porffor 0.2.0-cb647c8 → 0.2.0-dcc06c8

package/compiler/opt.js

@@ -2,6 +2,7 @@ import { Opcodes, Valtype } from "./wasmSpec.js";
 import { number } from "./embedding.js";
 import { read_signedLEB128, read_ieee754_binary64 } from "./encoding.js";
 import { log } from "./log.js";
+import Prefs from './prefs.js';
 
 const performWasmOp = (op, a, b) => {
   switch (op) {
@@ -11,21 +12,21 @@ const performWasmOp = (op, a, b) => {
   }
 };
 
-export default (funcs, globals, pages, tags) => {
+export default (funcs, globals, pages, tags, exceptions) => {
   const optLevel = parseInt(process.argv.find(x => x.startsWith('-O'))?.[2] ?? 1);
   if (optLevel === 0) return;
 
-  const tailCall = process.argv.includes('-tail-call');
+  const tailCall = Prefs.tailCall;
   if (tailCall) log.warning('opt', 'tail call proposal is not widely implemented! (you used -tail-call)');
 
-  if (optLevel >= 2 && !process.argv.includes('-opt-no-inline')) {
+  if (optLevel >= 2 && !Prefs.optNoInline) {
     // inline pass (very WIP)
     // get candidates for inlining
     // todo: pick smart in future (if func is used <N times? or?)
     const callsSelf = f => f.wasm.some(x => x[0] === Opcodes.call && x[1] === f.index);
     const suitableReturns = wasm => wasm.reduce((acc, x) => acc + (x[0] === Opcodes.return), 0) <= 1;
     const candidates = funcs.filter(x => x.name !== 'main' && Object.keys(x.locals).length === x.params.length && (x.returns.length === 0 || suitableReturns(x.wasm)) && !callsSelf(x) && !x.throws).reverse();
-    if (optLog) {
+    if (Prefs.optLog) {
       log('opt', `found inline candidates: ${candidates.map(x => x.name).join(', ')} (${candidates.length}/${funcs.length - 1})`);
 
       let reasons = {};
@@ -53,7 +54,7 @@ export default (funcs, globals, pages, tags) => {
         for (let i = 0; i < tWasm.length; i++) {
           const inst = tWasm[i];
           if (inst[0] === Opcodes.call && inst[1] === c.index) {
-            if (optLog) log('opt', `inlining call for ${c.name} (in ${t.name})`);
+            if (Prefs.optLog) log('opt', `inlining call for ${c.name} (in ${t.name})`);
             tWasm.splice(i, 1); // remove this call
 
             // add params as locals and set in reverse order
@@ -80,7 +81,7 @@ export default (funcs, globals, pages, tags) => {
             // adjust local operands to go to correct param index
             for (const inst of iWasm) {
               if ((inst[0] === Opcodes.local_get || inst[0] === Opcodes.local_set) && inst[1] < c.params.length) {
-                if (optLog) log('opt', `replacing local operand in inlined wasm (${inst[1]} -> ${paramIdx[inst[1]]})`);
+                if (Prefs.optLog) log('opt', `replacing local operand in inlined wasm (${inst[1]} -> ${paramIdx[inst[1]]})`);
                 inst[1] = paramIdx[inst[1]];
               }
             }
@@ -97,9 +98,10 @@ export default (funcs, globals, pages, tags) => {
     }
   }
 
-  if (process.argv.includes('-opt-inline-only')) return;
+  if (Prefs.optInlineOnly) return;
 
   const tagUse = tags.reduce((acc, x) => { acc[x.idx] = 0; return acc; }, {});
+  const exceptionUse = exceptions.reduce((acc, _, i) => { acc[i] = 0; return acc; }, {});
 
   // wasm transform pass
   for (const f of funcs) {
@@ -129,7 +131,12 @@ export default (funcs, globals, pages, tags) => {
         if (inst[0] === Opcodes.local_get) getCount[inst[1]]++;
         if (inst[0] === Opcodes.local_set || inst[0] === Opcodes.local_tee) setCount[inst[1]]++;
 
-        if (inst[0] === Opcodes.throw) tagUse[inst[1]]++;
+        if (inst[0] === Opcodes.throw) {
+          tagUse[inst[1]]++;
+
+          const exceptId = read_signedLEB128(wasm[i - 1].slice(1));
+          exceptionUse[exceptId]++;
+        }
 
         if (inst[0] === Opcodes.block) {
           // remove unneeded blocks (no brs inside)
@@ -160,7 +167,7 @@ export default (funcs, globals, pages, tags) => {
 
             wasm.splice(j - 1, 1); // remove end of this block
 
-            if (optLog) log('opt', `removed unneeded block in for loop`);
+            if (Prefs.optLog) log('opt', `removed unneeded block in for loop`);
           }
         }
 
@@ -209,7 +216,7 @@ export default (funcs, globals, pages, tags) => {
             i -= 4;
             inst = wasm[i];
 
-            if (optLog) log('opt', `removed unneeded typeswitch check`);
+            if (Prefs.optLog) log('opt', `removed unneeded typeswitch check`);
           }
         }
 
@@ -232,7 +239,7 @@ export default (funcs, globals, pages, tags) => {
             wasm.splice(j - 1, 2, [ Opcodes.drop ]); // remove typeswitch start
             wasm.splice(i - 1, 1); // remove this inst
 
-            if (optLog) log('opt', 'removed unneeded entire typeswitch');
+            if (Prefs.optLog) log('opt', 'removed unneeded entire typeswitch');
 
             if (i > 0) i--;
             continue;
@@ -261,7 +268,7 @@ export default (funcs, globals, pages, tags) => {
 
           getCount[inst[1]]--;
           i--;
-          // if (optLog) log('opt', `consolidated set, get -> tee`);
+          // if (Prefs.optLog) log('opt', `consolidated set, get -> tee`);
           continue;
         }
 
@@ -329,7 +336,7 @@ export default (funcs, globals, pages, tags) => {
 
           wasm.splice(i - 1, 2); // remove this inst and last
           i -= 2;
-          // if (optLog) log('opt', `removed redundant i32 -> i64 -> i32 conversion ops`);
+          // if (Prefs.optLog) log('opt', `removed redundant i32 -> i64 -> i32 conversion ops`);
           continue;
         }
 
@@ -342,7 +349,7 @@ export default (funcs, globals, pages, tags) => {
 
           wasm.splice(i - 1, 2); // remove this inst and last
           i -= 2;
-          // if (optLog) log('opt', `removed redundant i32 -> f64 -> i32 conversion ops`);
+          // if (Prefs.optLog) log('opt', `removed redundant i32 -> f64 -> i32 conversion ops`);
           continue;
         }
 
@@ -357,7 +364,7 @@ export default (funcs, globals, pages, tags) => {
 
           wasm.splice(i, 1); // remove this inst
           i--;
-          if (optLog) log('opt', `converted const -> i32 convert into i32 const`);
+          if (Prefs.optLog) log('opt', `converted const -> i32 convert into i32 const`);
           continue;
         }
 
@@ -372,7 +379,7 @@ export default (funcs, globals, pages, tags) => {
 
           wasm.splice(i, 1); // remove this inst
           i--;
-          if (optLog) log('opt', `converted i32 const -> convert into const`);
+          if (Prefs.optLog) log('opt', `converted i32 const -> convert into const`);
           continue;
         }
 
@@ -387,7 +394,7 @@ export default (funcs, globals, pages, tags) => {
 
           wasm.splice(i, 1); // remove this inst (return)
           i--;
-          if (optLog) log('opt', `tail called return, call`);
+          if (Prefs.optLog) log('opt', `tail called return, call`);
           continue;
         }
 
@@ -400,7 +407,7 @@ export default (funcs, globals, pages, tags) => {
 
           wasm.splice(i, 1); // remove this inst (return)
           i--;
-          // if (optLog) log('opt', `removed redundant return at end`);
+          // if (Prefs.optLog) log('opt', `removed redundant return at end`);
           continue;
         }
 
@@ -430,7 +437,7 @@ export default (funcs, globals, pages, tags) => {
           // <nothing>
 
           wasm.splice(i - 2, 3); // remove this, last, 2nd last insts
-          if (optLog) log('opt', `removed redundant inline param local handling`);
+          if (Prefs.optLog) log('opt', `removed redundant inline param local handling`);
           i -= 3;
           continue;
         }
@@ -438,12 +445,12 @@ export default (funcs, globals, pages, tags) => {
 
       if (optLevel < 2) continue;
 
-      if (optLog) log('opt', `get counts: ${Object.keys(f.locals).map(x => `${x} (${f.locals[x].idx}): ${getCount[f.locals[x].idx]}`).join(', ')}`);
+      if (Prefs.optLog) log('opt', `get counts: ${Object.keys(f.locals).map(x => `${x} (${f.locals[x].idx}): ${getCount[f.locals[x].idx]}`).join(', ')}`);
 
       // remove unneeded var: remove pass
       // locals only got once. we don't need to worry about sets/else as these are only candidates and we will check for matching set + get insts in wasm
       let unneededCandidates = Object.keys(getCount).filter(x => getCount[x] === 0 || (getCount[x] === 1 && setCount[x] === 0)).map(x => parseInt(x));
-      if (optLog) log('opt', `found unneeded locals candidates: ${unneededCandidates.join(', ')} (${unneededCandidates.length}/${Object.keys(getCount).length})`);
+      if (Prefs.optLog) log('opt', `found unneeded locals candidates: ${unneededCandidates.join(', ')} (${unneededCandidates.length}/${Object.keys(getCount).length})`);
 
       // note: disabled for now due to instability
       if (unneededCandidates.length > 0 && false) for (let i = 0; i < wasm.length; i++) {
@@ -461,7 +468,7 @@ export default (funcs, globals, pages, tags) => {
           wasm.splice(i - 1, 2); // remove insts
           i -= 2;
           delete f.locals[Object.keys(f.locals)[inst[1]]]; // remove from locals
-          if (optLog) log('opt', `removed redundant local (get set ${inst[1]})`);
+          if (Prefs.optLog) log('opt', `removed redundant local (get set ${inst[1]})`);
         }
 
         if (inst[0] === Opcodes.local_tee && unneededCandidates.includes(inst[1])) {
@@ -489,7 +496,7 @@ export default (funcs, globals, pages, tags) => {
           unneededCandidates.splice(unneededCandidates.indexOf(inst[1]), 1);
           unneededCandidates = unneededCandidates.map(x => x > removedIdx ? (x - 1) : x);
 
-          if (optLog) log('opt', `removed redundant local ${localName} (tee ${inst[1]})`);
+          if (Prefs.optLog) log('opt', `removed redundant local ${localName} (tee ${inst[1]})`);
         }
       }
 
@@ -525,7 +532,7 @@ export default (funcs, globals, pages, tags) => {
           let b = lastInst[1];
 
           const val = performWasmOp(inst[0], a, b);
-          if (optLog) log('opt', `inlined math op (${a} ${inst[0].toString(16)} ${b} -> ${val})`);
+          if (Prefs.optLog) log('opt', `inlined math op (${a} ${inst[0].toString(16)} ${b} -> ${val})`);
 
           wasm.splice(i - 2, 3, ...number(val)); // remove consts, math op and add new const
           i -= 2;
@@ -537,12 +544,12 @@ export default (funcs, globals, pages, tags) => {
       for (const x in useCount) {
         if (useCount[x] === 0) {
           const name = Object.keys(f.locals)[localIdxs.indexOf(parseInt(x))];
-          if (optLog) log('opt', `removed internal local ${x} (${name})`);
+          if (Prefs.optLog) log('opt', `removed internal local ${x} (${name})`);
           delete f.locals[name];
         }
       }
 
-      if (optLog) log('opt', `final use counts: ${Object.keys(f.locals).map(x => `${x} (${f.locals[x].idx}): ${useCount[f.locals[x].idx]}`).join(', ')}`);
+      if (Prefs.optLog) log('opt', `final use counts: ${Object.keys(f.locals).map(x => `${x} (${f.locals[x].idx}): ${useCount[f.locals[x].idx]}`).join(', ')}`);
     }
   }
 
@@ -553,5 +560,11 @@ export default (funcs, globals, pages, tags) => {
     }
   }
 
+  for (const x of Object.keys(exceptionUse).sort((a, b) => b - a)) {
+    if (exceptionUse[x] === 0) {
+      exceptions.splice(+x, 1);
+    }
+  }
+
   // return funcs;
 };

package/compiler/parse.js

@@ -1,4 +1,6 @@
 import { log } from "./log.js";
+import Prefs from './prefs.js';
+
 // import { parse } from 'acorn';
 
 // deno compat
@@ -8,8 +10,8 @@ if (typeof process === 'undefined' && typeof Deno !== 'undefined') {
 }
 
 // should we try to support types (while parsing)
-const types = process.argv.includes('-parse-types');
-globalThis.typedInput = types && process.argv.includes('-opt-types');
+const types = Prefs.parseTypes;
+globalThis.typedInput = types && Prefs.optTypes;
 
 // todo: review which to use by default
 // supported parsers:

package/compiler/precompile.js

@@ -0,0 +1,79 @@
+import fs from 'node:fs';
+import { join } from 'node:path';
+
+import { fileURLToPath } from 'node:url';
+const __dirname = fileURLToPath(new URL('.', import.meta.url));
+
+// import porfParse from './parse.js';
+// import porfCodegen from './codeGen.js';
+
+const argv = process.argv.slice();
+
+const compile = async (file, [ _funcs, _globals ]) => {
+  const source = fs.readFileSync(file, 'utf8');
+  const first = source.slice(0, source.indexOf('\n'));
+
+  let args = ['-bytestring'];
+  if (file.endsWith('.ts')) args.push('-parse-types', '-opt-types');
+  if (first.startsWith('// @porf')) {
+    args = args.concat(first.slice('// @porf '.length).split(' '));
+  }
+  process.argv = argv.concat(args);
+
+  // globalThis.optLog = process.argv.includes('-opt-log');
+  // globalThis.codeLog = process.argv.includes('-code-log');
+  // globalThis.allocLog = process.argv.includes('-alloc-log');
+  // globalThis.regexLog = process.argv.includes('-regex-log');
+
+  // const porfParse = (await import(`./parse.js?_=${Date.now()}`)).default;
+  // const porfCodegen = (await import(`./codeGen.js?_=${Date.now()}`)).default;
+
+  // let { funcs, globals, data } = porfCodegen(porfParse(source, ['module']));
+
+  const porfCompile = (await import(`./index.js?_=${Date.now()}`)).default;
+
+  let { funcs, globals, data, exceptions } = porfCompile(source, ['module']);
+
+  funcs = funcs.filter(x => x.export);
+  for (const x of funcs) {
+    if (x.data) x.data = x.data.map(x => data[x]);
+    if (x.exceptions) x.exceptions = x.exceptions.map(x => {
+      const obj = exceptions[x];
+      if (obj) obj.exceptId = x;
+      return obj;
+    }).filter(x => x);
+  }
+
+  _funcs.push(...funcs);
+  _globals.push(...Object.values(globals));
+};
+
+const precompile = async () => {
+  const dir = join(__dirname, 'builtins');
+
+  let funcs = [], globals = [];
+  for (const file of fs.readdirSync(dir)) {
+    if (file.endsWith('.d.ts')) continue;
+    await compile(join(dir, file), [ funcs, globals ]);
+  }
+
+  // todo: globals, exceptions, pages per func
+
+  return `// autogenerated by precompile.js
+
+export const BuiltinFuncs = function() {
+${funcs.map(x => `  this.${x.name} = {
+    wasm: ${JSON.stringify(x.wasm)},
+    params: ${JSON.stringify(x.params)},
+    typedParams: true,
+    returns: ${JSON.stringify(x.returns)},
+    typedReturns: true,
+    locals: ${JSON.stringify(Object.values(x.locals).slice(x.params.length * 2).map(x => x.type))},
+${x.pages && x.pages.size > 0 ? `    pages: ${JSON.stringify(Object.fromEntries(x.pages.entries()))},` : ''}
+${x.data && x.data.length > 0 ? `    data: ${JSON.stringify(x.data)},` : ''}
+${x.exceptions && x.exceptions.length > 0 ? `    exceptions: ${JSON.stringify(x.exceptions)},` : ''}
+  };`.replaceAll('\n\n', '\n')).join('\n')}
+}`;
+};
+
+console.log(await precompile());

package/compiler/prefs.js

@@ -0,0 +1,22 @@
+const cache = {};
+const obj = new Proxy({}, {
+  get(_, p) {
+    // intentionally misses with undefined values cached
+    if (cache[p]) return cache[p];
+
+    return cache[p] = (() => {
+      // fooBar -> foo-bar
+      const name = p[0] === '_' ? p : p.replace(/[A-Z]/g, c => `-${c.toLowerCase()}`);
+      if (process.argv.includes('-' + name)) return true;
+
+      const valArg = process.argv.find(x => x.startsWith(`-${name}=`));
+      if (valArg) return valArg.slice(name.length + 2);
+
+      return undefined;
+    })();
+  }
+});
+
+obj.uncache = () => cache = {};
+
+export default obj;

package/compiler/prototype.js

@@ -2,6 +2,7 @@ import { Opcodes, Blocktype, Valtype, ValtypeSize, PageSize } from "./wasmSpec.j
 import { number } from "./embedding.js";
 import { unsignedLEB128 } from "./encoding.js";
 import { UNDEFINED } from "./builtins.js";
+import Prefs from './prefs.js';
 
 // todo: do not duplicate this
 const TYPES = {
@@ -23,9 +24,10 @@ const TYPES = {
 // todo: turn these into built-ins once arrays and these become less hacky
 
 export const PrototypeFuncs = function() {
-  const noUnlikelyChecks = process.argv.includes('-funsafe-no-unlikely-proto-checks');
-  let zeroChecks = process.argv.find(x => x.startsWith('-funsafe-zero-proto-checks='));
-  if (zeroChecks) zeroChecks = zeroChecks.split('=')[1].split(',').reduce((acc, x) => { acc[x.toLowerCase()] = true; return acc; }, {});
+  const noUnlikelyChecks = Prefs.funsafeNoUnlikelyProtoChecks;
+
+  let zeroChecks;
+  if (Prefs.zeroChecks) zeroChecks = Prefs.zeroChecks.split('=')[1].split(',').reduce((acc, x) => { acc[x.toLowerCase()] = true; return acc; }, {});
     else zeroChecks = {};
 
   this[TYPES._array] = {
@@ -342,8 +344,8 @@ export const PrototypeFuncs = function() {
         ...number(0, Valtype.i32), // base 0 for store later
 
         ...wIndex,
-
         Opcodes.i32_to,
+
         ...number(ValtypeSize.i16, Valtype.i32),
         [ Opcodes.i32_mul ],
 
@@ -489,10 +491,10 @@ export const PrototypeFuncs = function() {
   this[TYPES.string].isWellFormed.local2 = Valtype.i32;
   this[TYPES.string].isWellFormed.returnType = TYPES.boolean;
 
-  if (process.argv.includes('-bytestring')) {
+  if (Prefs.bytestring) {
     this[TYPES._bytestring] = {
       at: (pointer, length, wIndex, iTmp, _, arrayShell) => {
-        const [ newOut, newPointer ] = arrayShell(1, 'i16');
+        const [ newOut, newPointer ] = arrayShell(1, 'i8');
 
         return [
           // setup new/out array
@@ -548,7 +550,7 @@ export const PrototypeFuncs = function() {
 
       // todo: out of bounds properly
       charAt: (pointer, length, wIndex, _1, _2, arrayShell) => {
-        const [ newOut, newPointer ] = arrayShell(1, 'i16');
+        const [ newOut, newPointer ] = arrayShell(1, 'i8');
 
         return [
           // setup new/out array
@@ -558,7 +560,6 @@ export const PrototypeFuncs = function() {
           ...number(0, Valtype.i32), // base 0 for store later
 
           ...wIndex,
-
           Opcodes.i32_to,
 
           ...pointer,

package/compiler/sections.js

@@ -3,6 +3,7 @@ import { encodeVector, encodeString, encodeLocal, unsignedLEB128, signedLEB128 }
 import { number } from './embedding.js';
 import { importedFuncs } from './builtins.js';
 import { log } from "./log.js";
+import Prefs from './prefs.js';
 
 const createSection = (type, data) => [
   type,
@@ -26,12 +27,12 @@ export default (funcs, globals, tags, pages, data, flags) => {
 
   const optLevel = parseInt(process.argv.find(x => x.startsWith('-O'))?.[2] ?? 1);
 
-  const compileHints = process.argv.includes('-compile-hints');
+  const compileHints = Prefs.compileHints;
   if (compileHints) log.warning('sections', 'compile hints is V8 only w/ experimental arg! (you used -compile-hints)');
 
   const getType = (params, returns) => {
     const hash = `${params.join(',')}_${returns.join(',')}`;
-    if (optLog) log('sections', `getType(${JSON.stringify(params)}, ${JSON.stringify(returns)}) -> ${hash} | cache: ${typeCache[hash]}`);
+    if (Prefs.optLog) log('sections', `getType(${JSON.stringify(params)}, ${JSON.stringify(returns)}) -> ${hash} | cache: ${typeCache[hash]}`);
     if (optLevel >= 1 && typeCache[hash] !== undefined) return typeCache[hash];
 
     const type = [ FuncType, ...encodeVector(params), ...encodeVector(returns) ];
@@ -79,7 +80,7 @@ export default (funcs, globals, tags, pages, data, flags) => {
   }
   globalThis.importFuncs = importFuncs;
 
-  if (optLog) log('sections', `treeshake: using ${importFuncs.length}/${importedFuncs.length} imports`);
+  if (Prefs.optLog) log('sections', `treeshake: using ${importFuncs.length}/${importedFuncs.length} imports`);
 
   const importSection = importFuncs.length === 0 ? [] : createSection(
     Section.import,
@@ -95,7 +96,7 @@ export default (funcs, globals, tags, pages, data, flags) => {
   // https://github.com/WebAssembly/design/issues/1473#issuecomment-1431274746
   const chSection = !compileHints ? [] : customSection(
     'compilationHints',
-    // for now just do everything as optimise eager
+    // for now just do everything as optimize eager
     encodeVector(funcs.map(_ => chHint(0x02, 0x02, 0x02)))
   );
 
@@ -106,7 +107,7 @@ 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 (process.argv.includes('-always-memory') && 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;
@@ -169,7 +170,7 @@ export default (funcs, globals, tags, pages, data, flags) => {
     unsignedLEB128(data.length)
   );
 
-  if (process.argv.includes('-sections')) console.log({
+  if (Prefs.sections) console.log({
     typeSection: typeSection.map(x => x.toString(16)),
     importSection: importSection.map(x => x.toString(16)),
     funcSection: funcSection.map(x => x.toString(16)),

package/demo.js

@@ -1,3 +1,3 @@
-foo();
+// foo();
 
-// console.log('Hello, World!');
+console.log('Hello, World!');

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-cb647c8",
+  "version": "0.2.0-dcc06c8",
   "author": "CanadaHonk",
   "license": "MIT",
   "dependencies": {

package/rhemyn/compile.js

@@ -2,6 +2,7 @@ import { Blocktype, Opcodes, Valtype, PageSize, ValtypeSize } from '../compiler/
 import { number } from '../compiler/embedding.js';
 import { signedLEB128, unsignedLEB128 } from '../compiler/encoding.js';
 import parse from './parse.js';
+import Prefs from '../compiler/prefs.js';
 
 // local indexes
 const BasePointer = 0; // base string pointer
@@ -80,7 +81,7 @@ const generate = (node, negated = false, get = true, func = 'test') => {
         })[func], Valtype.i32)
       ];
 
-      if (globalThis.regexLog) {
+      if (Prefs.regexLog) {
         const underline = x => `\u001b[4m\u001b[1m${x}\u001b[0m`;
         console.log(`\n${underline('ast')}`);
         console.log(node);