compiler stuff

This commit is contained in:
sfja 2025-09-10 02:04:13 +02:00
parent f3b064394e
commit 8b38e5e883
4 changed files with 352 additions and 100 deletions

View File

@ -1,7 +1,93 @@
(let text (call read_text_file "program.phi"))
(fn a (b c) (do
(let identChars "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890+-*/%&|=?!<>'_")
(fn tokenize (text) (do
(let text_len (call len text))
(let tokens ())
(let i 0)
(let line 1)
(loop (do
(if (>= i text_len) (break))
(let ch (call at text i))
(if (call contains " \t\r\n" ch) (do
(call println "line = %, ch = '%'" line ch)
(if (== ch "\n") (do
(+= line 1)
))
(+= i 1)
) (if (call slice_eq text i "//") (do
(loop (do
if (or (>= i text_len) (== (call at text i) "\n") (do
(break)
))
(+= i 1)
))
) (if (call contains "()" ch) (do
(call push tokens (ch line))
(+= i 1)
) (if (== ch "\"") (do
(+= i 1)
) (if (call contains identChars ch) (do
(let value "")
(loop (do
(= ch (call at text i))
(if (or (>= i text_len) (not (call contains identChars ch))) (do
(break)
))
(call push value ch)
(+= i 1)
))
(call push tokens ("ident" line value))
) (do
(call println "illegal char '%'" ch)
(+= i 1)
))))))
))
(return tokens)
))
(fn contains (text ch) (do
(let text_len (call len text))
(let i 0)
(loop (do
(if (>= i text_len) (break))
(if (== (call at text i) ch) (do
(return true)
))
(+= i 1)
))
(return false)
))
(fn slice_eq (str slice_idx substr) (do
(let str_len (call len str))
(let substr_len (call len substr))
(let i slice_idx)
(loop (do
(if (or (>= (+ slice_idx i) str_len) (>= i substr_len))
(return false))
(if (!= (call at str (+ slice_idx i)) (call at substr i))
(return false))
(+= i 1)
))
(return true)
))
(let text (call read_text_file "program.phi"))
(let tokens (call tokenize text))
(call println "=== text ===")
(call println text)
(call println "=== tokens ===")
(call println tokens)
(call println (+ 1 2))

View File

@ -1,4 +1,5 @@
{
"name": "phi-lang",
"version": "1.0.0",
"type": "module"
}

343
phi.js
View File

@ -1,6 +1,7 @@
"use strict";
import * as fs from "fs";
import * as fs from "node:fs";
import process from "node:process";
function main() {
const text = fs.readFileSync(process.argv[2]).toString();
@ -28,7 +29,7 @@ class Evaluator {
}
/**
* @param {Expr} expr
* @param {Expr} expr
*/
eval(expr) {
if (expr.type === "list") {
@ -36,22 +37,22 @@ class Evaluator {
} else if (expr.type === "int") {
return { type: "value", value: { type: "int", value: expr.value } };
} else if (expr.type === "string") {
return { type: "value", value: { type: "string", value: expr.value } };
return {
type: "value",
value: { type: "string", value: expr.value },
};
} else if (expr.type === "ident") {
const findInTree = (syms, ident) => {
if (syms.map.has(ident))
return syms.map.get(ident);
else if (syms.parent)
return findInTree(syms.parent, ident);
else
return undefined;
const sym = this.findSym(expr.value);
if (!sym) {
throw new Error(
`could not find symbol '${expr.value}' on line ${expr.line}`,
);
}
const sym = findInTree(this.syms, expr.value);
if (!sym)
throw new Error(`could not find symbol '${expr.value}' on line ${expr.line}`);
return { type: "value", value: sym };
} else {
throw new Error(`unknown expr type '${expr.type}' on line ${expr.line}`);
throw new Error(
`unknown expr type '${expr.type}' on line ${expr.line}`,
);
}
}
@ -65,19 +66,67 @@ class Evaluator {
evalList(expr) {
const s = expr.values;
const id = s[0]?.value ?? undefined;
const id = s[0]?.type === "ident" ? s[0].value : undefined;
if (id === "fn") {
const name = s[1].value;
this.syms.map.set(name, {
type: "fn",
name,
params: s[2].values.map(ident => ident.value),
params: s[2].values.map((ident) => ident.value),
body: s[3],
syms: this.syms,
});
return { type: "value", value: { type: "null" } };
} else if (id === "return") {
return {
type: "return",
value: s[1] ? this.evalToValue(s[1]) : { type: "null" },
};
} else if (id === "let") {
const value = this.evalToValue(s[2]);
this.syms.map.set(s[1].value, value);
return { type: "value", value: { type: "null" } };
} else if (id === "if") {
const cond = this.evalToValue(s[1]);
if (cond.type !== "bool") {
throw new Error(
`expected bool on line ${expr.line}`,
);
}
if (cond.value) {
return this.eval(s[2]);
} else if (s[3]) {
return this.eval(s[3]);
} else {
return { type: "value", value: "null" };
}
} else if (id === "loop") {
while (true) {
const result = this.eval(s[1]);
if (result.type === "break") {
return { type: "value", value: result.value };
} else if (result.type !== "value") {
return result;
}
}
} else if (id === "break") {
return {
type: "break",
value: s[1] ? this.evalToValue(s[1]) : { type: "null" },
};
} else if (id === "do") {
let lastValue = { type: "null" };
for (const expr of s.slice(1)) {
const result = this.eval(expr);
if (result.type !== "value") {
return result;
}
lastValue = result.value;
}
return { type: "value", value: lastValue };
} else if (id === "call") {
const args = s.slice(2).map(arg => this.evalToValue(arg));
const args = s.slice(2).map((arg) => this.evalToValue(arg));
const fnValue = this.evalToValue(s[1]);
if (fnValue.type === "builtin") {
@ -91,7 +140,9 @@ class Evaluator {
map: new Map(),
};
if (fnValue.params.length !== args.length) {
throw new Error(`incorrect amount of arguments on line ${line}`);
throw new Error(
`incorrect amount of arguments on line ${line}`,
);
}
for (let i = 0; i < fnValue.params.length; ++i) {
this.syms.map.set(fnValue.params[i], args[i]);
@ -103,70 +154,159 @@ class Evaluator {
if (result.type === "value" || result.type === "return") {
returnValue = result.value;
} else {
throw new Error(`illegal ${result.type} across boundry`)
throw new Error(`illegal ${result.type} across boundry`);
}
this.syms = callerSyms;
return { type: "value", value: returnValue };
} else if (id === "return") {
return { type: "return", value: s[1] ? this.evalToValue(s[1]) : { type: "null" } };
} else if (id === "let") {
const value = this.evalToValue(s[2]);
this.syms.map.set(s[1].value, value);
return { type: "value", value: { type: "null" } };
} else if (id === "do") {
let lastValue = { type: "null" };
for (const expr of s.slice(1)) {
const result = this.eval(expr);
if (result.type !== "value") {
break;
}
lastValue = result.value;
}
return { type: "value", value: lastValue };
} else if (s[0] === "if") {
const cond = this.evalToValue(s[1]);
if (cond.type !== "bool") {
throw new Error(`expected bool on line ${line}`);
}
if (cond.value) {
return this.eval(s[2]);
} else if (s[3]) {
return this.eval(s[3]);
} else if (id === "not") {
const value = this.evalToValue(s[1]);
return {
type: "value",
value: { type: "bool", value: !value.value },
};
} else if (id === "or") {
const left = this.evalToValue(s[1]);
if (left.value) {
return { type: "value", value: left };
} else {
return { type: "value", value: "null" };
const right = this.evalToValue(s[2]);
return { type: "value", value: right };
}
} else if (s[0] === "loop") {
while (true) {
const result = this.eval(s[1]);
if (result.type === "break") {
return { type: "value", value: result.value };
} else if (result.type !== "value") {
return result;
} else if (id === "and") {
const left = this.evalToValue(s[1]);
if (left.value) {
const right = this.evalToValue(s[2]);
return { type: "value", value: right };
} else {
return { type: "value", value: left };
}
} else if (id in artithmeticOps) {
const left = this.evalToValue(s[1]);
const right = this.evalToValue(s[2]);
return {
type: "value",
value: {
type: "int",
value: artithmeticOps[id](left.value, right.value),
},
};
} else if (id in comparisonOps) {
const left = this.evalToValue(s[1]);
const right = this.evalToValue(s[2]);
return {
type: "value",
value: {
type: "bool",
value: comparisonOps[id](left.value, right.value),
},
};
} else if (id in assignOps) {
if (s[1].type === "ident") {
const sym = this.findSym(s[1].value);
if (!sym) {
throw new Error(
`could not find symbol '${expr.value}' on line ${expr.line}`,
);
}
const right = this.evalToValue(s[2]);
const newValue = assignOps[id](sym, right);
sym.type = newValue.type;
sym.value = newValue.value;
} else {
throw new Error(
`cannot assign to expression on line ${expr.line}`,
);
}
} else if (s[0] === "break") {
return { type: "break", value: s[1] ? this.evalToValue(s[1]) : { type: "null" } };
return { type: "value", value: { type: "null" } };
} else {
return { type: "value", value: { type: "list", values: s.map(expr => this.evalToValue(expr)) } };
return {
type: "value",
value: {
type: "list",
values: s.map((expr) => this.evalToValue(expr)),
},
};
}
}
findSym(ident, syms = this.syms) {
if (syms.map.has(ident)) {
return syms.map.get(ident);
} else if (syms.parent) {
return this.findSym(ident, syms.parent);
} else {
return undefined;
}
}
}
const artithmeticOps = {
"+": (left, right) => right + left,
"-": (left, right) => right - left,
};
const comparisonOps = {
"==": (left, right) => left === right,
"!=": (left, right) => left !== right,
"<": (left, right) => left < right,
">": (left, right) => left > right,
"<=": (left, right) => left <= right,
">=": (left, right) => left >= right,
};
const assignOps = {
"=": (_, right) => right,
"+=": (left, right) => ({ type: "int", value: left.value + right.value }),
"-=": (left, right) => ({ type: "int", value: left.value - right.value }),
};
const builtinFns = {
println(msg) {
console.log(valueToPrint(msg));
println(msg, ...args) {
let text = valueToPrint(msg);
for (const arg of args) {
text = text.replace("%", valueToPrint(arg));
}
console.log(text);
return { type: "null" };
},
read_text_file(path) {
const text = fs.readFileSync(path.value);
const text = fs.readFileSync(path.value).toString();
return { type: "string", value: text };
}
},
push(list, value) {
if (list.type === "string") {
list.value += value.value;
return list;
}
list.values.push(value);
return list;
},
at(value, index) {
if (value.type === "string") {
return { type: "string", value: value.value[index.value] };
}
return value.values[index.value];
},
len(value) {
if (value.type === "string") {
return { type: "int", value: value.value.length };
}
return { type: "int", value: value.values.length };
},
};
const builtins = Object.entries(builtinFns)
.map(([key, fn]) => [key, { type: "builtin", fn }]);
const consts = {
"null": { type: "null" },
"false": { type: "bool", value: false },
"true": { type: "bool", value: true },
};
const builtins = [
...Object.entries(builtinFns)
.map(([key, fn]) => [key, { type: "builtin", fn }]),
...Object.entries(consts),
];
function valueToPrint(value) {
if (value.type === "null") {
@ -178,7 +318,7 @@ function valueToPrint(value) {
} else if (value.type === "string") {
return `${value.value}`;
} else if (value.type === "list") {
return `(${value.values.map(v => valueToString(v)).join(" ")})`;
return `(${value.values.map((v) => valueToString(v)).join(" ")})`;
} else {
throw new Error(`unknown value type ${value.type}`);
}
@ -194,7 +334,7 @@ function valueToString(value) {
} else if (value.type === "string") {
return `"${value.value}"`;
} else if (value.type === "list") {
return `(${value.values.map(v => valueToString(v)).join(" ")})`;
return `(${value.values.map((v) => valueToString(v)).join(" ")})`;
} else {
throw new Error(`unknown value type ${value.type}`);
}
@ -210,14 +350,14 @@ function valueToJs(value) {
} else if (value.type === "string") {
return value.value;
} else if (value.type === "list") {
return value.values.map(v => valueToJs(v));
return value.values.map((v) => valueToJs(v));
} else {
throw new Error(`unknown value type ${value.type}`);
}
}
/**
* @param {Expr} expr
* @param {Expr} expr
* @returns {string}
*/
function exprToString(expr) {
@ -228,7 +368,7 @@ function exprToString(expr) {
} else if (expr.type === "string") {
return `"${expr.value}"`;
} else if (expr.type === "list") {
return `(${expr.values.map(v => exprToString(v)).join(" ")})`;
return `(${expr.values.map((v) => exprToString(v)).join(" ")})`;
} else {
throw new Error(`unknown value type ${expr.type}`);
}
@ -248,18 +388,18 @@ class Parser {
.replace(/\/\/.*?$/mg, "")
.replace(/([\(\)\n])/g, " $1 ")
.split(/[ \t\r]/)
.filter(tok => tok !== "");
.filter((tok) => tok !== "");
this.idx = 0;
this.line = 1;
}
/**
*
* @returns {Expr[]}
*/
parse() {
if (this.curr === "\n")
if (this.curr === "\n") {
this.step();
}
const exprs = [];
while (!this.done) {
@ -276,62 +416,66 @@ class Parser {
values.push(this.parseExpr());
}
if (!this.test(")")) {
throw new Error(`expected ')'`)
throw new Error(`expected ')'`);
}
this.step();
return { type: "list", line, values };
}
else if (this.test(/STRING_\d+/)) {
} else if (this.test(/STRING_\d+/)) {
const id = Number(this.curr.match(/STRING_(\d+)/)[1]);
this.step();
return { type: "string", line, value: this.strings[id] };
}
else if (this.test(/0|(:?[1-9][0-9]*)/)) {
} else if (this.test(/0|(:?[1-9][0-9]*)/)) {
const value = Number(this.curr);
this.step();
return { type: "int", line, value };
}
else if (this.test(/[a-zA-Z0-9\+\-\*/%&\|=\?\!<>'_]+/)) {
} else if (this.test(/[a-zA-Z0-9\+\-\*/%&\|=\?\!<>'_]+/)) {
const value = this.curr;
this.step();
return { type: "ident", line, value };
}
else {
throw new Error(`expected expression, got ${this.curr}`)
} else {
throw new Error(
`expected expression, got ${this.curr} on line ${this.line}`,
);
}
}
eat(tok) {
if (!this.test(tok))
if (!this.test(tok)) {
return false;
this.step()
}
this.step();
return true;
}
test(tok) {
if (this.done)
if (this.done) {
return false;
if (typeof tok === "string")
}
if (typeof tok === "string") {
return this.curr === tok;
else if (tok instanceof RegExp)
} else if (tok instanceof RegExp) {
return new RegExp(`^${tok.source}$`)
.test(this.curr);
else
throw new Error()
} else {
throw new Error();
}
}
step() {
do {
this.idx += 1;
if (!this.done && this.curr === "\n") {
this.line += 1;
}
this.idx += 1;
} while (!this.done && this.curr === "\n");
}
get done() { return this.idx >= this.tokens.length; }
get curr() { return this.tokens[this.idx]; }
get done() {
return this.idx >= this.tokens.length;
}
get curr() {
return this.tokens[this.idx];
}
}
class StringExtractor {
constructor(text) {
this.text = text;
@ -342,8 +486,8 @@ class StringExtractor {
extract() {
while (this.idx < this.text.length) {
if (this.text[this.idx] == '\"') {
this.extractString()
if (this.text[this.idx] == '"') {
this.extractString();
} else {
this.outputText += this.text[this.idx];
this.idx += 1;
@ -355,10 +499,11 @@ class StringExtractor {
this.idx += 1;
let value = "";
while (this.idx < this.text.length && this.text[this.idx] != '"') {
if (this.text[this.idx] == '\\') {
if (this.text[this.idx] == "\\") {
this.idx += 1;
if (this.idx > this.text.length)
if (this.idx > this.text.length) {
break;
}
const ch = this.text[this.idx];
value += {
"0": "\0",
@ -380,8 +525,12 @@ class StringExtractor {
this.outputText += `STRING_${id}`;
}
getStrings() { return this.strings; }
getOutputText() { return this.outputText; }
getStrings() {
return this.strings;
}
getOutputText() {
return this.outputText;
}
}
main();

View File

@ -9,9 +9,25 @@ endif
syn keyword Keyword fn call return loop break if let do
syn keyword Operator and or not
syn keyword Special null
syn keyword Boolean true false
syn match Operator '+'
syn match Operator '-'
syn match Operator '\*'
syn match Operator '/'
syn match Operator '='
syn match Operator '+='
syn match Operator '-='
syn match Operator '=='
syn match Operator '!='
syn match Operator '<'
syn match Operator '>'
syn match Operator '<='
syn match Operator '>='
syn match Number '0'
syn match Number '[1-9][0-9]*'