if (onGoPos == -1) throw ParserError("Malformed ON Statement");
@@ -160,7 +207,7 @@ bF._parseStmt = function(lnum, tokens, states, recDepth) {
// recursively parse expressions // recursively parse expressions
treeHead.astLeaves = [testvalue, functionname].concat(onArgPos.map((x,i) => { treeHead.astLeaves = [testvalue, functionname].concat(onArgPos.map((x,i) => {
printdbgline('$', 'ON GOTO/GOSUB Arguments #'+i, lnum, recDepth); bF.parserPrintdbgline('$', 'ON GOTO/GOSUB Arguments #'+(i+1), lnum, recDepth);
// check for empty tokens // check for empty tokens
if (x.end - x.start <= 0) throw new ParserError("Malformed ON arguments"); if (x.end - x.start <= 0) throw new ParserError("Malformed ON arguments");
@@ -181,7 +228,7 @@ bF._parseStmt = function(lnum, tokens, states, recDepth) {
// ## case for: // ## case for:
// | "(" , stmt , ")" // | "(" , stmt , ")"
if (parenStart == 0 && parenEnd == tokens.length - 1) { if (parenStart == 0 && parenEnd == tokens.length - 1) {
printdbgline('$', '( Stmt )', lnum, recDepth); bF.parserPrintdbgline('$', '( Stmt )', lnum, recDepth);
return bF._parseStmt(lnum, return bF._parseStmt(lnum,
tokens.slice(parenStart + 1, parenEnd), tokens.slice(parenStart + 1, parenEnd),
states.slice(parenStart + 1, parenEnd), states.slice(parenStart + 1, parenEnd),
@@ -194,11 +241,11 @@ bF._parseStmt = function(lnum, tokens, states, recDepth) {
// ## case for: // ## case for:
// | expr ; // | expr ;
try { try {
printdbgline('$', 'Trying Expression Call...', lnum, recDepth); bF.parserPrintdbgline('$', 'Trying Expression Call...', lnum, recDepth);
return bF._parseExpr(lnum, tokens, states, recDepth + 1); return bF._parseExpr(lnum, tokens, states, recDepth + 1);
} }
catch (e) { catch (e) {
printdbgline('$', 'Error!', lnum, recDepth); bF.parserPrintdbgline('$', 'Error!', lnum, recDepth);
throw new ParserError("Statement cannot be parsed: "+e.stack); throw new ParserError("Statement cannot be parsed: "+e.stack);
} }
@@ -214,7 +261,7 @@ bF._parseStmt = function(lnum, tokens, states, recDepth) {
* @return: BasicAST * @return: BasicAST
*/ */
bF._parseIfMode = function(lnum, tokens, states, recDepth, exprMode) { bF._parseIfMode = function(lnum, tokens, states, recDepth, exprMode) {
printdbg2('/', lnum, tokens, states, recDepth); bF.parserPrintdbg2('/', lnum, tokens, states, recDepth);
/*************************************************************************/ /*************************************************************************/
@@ -304,7 +351,7 @@ expr = (* this basically blocks some funny attemps such as using DEFUN as anon f
* @return: BasicAST * @return: BasicAST
*/ */
bF._parseExpr = function(lnum, tokens, states, recDepth, ifMode) { bF._parseExpr = function(lnum, tokens, states, recDepth, ifMode) {
printdbg2('E', lnum, tokens, states, recDepth); bF.parserPrintdbg2('E', lnum, tokens, states, recDepth);
/*************************************************************************/ /*************************************************************************/
@@ -312,7 +359,7 @@ bF._parseExpr = function(lnum, tokens, states, recDepth, ifMode) {
// lit // lit
let headTkn = tokens[0].toUpperCase(); let headTkn = tokens[0].toUpperCase();
if (!bF._EquationIllegalTokens.includes(headTkn) && tokens.length == 1) { if (!bF._EquationIllegalTokens.includes(headTkn) && tokens.length == 1) {
printdbgline('E', 'Literal Call', lnum, recDepth); bF.parserPrintdbgline('E', 'Literal Call', lnum, recDepth);
return bF._parseLit(lnum, tokens, states, recDepth + 1); return bF._parseLit(lnum, tokens, states, recDepth + 1);
} }
@@ -365,7 +412,7 @@ bF._parseExpr = function(lnum, tokens, states, recDepth, ifMode) {
// ## case for: // ## case for:
// | "(" , expr , ")" // | "(" , expr , ")"
if (parenStart == 0 && parenEnd == tokens.length - 1) { if (parenStart == 0 && parenEnd == tokens.length - 1) {
printdbgline('E', '( Expr )', lnum, recDepth); bF.parserPrintdbgline('E', '( Expr )', lnum, recDepth);
return bF._parseEquation(lnum, return bF._parseEquation(lnum,
tokens.slice(parenStart + 1, parenEnd), tokens.slice(parenStart + 1, parenEnd),
@@ -379,12 +426,12 @@ bF._parseExpr = function(lnum, tokens, states, recDepth, ifMode) {
// ## case for: // ## case for:
// | "IF" , expr_sans_asgn , "THEN" , expr , ["ELSE" , expr] // | "IF" , expr_sans_asgn , "THEN" , expr , ["ELSE" , expr]
try { try {
printdbgline('E', "Trying IF Expression...", lnum, recDepth); bF.parserPrintdbgline('E', "Trying IF Expression...", lnum, recDepth);
return bF._parseIfMode(lnum, tokens, states, recDepth + 1, false); return bF._parseIfMode(lnum, tokens, states, recDepth + 1, false);
} }
// if ParserError is raised, continue to apply other rules // if ParserError is raised, continue to apply other rules
catch (e) { catch (e) {
printdbgline('E', 'It was NOT!', lnum, recDepth); bF.parserPrintdbgline('E', 'It was NOT!', lnum, recDepth);
if (!(e instanceof ParserError)) throw e; if (!(e instanceof ParserError)) throw e;
} }
@@ -395,7 +442,7 @@ bF._parseExpr = function(lnum, tokens, states, recDepth, ifMode) {
// | op_uni , expr // | op_uni , expr
// if operator is found, split by the operator and recursively parse the LH and RH // if operator is found, split by the operator and recursively parse the LH and RH
if (topmostOp !== undefined) { if (topmostOp !== undefined) {
printdbgline('E', 'Operators', lnum, recDepth); bF.parserPrintdbgline('E', 'Operators', lnum, recDepth);
if (_debugSyntaxAnalysis) serial.println("operator: "+topmostOp+", pos: "+operatorPos); if (_debugSyntaxAnalysis) serial.println("operator: "+topmostOp+", pos: "+operatorPos);
@@ -437,7 +484,7 @@ bF._parseExpr = function(lnum, tokens, states, recDepth, ifMode) {
// ## case for: // ## case for:
// | function_call ; // | function_call ;
try { try {
printdbgline('E', "Trying Function Call...", lnum, recDepth); bF.parserPrintdbgline('E', "Trying Function Call...", lnum, recDepth);
return bF._parseFunctionCall(lnum, tokens, states, recDepth + 1); return bF._parseFunctionCall(lnum, tokens, states, recDepth + 1);
} }
catch (e) { catch (e) {
@@ -458,7 +505,7 @@ function_call =
* @return: BasicAST * @return: BasicAST
*/ */
bF._parseFunctionCall = function(lnum, tokens, states, recDepth) { bF._parseFunctionCall = function(lnum, tokens, states, recDepth) {
printdbg2(String.fromCharCode(0x192), lnum, tokens, states, recDepth); bF.parserPrintdbg2(String.fromCharCode(0x192), lnum, tokens, states, recDepth);
/*************************************************************************/ /*************************************************************************/
@@ -497,7 +544,7 @@ bF._parseFunctionCall = function(lnum, tokens, states, recDepth) {
// ## case for: // ## case for:
// ident , "(" , [expr , {argsep , expr} , [argsep]] , ")" // ident , "(" , [expr , {argsep , expr} , [argsep]] , ")"
// | ident , expr , {argsep , expr} , [argsep] ; // | ident , expr , {argsep , expr} , [argsep] ;
printdbgline(String.fromCharCode(0x192), 'Function Call', lnum, recDepth); bF.parserPrintdbgline(String.fromCharCode(0x192), 'Function Call', lnum, recDepth);
let treeHead = new BasicAST(); let treeHead = new BasicAST();
treeHead.astLnum = lnum; treeHead.astLnum = lnum;
@@ -519,7 +566,7 @@ bF._parseFunctionCall = function(lnum, tokens, states, recDepth) {
// recursively parse function arguments // recursively parse function arguments
treeHead.astLeaves = argPos.map((x,i) => { treeHead.astLeaves = argPos.map((x,i) => {
printdbgline(String.fromCharCode(0x192), 'Function Arguments #'+i, lnum, recDepth); bF.parserPrintdbgline(String.fromCharCode(0x192), 'Function Arguments #'+(i+1), lnum, recDepth);
// check for empty tokens // check for empty tokens
if (x.end - x.start <= 0) throw new ParserError("not a function call because it's malformed"); if (x.end - x.start <= 0) throw new ParserError("not a function call because it's malformed");
@@ -538,7 +585,7 @@ bF._parseFunctionCall = function(lnum, tokens, states, recDepth) {
bF._parseIdent = function(lnum, tokens, states, recDepth) { bF._parseIdent = function(lnum, tokens, states, recDepth) {
printdbg2('i', lnum, tokens, states, recDepth); bF.parserPrintdbg2('i', lnum, tokens, states, recDepth);
if (!Array.isArray(tokens) && !Array.isArray(states)) throw new ParserError("Tokens and states are not array"); if (!Array.isArray(tokens) && !Array.isArray(states)) throw new ParserError("Tokens and states are not array");
if (tokens.length != 1 || states[0] != "lit") throw new ParserError(`illegal tokens '${tokens}' with states '${states}' in ${lnum}`); if (tokens.length != 1 || states[0] != "lit") throw new ParserError(`illegal tokens '${tokens}' with states '${states}' in ${lnum}`);
@@ -554,7 +601,7 @@ bF._parseIdent = function(lnum, tokens, states, recDepth) {
* @return: BasicAST * @return: BasicAST
*/ */
bF._parseLit = function(lnum, tokens, states, recDepth) { bF._parseLit = function(lnum, tokens, states, recDepth) {
printdbg2(String.fromCharCode(0xA2), lnum, tokens, states, recDepth); bF.parserPrintdbg2(String.fromCharCode(0xA2), lnum, tokens, states, recDepth);
if (!Array.isArray(tokens) && !Array.isArray(states)) throw new ParserError("Tokens and states are not array"); if (!Array.isArray(tokens) && !Array.isArray(states)) throw new ParserError("Tokens and states are not array");
if (tokens.length != 1) throw new ParserError("parseLit 1"); if (tokens.length != 1) throw new ParserError("parseLit 1");
@@ -632,30 +679,36 @@ let lnum = 10;
// FIXME print's last (;) gets parsed but ignored // FIXME print's last (;) gets parsed but ignored
// if s<2 then (nop1) else (if s < 9999 then nop2 else nop3) // if s<2 then (nop1) else (if s < 9999 then nop2 else nop3)
//let tokens = ["if","s","<","2","then","(","nop1",")","else","(","if","s","<","9999","then","nop2","else","nop3",")"]; let tokens1 = ["if","s","<","2","then","(","nop1",")","else","(","if","s","<","9999","then","nop2","else","nop3",")"];
//let states = ["lit","lit","op","num","lit","paren","lit","paren","lit","paren","lit","lit","op","num","lit","lit","lit","lit","paren"]; let states1 = ["lit","lit","op","num","lit","paren","lit","paren","lit","paren","lit","lit","op","num","lit","lit","lit","lit","paren"];
// DEFUN HYPOT(X,Y) = SQR(X*X+Y*Y) // DEFUN HYPOT(X,Y) = SQR(X*X+Y*Y)
//let tokens = ["defun","HYPOT","(","X",",","Y",")","=","SQR","(","X","*","X","+","Y","*","Y",")"]; let tokens2 = ["defun","HYPOT","(","X",",","Y",")","=","SQR","(","X","*","X","+","Y","*","Y",")"];
//let states = ["lit","lit","paren","lit","sep","lit","paren","op","lit","paren","lit","op","lit","op","lit","op","lit","paren"]; let states2 = ["lit","lit","paren","lit","sep","lit","paren","op","lit","paren","lit","op","lit","op","lit","op","lit","paren"];
// DEFUN SINC(X) = SIN(X) / X // DEFUN SINC(X) = SIN(X) / X
//let tokens = ["DEFUN","SINC","(","X",")","=","SIN","(","X",")","/","X"]; let tokens3 = ["DEFUN","SINC","(","X",")","=","SIN","(","X",")","/","X"];
//let states = ["lit","lit","paren","lit","paren","op","lit","paren","lit","paren","op","lit"]; let states3 = ["lit","lit","paren","lit","paren","op","lit","paren","lit","paren","op","lit"];
// PRINT(IF S<2 THEN "111" ELSE IF S<3 THEN "222" ELSE "333") // PRINT(IF S<2 THEN "111" ELSE IF S<3 THEN "222" ELSE "333")
//let tokens = ["PRINT","(","IF","S","<","2","THEN","111","ELSE","IF","S","<","3","THEN","222","ELSE","333",")"]; let tokens4 = ["PRINT","(","IF","S","<","2","THEN","111","ELSE","IF","S","<","3","THEN","222","ELSE","333",")"];
//let states = ["lit","paren","lit","lit","op","lit","lit","qot","lit","lit","lit","op","lit","lit","qot","lit","qot","paren"]; let states4 = ["lit","paren","lit","lit","op","lit","lit","qot","lit","lit","lit","op","lit","lit","qot","lit","qot","paren"];
// ON 6*SQR(X-3) GOTO X+1, X+2, X+3 // ON 6*SQR(X-3) GOTO X+1, X+2, X+3
let tokens = ["ON","6","*","SQR","(","X","-","3",")","GOTO","X","+","1",",","X","+","2",",","X","+","3"]; let tokens5 = ["ON","6","*","SQR","(","X","-","3",")","GOTO","X","+","1",",","X","+","2",",","X","+","3"];
let states = ["lit","num","op","lit","paren","lit","op","num","paren","lit","lit","op","num","sep","lit","op","num","sep","lit","op","num"]; let states5 = ["lit","num","op","lit","paren","lit","op","num","paren","lit","lit","op","num","sep","lit","op","num","sep","lit","op","num"];
let _debugSyntaxAnalysis = false; let _debugSyntaxAnalysis = false;
try { try {
let tree = bF._parseStmt(lnum, tokens, states, 0); let trees = bF._parseTokens(lnum,
serial.println(astToString(tree)); tokens5.concat([":"], tokens4),
states5.concat(["seq"], states4)
);
trees.forEach((t,i) => {
serial.println("\nParsed Statement #"+(i+1));
serial.println(astToString(t));
});
} }
catch (e) { catch (e) {
serial.printerr(e); serial.printerr(e);