6502/8bitworkshop
1
0
Fork 0
mirror of https://github.com/sehugg/8bitworkshop.git synced 2026-04-25 18:47:56 +00:00
Code Issues Projects Releases Wiki Activity

basic: fuzz test fixes, DEF cycle detector

Browse Source
This commit is contained in:
Steven Hugg
2020-08-15 15:03:56 -05:00
parent 295f1ef9de
commit 88fa924507
6 changed files with 1107 additions and 87 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/common/basic/compiler.ts
+101 -32
View File
@@ -6,8 +6,8 @@ export interface BASICOptions {
asciiOnly : boolean; // reject non-ASCII chars?
uppercaseOnly : boolean; // convert everything to uppercase?
optionalLabels : boolean; // can omit line numbers and use labels?
optionalWhitespace : boolean; // can "crunch" keywords?
varNaming : 'A'|'A1'|'AA'|'*'; // only allow A0-9 for numerics, single letter for arrays/strings
optionalWhitespace : boolean; // can "crunch" keywords? also, eat extra ":" delims
varNaming : 'A'|'A1'|'AA'|'*'; // only allow A0-9 for numerics, single letter for arrays/strings
squareBrackets : boolean; // "[" and "]" interchangable with "(" and ")"?
tickComments : boolean; // support 'comments?
hexOctalConsts : boolean; // support &H and &O integer constants?
@@ -43,6 +43,8 @@ export interface BASICOptions {
endStmtRequired : boolean; // need END at end?
// MISC
commandsPerSec? : number; // how many commands per second?
maxLinesPerFile? : number; // limit on # of lines
maxArrayElements? : number; // max array elements (all dimensions)
}
export interface SourceLocated {
@@ -59,8 +61,8 @@ export class CompileError extends Error {
}
// Lexer regular expression -- each (capture group) handles a different token type
// FLOAT INT HEXOCTAL REMARK IDENT STRING RELOP EXP OPERATORS OTHER WS
const re_toks = /([0-9.]+[E][+-]?\d+|\d+[.][E0-9]*|[.][E0-9]+)|[0]*(\d+)|&([OH][0-9A-F]+)|(['].*)|(\w+[$]?)|(".*?")|([<>]?[=<>#])|(\*\*)|([-+*/^,;:()\[\]\?\\])|(\S+)|(\s+)/gi;
// FLOAT INT HEXOCTAL REMARK IDENT STRING RELOP EXP OPERATORS OTHER WS
const re_toks = /([0-9.]+[E][+-]?\d+|\d+[.][E0-9]*|[.][E0-9]+)|[0]*(\d+)|&([OH][0-9A-F]+)|(['].*)|([A-Z_]\w*[$]?)|(".*?")|([<>]?[=<>#])|(\*\*)|([-+*/^,;:()\[\]\?\\])|(\S+)|(\s+)/gi;
export enum TokenType {
EOL = 0,
@@ -302,6 +304,19 @@ function stripQuotes(s: string) {
return s.substr(1, s.length-2);
}
function isLiteral(arg: Expr): arg is Literal {
return (arg as any).value != null;
}
function isLookup(arg: Expr): arg is IndOp {
return (arg as any).name != null;
}
function isBinOp(arg: Expr): arg is BinOp {
return (arg as any).op != null && (arg as any).left != null && (arg as any).right != null;
}
function isUnOp(arg: Expr): arg is UnOp {
return (arg as any).op != null && (arg as any).expr != null;
}
///// BASIC PARSER
export class BASICParser {
@@ -311,7 +326,9 @@ export class BASICParser {
listings: CodeListingMap;
labels: { [label: string]: BASICLine };
targets: { [targetlabel: string]: SourceLocation };
refs: { [name: string]: SourceLocation }; // references
varrefs: { [name: string]: SourceLocation }; // references
fnrefs: { [name: string]: string[] }; // DEF FN call graph
maxlinelen : number = 255; // maximum line length
path : string;
lineno : number;
@@ -327,7 +344,8 @@ export class BASICParser {
this.lineno = 0;
this.curlabel = null;
this.listings = {};
this.refs = {};
this.varrefs = {};
this.fnrefs = {};
this.optionCount = 0;
}
addError(msg: string, loc?: SourceLocation) {
@@ -382,10 +400,7 @@ export class BASICParser {
}
} else this.dialectError(`optional line numbers`);
case TokenType.Int:
if (this.labels[tok.str] != null) this.compileError(`There's a duplicated label "${tok.str}".`);
this.labels[tok.str] = line;
line.label = tok.str;
this.curlabel = tok.str;
this.setCurrentLabel(line, tok.str);
break;
case TokenType.HexOctalInt:
case TokenType.Float:
@@ -398,9 +413,16 @@ export class BASICParser {
break;
}
}
setCurrentLabel(line: BASICLine, str: string) {
if (this.labels[str] != null) this.compileError(`There's a duplicated label "${str}".`);
this.labels[str] = line;
line.label = str;
this.curlabel = str;
this.tokens.forEach((tok) => tok.$loc.label = str);
}
parseFile(file: string, path: string) : BASICProgram {
this.path = path;
var txtlines = file.split(/\n|\r\n/);
var txtlines = file.split(/\n|\r\n?/);
var pgmlines = txtlines.map((line) => this.parseLine(line));
var program = { opts: this.opts, lines: pgmlines };
this.checkAll(program);
@@ -416,9 +438,7 @@ export class BASICParser {
return {label:null, stmts:[]};
}
}
tokenize(line: string) : void {
this.lineno++;
this.tokens = [];
_tokenize(line: string) : void {
// split identifier regex (if token-crunching enabled)
let splitre = this.opts.optionalWhitespace && new RegExp('('+this.opts.validKeywords.map(s => `${s}`).join('|')+')');
// iterate over each token via re_toks regex
@@ -428,7 +448,7 @@ export class BASICParser {
for (var i = 1; i <= lastTokType; i++) {
let s : string = m[i];
if (s != null) {
let loc = { path: this.path, line: this.lineno, start: m.index, end: m.index+s.length, label: this.curlabel };
let loc = { path: this.path, line: this.lineno, start: m.index, end: m.index+s.length };
// maybe we don't support unicode in 1975?
if (this.opts.asciiOnly && !/^[\x00-\x7F]*$/.test(s))
this.dialectError(`non-ASCII characters`);
@@ -437,6 +457,9 @@ export class BASICParser {
s = s.toUpperCase();
// DATA statement captures whitespace too
if (s == 'DATA') lastTokType = TokenType.Whitespace;
// certain keywords shouldn't split for rest of line
if (s == 'DATA') splitre = null;
if (s == 'OPTION') splitre = null;
// REM means ignore rest of statement
if (lastTokType == TokenType.CatchAll && s.startsWith('REM')) {
s = 'REM';
@@ -451,26 +474,31 @@ export class BASICParser {
}
// un-crunch tokens?
if (splitre && i == TokenType.Ident) {
var splittoks = s.split(splitre);
splittoks.forEach((ss) => {
if (ss != '') {
// leftover might be integer
i = /^[0-9]+$/.test(ss) ? TokenType.Int : TokenType.Ident;
// disable crunching after this token?
if (ss == 'DATA' || ss == 'OPTION')
splitre = null;
var splittoks = s.split(splitre).filter((s) => s != ''); // only non-empties
if (splittoks.length > 1) {
splittoks.forEach((ss) => {
// check to see if leftover might be integer, or identifier
if (/^[0-9]+$/.test(ss)) i = TokenType.Int;
else if (/^[A-Z_]\w*[$]?$/.test(ss)) i = TokenType.Ident;
else this.compileError(`Try adding whitespace before "${ss}".`);
this.tokens.push({str: ss, type: i, $loc:loc});
}
});
} else {
// add token to list
this.tokens.push({str: s, type: i, $loc:loc});
});
s = null;
}
}
// add token to list
if (s) this.tokens.push({str: s, type: i, $loc:loc});
break;
}
}
}
this.eol = { type: TokenType.EOL, str: "", $loc: { path: this.path, line: this.lineno, start: line.length, label: this.curlabel } };
}
tokenize(line: string) : void {
this.lineno++;
this.tokens = []; // can't have errors until this is set
this.eol = { type: TokenType.EOL, str: "", $loc: { path: this.path, line: this.lineno, start: line.length } };
if (line.length > this.maxlinelen) this.compileError(`A line should be no more than ${this.maxlinelen} characters long.`);
this._tokenize(line);
}
parse() : BASICLine {
var line = {label: null, stmts: []};
@@ -502,6 +530,9 @@ export class BASICParser {
return false;
}
parseStatement(): Statement | null {
// eat extra ":" (should have separate property for this)
if (this.opts.optionalWhitespace && this.peekToken().str == ':') return null;
// get the command word
var cmdtok = this.consumeToken();
var cmd = cmdtok.str;
var stmt;
@@ -510,6 +541,7 @@ export class BASICParser {
if (cmdtok.str.startsWith("'") && !this.opts.tickComments) this.dialectError(`tick remarks`);
return null;
case TokenType.Operator:
// "?" is alias for "PRINT" on some platforms
if (cmd == this.validKeyword('?')) cmd = 'PRINT';
case TokenType.Ident:
// ignore remarks
@@ -534,8 +566,11 @@ export class BASICParser {
this.pushbackToken(cmdtok);
stmt = this.stmt__LET();
break;
} else {
this.compileError(`I don't understand the command "${cmd}".`);
}
case TokenType.EOL:
if (this.opts.optionalWhitespace) return null;
default:
this.compileError(`There should be a command here.`);
return null;
@@ -547,7 +582,7 @@ export class BASICParser {
var tok = this.consumeToken();
switch (tok.type) {
case TokenType.Ident:
this.refs[tok.str] = tok.$loc;
this.varrefs[tok.str] = tok.$loc;
let args = null;
if (this.peekToken().str == '(') {
this.expectToken('(');
@@ -745,6 +780,20 @@ export class BASICParser {
break;
}
}
visitExpr(expr: Expr, callback: (expr:Expr) => void) {
if (isBinOp(expr)) {
this.visitExpr(expr.left, callback);
this.visitExpr(expr.right, callback);
}
if (isUnOp(expr)) {
this.visitExpr(expr.expr, callback);
}
if (isLookup(expr) && expr.args != null) {
for (var arg of expr.args)
this.visitExpr(arg, callback);
}
callback(expr);
}
//// STATEMENTS
@@ -865,6 +914,10 @@ export class BASICParser {
this.compileError(`An array defined by DIM must have at least one dimension.`)
else if (arr.args.length > this.opts.maxDimensions)
this.dialectError(`more than ${this.opts.maxDimensions} dimensional arrays`);
for (var arrdim of arr.args) {
if (isLiteral(arrdim) && arrdim.value < this.opts.defaultArrayBase)
this.compileError(`An array dimension cannot be less than ${this.opts.defaultArrayBase}.`);
}
});
return { command:'DIM', args:lexprs };
}
@@ -925,8 +978,26 @@ export class BASICParser {
if (!lexpr.name.startsWith('FN')) this.compileError(`Functions defined with DEF must begin with the letters "FN".`)
this.expectToken("=");
var func = this.parseExpr();
// build call graph to detect cycles
this.visitExpr(func, (expr:Expr) => {
if (isLookup(expr) && expr.name.startsWith('FN')) {
if (!this.fnrefs[lexpr.name])
this.fnrefs[lexpr.name] = [];
this.fnrefs[lexpr.name].push(expr.name);
}
});
this.checkCallGraph(lexpr.name, new Set());
return { command:'DEF', lexpr:lexpr, def:func };
}
// detect cycles in call graph starting at function 'name'
checkCallGraph(name: string, visited: Set<string>) {
if (visited.has(name)) this.compileError(`There was a cycle in the function definition graph for ${name}.`);
visited.add(name);
var refs = this.fnrefs[name] || [];
for (var ref of refs)
this.checkCallGraph(ref, visited); // recurse
visited.delete(name);
}
stmt__POP() : NoArgStatement {
return { command:'POP' };
}
@@ -1550,9 +1621,7 @@ export const MODERN_BASIC : BASICOptions = {
// TODO: integer vars
// TODO: DEFINT/DEFSTR
// TODO: superfluous ":" ignored on MS basics only?
// TODO: excess INPUT ignored, error msg
// TODO: max line len?
export const DIALECTS = {
"DEFAULT": ALTAIR_BASIC41,