porffor 0.2.0-ef043de → 0.2.0-f7ea197

package/README.md

@@ -23,13 +23,13 @@ Expect nothing to work! Only very limited JS is currently supported. See files i
 **`porf path/to/script.js`**
 
 ### Compiling to Wasm
-**`porf compile path/to/script.js out.wasm`**. Currently it does not use an import standard like WASI, so it is mostly unusable on its own.
+**`porf wasm path/to/script.js out.wasm`**. Currently it does not use an import standard like WASI, so it is mostly unusable on its own.
 
 ### Compiling to native binaries
 > [!WARNING]
 > Compiling to native binaries uses [2c](#2c), Porffor's own Wasm -> C compiler, which is experimental.
 
-**`porf native path/to/script.js out(.exe)`**. You can specify the compiler with `-compiler=clang/zig/gcc`, and which opt level to use with `-cO=O3` (`Ofast` by default). Output binaries are also stripped by default.
+**`porf native path/to/script.js out(.exe)`**. You can specify the compiler with `--compiler=clang/zig/gcc`, and which opt level to use with `--cO=O3` (`Ofast` by default). Output binaries are also stripped by default.
 
 ### Compiling to C
 > [!WARNING]
@@ -57,31 +57,25 @@ Expect nothing to work! Only very limited JS is currently supported. See files i
 
 
 ### Options
-- `-target=wasm|c|native` (default: `wasm`) to set target output (native compiles c output to binary, see args below)
-- `-target=c|native` only:
-  - `-o=out.c|out.exe|out` to set file to output c or binary
-- `-target=native` only:
-  - `-compiler=clang` to set compiler binary (path/name) to use to compile
-  - `-cO=O3` to set compiler opt argument
-- `-parser=acorn|@babel/parser|meriyah|hermes-parser` (default: `acorn`) to set which parser to use
-- `-parse-types` to enable parsing type annotations/typescript. if `-parser` is unset, changes default to `@babel/parser`. does not type check
-- `-opt-types` to perform optimizations using type annotations as compiler hints. does not type check
-- `-valtype=i32|i64|f64` (default: `f64`) to set valtype
+- `--parser=acorn|@babel/parser|meriyah|hermes-parser` (default: `acorn`) to set which parser to use
+- `--parse-types` to enable parsing type annotations/typescript. if `-parser` is unset, changes default to `@babel/parser`. does not type check
+- `--opt-types` to perform optimizations using type annotations as compiler hints. does not type check
+- `--valtype=i32|i64|f64` (default: `f64`) to set valtype
 - `-O0` to disable opt
 - `-O1` (default) to enable basic opt (simplify insts, treeshake wasm imports)
 - `-O2` to enable advanced opt (inlining). unstable
 - `-O3` to enable advanceder opt (precompute const math). unstable
-- `-no-run` to not run wasm output, just compile
-- `-opt-log` to log some opts
-- `-code-log` to log some codegen (you probably want `-funcs`)
-- `-regex-log` to log some regex
-- `-funcs` to log funcs
-- `-ast-log` to log AST
-- `-opt-funcs` to log funcs after opt
-- `-sections` to log sections as hex
-- `-opt-no-inline` to not inline any funcs
-- `-tail-call` to enable tail calls (experimental + not widely implemented)
-- `-compile-hints` to enable V8 compilation hints (experimental + doesn't seem to do much?)
+- `--no-run` to not run wasm output, just compile
+- `--opt-log` to log some opts
+- `--code-log` to log some codegen (you probably want `-funcs`)
+- `--regex-log` to log some regex
+- `--funcs` to log funcs
+- `--ast-log` to log AST
+- `--opt-funcs` to log funcs after opt
+- `--sections` to log sections as hex
+- `--opt-no-inline` to not inline any funcs
+- `--tail-call` to enable tail calls (experimental + not widely implemented)
+- `--compile-hints` to enable V8 compilation hints (experimental + doesn't seem to do much?)
 
 ### Running in the repo
 The repo comes with easy alias files for Unix and Windows, which you can use like so:
@@ -211,7 +205,7 @@ Mostly for reducing size. I do not really care about compiler perf/time as long
 ### Traditional opts
 - Inlining functions (WIP, limited)
 - Inline const math ops
-- Tail calls (behind flag `-tail-call`)
+- Tail calls (behind flag `--tail-call`)
 
 ### Wasm transforms
 - `local.set`, `local.get` -> `local.tee`

package/asur/index.js

@@ -1244,7 +1244,7 @@ paused = _paused;`);
 });
 
 export const instantiate = async (binary, importImpls) => {
-  const _vm = process?.argv?.includes('-wasm-debug') ? await wasmDebugVm() : vm;
+  const _vm = process?.argv?.includes('--wasm-debug') ? await wasmDebugVm() : vm;
 
   const parsed = parse(binary);
   const exports = {};

package/compiler/assemble.js

@@ -154,7 +154,7 @@ export default (funcs, globals, tags, pages, data, flags) => {
 
   const exports = funcs.filter(x => x.export).map((x, i) => [ ...encodeString(x.name === 'main' ? 'm' : x.name), ExportDesc.func, x.index ]);
 
-  if (Prefs.alwaysMemory && pages.size === 0) pages.set('-always-memory', 0);
+  if (Prefs.alwaysMemory && pages.size === 0) pages.set('--always-memory', 0);
   if (optLevel === 0) pages.set('O0 precaution', 0);
 
   const usesMemory = pages.size > 0;

package/compiler/codegen.js

@@ -1646,18 +1646,25 @@ const generateCall = (scope, decl, _global, _name, unusedValue = false) => {
     // megahack for /regex/.func()
     const funcName = decl.callee.property.name;
     if (decl.callee.object.regex && Object.hasOwn(Rhemyn, funcName)) {
-      const func = Rhemyn[funcName](decl.callee.object.regex.pattern, currentFuncIndex++);
+      const regex = decl.callee.object.regex.pattern;
+      const rhemynName = `regex_${funcName}_${regex}`;
 
-      funcIndex[func.name] = func.index;
-      funcs.push(func);
+      if (!funcIndex[rhemynName]) {
+        const func = Rhemyn[funcName](regex, currentFuncIndex++, rhemynName);
 
+        funcIndex[func.name] = func.index;
+        funcs.push(func);
+      }
+
+      const idx = funcIndex[rhemynName];
       return [
         // make string arg
         ...generate(scope, decl.arguments[0]),
+        Opcodes.i32_to_u,
+        ...getNodeType(scope, decl.arguments[0]),
 
         // call regex func
-        Opcodes.i32_to_u,
-        [ Opcodes.call, func.index ],
+        [ Opcodes.call, idx ],
         Opcodes.i32_from_u,
 
         ...number(TYPES.boolean, Valtype.i32),
@@ -3649,7 +3656,7 @@ export default program => {
 
   globalThis.valtype = 'f64';
 
-  const valtypeOpt = process.argv.find(x => x.startsWith('-valtype='));
+  const valtypeOpt = process.argv.find(x => x.startsWith('--valtype='));
   if (valtypeOpt) valtype = valtypeOpt.split('=')[1];
 
   globalThis.valtypeBinary = Valtype[valtype];
@@ -3657,7 +3664,7 @@ export default program => {
   const valtypeInd = ['i32', 'i64', 'f64'].indexOf(valtype);
 
   globalThis.pageSize = PageSize;
-  const pageSizeOpt = process.argv.find(x => x.startsWith('-page-size='));
+  const pageSizeOpt = process.argv.find(x => x.startsWith('--page-size='));
   if (pageSizeOpt) pageSize = parseInt(pageSizeOpt.split('=')[1]) * 1024;
 
   // set generic opcodes for current valtype

package/compiler/parse.js

@@ -23,7 +23,7 @@ globalThis.typedInput = types && Prefs.optTypes;
 globalThis.parser = '';
 let parse;
 const loadParser = async (fallbackParser = 'acorn', forceParser) => {
-  parser = forceParser ?? process.argv.find(x => x.startsWith('-parser='))?.split('=')?.[1] ?? fallbackParser;
+  parser = forceParser ?? process.argv.find(x => x.startsWith('--parser='))?.split('=')?.[1] ?? fallbackParser;
   0, { parse } = (await import((globalThis.document || globalThis.Deno ? 'https://esm.sh/' : '') + parser));
 };
 globalThis._porf_loadParser = loadParser;

package/compiler/precompile.js

@@ -21,7 +21,7 @@ const compile = async (file, [ _funcs, _globals ]) => {
     first = source.slice(0, source.indexOf('\n'));
   }
 
-  let args = ['-bytestring', '-todo-time=compile', '-no-aot-pointer-opt', '-no-treeshake-wasm-imports', '-scoped-page-names', '-parse-types', '-opt-types'];
+  let args = ['--bytestring', '--todo-time=compile', '--no-aot-pointer-opt', '--no-treeshake-wasm-imports', '--scoped-page-names', '--parse-types', '--opt-types'];
   if (first.startsWith('// @porf')) {
     args = args.concat(first.slice('// @porf '.length).split(' '));
   }

package/compiler/prefs.js

@@ -9,11 +9,12 @@ const obj = new Proxy({}, {
     return cache[p] = (() => {
       // fooBar -> foo-bar
       const name = p[0] === '_' ? p : p.replace(/[A-Z]/g, c => `-${c.toLowerCase()}`);
-      if (process.argv.includes('-' + name)) return true;
-      if (process.argv.includes('-no-' + name)) return false;
+      const prefix = name.length === 1 ? '-' : '--';
+      if (process.argv.includes(prefix + name)) return true;
+      if (process.argv.includes(prefix + 'no-' + name)) return false;
 
-      const valArg = process.argv.find(x => x.startsWith(`-${name}=`));
-      if (valArg) return valArg.slice(name.length + 2);
+      const valArg = process.argv.find(x => x.startsWith(`${prefix}${name}=`));
+      if (valArg) return valArg.slice(name.length + 1 + prefix.length);
 
       if (onByDefault.includes(p)) return true;
       return undefined;

package/fib.js

@@ -1,7 +1,10 @@
-export const fib = n => {
+const fib = n => {
   if (n < 2) {
     return n;
   } else {
     return fib(n - 1) + fib(n - 2);
   }
-};
+};
+
+console.log(fib(7));
+

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "porffor",
   "description": "a basic experimental wip aot optimizing js -> wasm engine/compiler/runtime in js",
-  "version": "0.2.0-ef043de",
+  "version": "0.2.0-f7ea197",
   "author": "CanadaHonk",
   "license": "MIT",
   "scripts": {

package/rhemyn/compile.js

@@ -1,8 +1,8 @@
-import { Blocktype, Opcodes, Valtype, PageSize, ValtypeSize } from '../compiler/wasmSpec.js';
+import { Blocktype, Opcodes, Valtype, ValtypeSize } from '../compiler/wasmSpec.js';
 import { number } from '../compiler/embedding.js';
-import { signedLEB128, unsignedLEB128 } from '../compiler/encoding.js';
 import parse from './parse.js';
 import Prefs from '../compiler/prefs.js';
+import { TYPES } from '../compiler/types.js';
 
 // local indexes
 const BasePointer = 0; // base string pointer
@@ -14,7 +14,7 @@ const Length = 5;
 const Tmp = 6;
 
 let exprLastGet = false;
-const generate = (node, negated = false, get = true, func = 'test') => {
+const generate = (node, negated = false, get = true, stringSize = 2, func = 'test') => {
   let out = [];
   switch (node.type) {
     case 'Expression':
@@ -42,7 +42,7 @@ const generate = (node, negated = false, get = true, func = 'test') => {
         // generate checks
         ...node.body.flatMap((x, i) => {
           exprLastGet = x.type !== 'Group' && i === (node.body.length - 1);
-          return generate(x, negated);
+          return generate(x, negated, true, stringSize, func);
         }),
 
         // reached end without branching out, successful match
@@ -56,9 +56,9 @@ const generate = (node, negated = false, get = true, func = 'test') => {
 
         [ Opcodes.end ],
 
-        // increment iter pointer by sizeof i16
+        // increment iter pointer by string size
         [ Opcodes.local_get, IterPointer ],
-        ...number(ValtypeSize.i16, Valtype.i32),
+        ...number(stringSize, Valtype.i32),
         [ Opcodes.i32_add ],
         [ Opcodes.local_set, IterPointer ],
 
@@ -91,34 +91,34 @@ const generate = (node, negated = false, get = true, func = 'test') => {
       break;
 
     case 'Character':
-      out = generateChar(node, node.negated ^ negated, get);
+      out = generateChar(node, node.negated ^ negated, get, stringSize);
       break;
 
     case 'Set':
-      out = generateSet(node, node.negated, get);
+      out = generateSet(node, node.negated, get, stringSize);
       break;
 
     case 'Group':
-      out = generateGroup(node, negated, get);
+      out = generateGroup(node, negated, get, stringSize);
       break;
 
     case 'Range':
-      out = generateRange(node, negated, get);
+      out = generateRange(node, negated, get, stringSize);
       break;
   }
 
   return out;
 };
 
-const getNextChar = () => [
+const getNextChar = (stringSize) => [
   // get char from pointer
   [ Opcodes.local_get, Pointer ],
-  [ Opcodes.i32_load16_u, Math.log2(ValtypeSize.i16) - 1, ...unsignedLEB128(0) ],
+  [ stringSize == 2 ? Opcodes.i32_load16_u : Opcodes.i32_load8_u, 0, 0 ],
 
   ...(exprLastGet ? [] : [
-    // pointer += sizeof i16
+    // pointer += string size
     [ Opcodes.local_get, Pointer ],
-    ...number(ValtypeSize.i16, Valtype.i32),
+    ...number(stringSize, Valtype.i32),
     [ Opcodes.i32_add ],
     [ Opcodes.local_set, Pointer ]
   ])
@@ -134,21 +134,21 @@ const checkFailure = () => [
   [ Opcodes.br_if, 0 ]
 ];
 
-const generateChar = (node, negated, get) => {
+const generateChar = (node, negated, get, stringSize) => {
   return [
-    ...(get ? getNextChar() : []),
+    ...(get ? getNextChar(stringSize) : []),
     ...number(node.char.charCodeAt(0), Valtype.i32),
     negated ? [ Opcodes.i32_eq ] : [ Opcodes.i32_ne ],
     ...(get ? checkFailure(): [])
   ];
 };
 
-const generateSet = (node, negated, get) => {
+const generateSet = (node, negated, get, stringSize) => {
   // for a single char we do not need a tmp, it is like just
   const singleChar = node.body.length === 1 && node.body[0].type === 'Character';
 
   let out = [
-    ...(get ? getNextChar() : []),
+    ...(get ? getNextChar(stringSize) : []),
     ...(singleChar ? [] : [ [ Opcodes.local_set, Tmp ] ]),
   ];
 
@@ -156,7 +156,7 @@ const generateSet = (node, negated, get) => {
     out = [
       ...out,
       ...(singleChar ? [] : [ [ Opcodes.local_get, Tmp ] ]),
-      ...generate(x, negated, false)
+      ...generate(x, negated, false, stringSize)
     ];
   }
 
@@ -168,9 +168,9 @@ const generateSet = (node, negated, get) => {
   ];
 };
 
-const generateRange = (node, negated, get) => {
+const generateRange = (node, negated, get, stringSize) => {
   return [
-    ...(get ? getNextChar() : []),
+    ...(get ? getNextChar(stringSize) : []),
     ...(get ? [ [ Opcodes.local_tee, Tmp ] ] : []),
 
     ...number(node.from.charCodeAt(0), Valtype.i32),
@@ -192,8 +192,25 @@ const generateGroup = (node, negated, get) => {
   return [];
 };
 
-export const test = (regex, index = 0, name = 'regex_test_' + regex) => outputFunc(generate(parse(regex), false, true, 'test'), name, index);
-export const search = (regex, index = 0, name = 'regex_search_' + regex) => outputFunc(generate(parse(regex), false, true, 'search'), name, index);
+const wrapFunc = (regex, func, name, index) => {
+  const parsed = parse(regex);
+
+  return outputFunc([
+    [ Opcodes.local_get, 1 ],
+    ...number(TYPES.string, Valtype.i32),
+    [ Opcodes.i32_eq ],
+    [ Opcodes.if, Valtype.i32 ],
+    // string
+    ...generate(parsed, false, true, 2, func),
+    [ Opcodes.else ],
+    // bytestring
+    ...generate(parsed, false, true, 1, func),
+    [ Opcodes.end ]
+  ], name, index);
+};
+
+export const test = (regex, index = 0, name = 'regex_test_' + regex) => wrapFunc(regex, 'test', name, index);
+export const search = (regex, index = 0, name = 'regex_search_' + regex) => wrapFunc(regex, 'search', name, index);
 
 const outputFunc = (wasm, name, index) => ({
   name,
@@ -201,9 +218,9 @@ const outputFunc = (wasm, name, index) => ({
   wasm,
 
   export: true,
-  params: [ Valtype.i32 ],
+  params: [ Valtype.i32, Valtype.i32 ],
   returns: [ Valtype.i32 ],
-  returnType: 0xffffffffffff1, // boolean - todo: do not hardcode this
+  returnType: TYPES.boolean,
   locals: {
     basePointer: { idx: 0, type: Valtype.i32 },
     iterPointer: { idx: 1, type: Valtype.i32 },

package/rhemyn/parse.js

@@ -11,7 +11,7 @@ const Quantifiers = {
 const QuantifierKeys = Object.keys(Quantifiers);
 
 const getArg = (name, def) => {
-  const arg = (typeof process !== 'undefined' ? process.argv : Deno.args).find(x => x.startsWith(`-${name}=`));
+  const arg = (typeof process !== 'undefined' ? process.argv : Deno.args).find(x => x.startsWith(`--${name}=`));
   if (arg) return arg.split('=')[0];
 
   return def;
@@ -20,8 +20,7 @@ const getArg = (name, def) => {
 // full is spec-compliant but slower. not needed most of the time. (evil)
 const DotChars = () => ({
   full: [ '\n', '\r', '\u2028', '\u2029' ],
-  simple: [ '\n', '\r' ],
-  fast: [ '\n' ]
+  fast: [ '\n', '\r' ]
 })[getArg('regex-dot', 'fast')];
 
 const WordChars = () => ({
@@ -31,8 +30,8 @@ const WordChars = () => ({
 
 const WhitespaceChars = () => ({
   full: [ ' ', '\t', '\n', '\r', '\u2028', '\u2029' ],
-  simple: [ ' ', '\t', '\n', '\r' ]
-})[getArg('regex-ws', 'simple')];
+  fast: [ ' ', '\t', '\n', '\r' ]
+})[getArg('regex-ws', 'fast')];
 
 const _Metachars = () => ({
   unescaped: {

package/runner/index.js

@@ -5,7 +5,7 @@ import fs from 'node:fs';
 
 const start = performance.now();
 
-if (process.argv.includes('-compile-hints')) {
+if (process.argv.includes('--compile-hints')) {
   const v8 = await import('node:v8');
   v8.setFlagsFromString(`--experimental-wasm-compilation-hints`);
 
@@ -22,11 +22,39 @@ if (process.argv.includes('--help')) {
   console.log(`\x1B[1m\x1B[35mPorffor\x1B[0m is a JavaScript engine/runtime/compiler. \x1B[90m(${(await import('./version.js')).default})\x1B[0m`);
 
   // basic usage
-  console.log(`\x1B[1mporf [command] path/to/script.js [...prefs] [...args]\x1B[0m`);
+  console.log(`Usage: \x1B[1mporf [command] path/to/script.js [...prefs] [...args]\x1B[0m`);
 
   // commands
-  console.log(`\x1B[1mporf [subcommand] path/to/script.js [...prefs] [...args]\x1B[0m`);
+  console.log(`\n\u001b[4mCommands\x1B[0m`);
+  for (const [ cmd, [ color, desc ] ] of Object.entries({
+    run: [ 34, 'Run a JS file' ],
+    wasm: [ 34, 'Compile a JS file to a Wasm binary\n' ],
+    c: [ 31, 'Compile a JS file to C source code' ],
+    native: [ 31, 'Compile a JS file to a native binary\n' ],
+    profile: [ 33, 'Profile a JS file' ],
+    debug: [ 33, 'Debug a JS file' ],
+    'debug-wasm': [ 33, 'Debug the compiled Wasm of a JS file' ]
+  })) {
+    console.log(`  \x1B[1m\x1B[${color}m${cmd}\x1B[0m${' '.repeat(20 - cmd.length - (desc.startsWith('🧪') ? 3 : 0))}${desc}`);
+  }
 
+  // console.log();
+
+  // // options
+  // console.log(`\n\u001b[4mCommands\x1B[0m`);
+  // for (const [ cmd, [ color, desc ] ] of Object.entries({
+  //   run: [ 34, 'Run a JS file' ],
+  //   wasm: [ 34, 'Compile a JS file to a Wasm binary\n' ],
+  //   c: [ 31, 'Compile a JS file to C source code' ],
+  //   native: [ 31, 'Compile a JS file to a native binary\n' ],
+  //   profile: [ 33, 'Profile a JS file' ],
+  //   debug: [ 33, 'Debug a JS file' ],
+  //   'debug-wasm': [ 33, 'Debug the compiled Wasm of a JS file' ]
+  // })) {
+  //   console.log(`  \x1B[1m\x1B[${color}m${cmd}\x1B[0m${' '.repeat(20 - cmd.length - (desc.startsWith('🧪') ? 3 : 0))}${desc}`);
+  // }
+
+  console.log();
   process.exit(0);
 }
 
@@ -39,11 +67,11 @@ if (['run', 'wasm', 'native', 'c', 'profile', 'debug', 'debug-wasm'].includes(fi
   }
 
   if (['wasm', 'native', 'c'].includes(file)) {
-    process.argv.push(`-target=${file}`);
+    process.argv.push(`--target=${file}`);
   }
 
   if (file === 'debug-wasm') {
-    process.argv.push('-asur', '-wasm-debug');
+    process.argv.push('--asur', '--wasm-debug');
   }
 
   if (file === 'debug') {
@@ -90,7 +118,7 @@ try {
   if (process.argv.includes('-b')) {
     const { wasm, exports } = await compile(source, process.argv.includes('--module') ? [ 'module' ] : [], {}, print);
 
-    if (!process.argv.includes('-no-run')) exports.main();
+    if (!process.argv.includes('--no-run')) exports.main();
 
     console.log(`\n\nwasm size: ${wasm.byteLength} bytes`);
   } else {

package/runner/repl.js

@@ -21,7 +21,7 @@ try {
 
 globalThis.valtype = 'f64';
 
-const valtypeOpt = process.argv.find(x => x.startsWith('-valtype='));
+const valtypeOpt = process.argv.find(x => x.startsWith('--valtype='));
 if (valtypeOpt) valtype = valtypeOpt.split('=')[1];
 
 let host = globalThis?.navigator?.userAgent;
@@ -116,7 +116,7 @@ replServer.defineCommand('asm', {
     this.clearBufferedCommand();
 
     try {
-      process.argv.push('-opt-funcs');
+      process.argv.push('--opt-funcs');
       await run('', null, null, () => {}, false);
       process.argv.pop();
     } catch { }

package/runner/sizes.js

@@ -18,7 +18,7 @@ const perform = async (file, args) => {
   console.log(label, ' '.repeat(40 - label.length), `${size}b`);
 };
 
-const argsValtypes = [ '-valtype=i32', '-valtype=i64', '-valtype=f64' ];
+const argsValtypes = [ '--valtype=i32', '--valtype=i64', '--valtype=f64' ];
 const argsOptlevels = [ '-O0', '-O1', '-O2', '-O3' ];
 
 for (const file of [ 'bench/prime_basic.js', 'bench/fib_iter.js', 'test/math_1.js', 'test/math_3.js', 'test/while_1.js', 'test/for_2.js', 'test/unary_3.js', 'test/updateexp_1.js', 'test/eq_3.js', 'test/empty.js' ]) {