diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 26a38482a..edb2b5731 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -217,17 +217,22 @@ jobs: # worker budget covers both processes; file and case shard axes stay # separate so every fixture still runs exactly once in each lane. # The tests run in a transient scope capped below the runner's 16 GB, so - # the runner itself keeps memory: past 14 GB of RAM the scope swaps, as - # the whole runner did before (a sanitized shard peaks near 16 GB), and - # when RAM and swap run out the kernel kills the scope's largest process. + # the runner itself keeps memory: past 15 GB of RAM the scope swaps, as + # the whole runner did before, and when RAM and swap run out the kernel + # kills the scope's largest process. A sanitized shard peaks near 16 GB, + # and the sanitized self-hosting native-frontend build alone needs 15 GB + # (its clang compile of the frontend takes 10.5 GB). # OOMPolicy=continue leaves the rest of the scope running, so vitest - # reports the killed worker and the job log names the test. + # reports the killed worker and the job log names the test. Node sizes + # its default heap from the memory it may use, which the scope lowers + # below 4 GB; the workers keep the 4 GB heap they had on the bare + # runner. - name: Tests (corpus and integration suites) run: | ulimit -c 0 if [ -d /run/systemd/system ]; then - exec sudo --preserve-env systemd-run --scope --quiet -p MemoryMax=14G -p OOMPolicy=continue \ - --uid="$(id -u)" --gid="$(id -g)" -- env PATH="$PATH" HOME="$HOME" node scripts/ci-test.mjs + exec sudo --preserve-env systemd-run --scope --quiet -p MemoryMax=15G -p OOMPolicy=continue \ + --uid="$(id -u)" --gid="$(id -g)" -- env PATH="$PATH" HOME="$HOME" NODE_OPTIONS="--max-old-space-size=4096" node scripts/ci-test.mjs fi node scripts/ci-test.mjs diff --git a/benchmarks/runtime/README.md b/benchmarks/runtime/README.md index 4f152e4bb..03238871d 100644 --- a/benchmarks/runtime/README.md +++ b/benchmarks/runtime/README.md @@ -13,9 +13,12 @@ Application-shaped workloads for measuring the speed of compiled executables. Ea | `cyclic-graphs` | cpu | object graphs with back references built and dropped: trees with parent and sibling links, rings of records sharing a long-lived registry, closures capturing their owner | | `dictionary-walks` | cpu | `Record` dictionaries walked with for-in and `Object.keys`/`values`/`entries` while entries are added, updated and deleted | | `dynamic-methods` | cpu | untyped JavaScript calling builtin methods on values it does not type: string `slice`/`indexOf`/`toUpperCase`/`padStart`, array `push`/`pop`/`includes`, `Map` and `Set`, and an object's own methods | +| `dynamic-properties` | cpu | untyped JavaScript reading and writing properties of values it does not type: records built by object literals in two key orders, nested records, and instances of a class and its subclass | | `functional-pipeline` | cpu | `filter`/`map`/`reduce`/`flatMap` chains, closures, small records | | `intern-table`, `intern-table-map`, `intern-table-1`, `intern-table-map-1` | cpu | worker threads interning the same type-like objects under string keys and records under number keys, then hits and misses: one shared `InternTable` per kind of key, or a module-level `Map` per thread (the `-map` entries), with 8 threads or 1 (the `-1` entries); under Node.js every thread has its own tables | | `json-records` | cpu | `JSON.stringify`, typed `JSON.parse`, Map aggregation | +| `json-unions` | cpu | `JSON.parse` into discriminated unions: an event log with eight kinds and nested expression trees with nine, evaluated after the cast | +| `keyed-option-writes` | cpu | class methods whose names are also keys of option records written through element accesses: `enabled[name] = true` over a union of literal names and `totals["count"] = n` | | `log-lines` | cpu | 200k immediate `console.log`/`process.stdout.write` lines into the runner's pipe | | `log-summary`, `inventory-report` | cpu | the multi-module applications from `benchmarks/builds`, on 300k-line inputs | | `number-maps` | cpu | number-keyed Map/Set: a dense-id table, sparse ids in many small maps, get/has/set mixes | diff --git a/benchmarks/runtime/workloads/dynamic-properties.json b/benchmarks/runtime/workloads/dynamic-properties.json new file mode 100644 index 000000000..398981ef3 --- /dev/null +++ b/benchmarks/runtime/workloads/dynamic-properties.json @@ -0,0 +1,6 @@ +{ + "entry": "workloads/dynamic-properties.mjs", + "args": [ + "1000000" + ] +} diff --git a/benchmarks/runtime/workloads/dynamic-properties.mjs b/benchmarks/runtime/workloads/dynamic-properties.mjs new file mode 100644 index 000000000..d0ee3802c --- /dev/null +++ b/benchmarks/runtime/workloads/dynamic-properties.mjs @@ -0,0 +1,56 @@ +// Untyped JavaScript reading and writing properties of values it never +// declares types for: records built by object literals in two key orders, +// nested records, property updates, and instances of a class and its +// subclass read and written the same way. Argument: the number of +// iterations. +const rounds = Number(process.argv[2] ?? "1000000"); +class Cursor { + constructor(x, y) { + this.x = x; + this.y = y; + } + step(dx, dy) { + this.x += dx; + this.y += dy; + return this; + } +} +class Marker extends Cursor { + constructor(x, y, label) { + super(x, y); + this.label = label; + } +} +function makeItem(i) { + if (i % 3 === 0) + return { id: i, price: i % 17, qty: i % 5, meta: { tag: i % 7, weight: i % 11 } }; + return { id: i, qty: i % 5, price: i % 13, meta: { weight: i % 9, tag: i % 5 } }; +} +function cost(item) { + return item.price * item.qty + item.meta.weight; +} +function restock(item, round) { + item.qty = (item.qty + round) % 7; + item.meta.tag = (item.meta.tag + 1) % 7; +} +function distance(a, b) { + return Math.abs(a.x - b.x) + Math.abs(a.y - b.y); +} +const items = []; +for (let i = 0; i < 1024; i++) items.push(makeItem(i)); +const cursors = []; +for (let i = 0; i < 64; i++) + cursors.push(i % 2 === 0 ? new Cursor(i - 32, 7 - i) : new Marker(i - 32, 7 - i, "m" + i)); +const origin = { x: 0, y: 0 }; +let checksum = 0; +for (let r = 0; r < rounds; r++) { + const item = items[r & 1023]; + checksum += cost(item); + restock(item, r & 3); + checksum += item.meta.tag; + const cursor = cursors[r & 63]; + checksum += distance(cursor.step((r % 3) - 1, (r % 5) - 2), origin); + if (cursor.x > 100 || cursor.x < -100) cursor.x = 0; + if (cursor.y > 100 || cursor.y < -100) cursor.y = 0; +} +console.log(checksum, items[5].qty, items[6].meta.tag, cursors[3].x, cursors[3].y, cursors[3].label); diff --git a/benchmarks/runtime/workloads/json-unions.json b/benchmarks/runtime/workloads/json-unions.json new file mode 100644 index 000000000..925e8457b --- /dev/null +++ b/benchmarks/runtime/workloads/json-unions.json @@ -0,0 +1,6 @@ +{ + "entry": "workloads/json-unions.ts", + "args": [ + "6" + ] +} diff --git a/benchmarks/runtime/workloads/json-unions.ts b/benchmarks/runtime/workloads/json-unions.ts new file mode 100644 index 000000000..0f684c35f --- /dev/null +++ b/benchmarks/runtime/workloads/json-unions.ts @@ -0,0 +1,155 @@ +// JSON text parsed into discriminated unions: an event log with eight kinds +// of event and a nested expression tree with nine kinds of node, cast with +// `JSON.parse(text) as T`, then aggregated and evaluated. Argument: the +// number of rounds. +type Event = + | { kind: "click"; target: string; x: number; y: number } + | { key: string; kind: "key"; repeat: boolean } + | { kind: "scroll"; offset: number } + | { kind: "resize"; height: number; width: number } + | { kind: "focus"; target: string } + | { kind: "blur"; target: string } + | { kind: "load"; ms: number; url: string } + | { code: number; kind: "error"; message: string }; + +type Expr = + | { kind: "num"; value: number } + | { kind: "var"; name: string } + | { kind: "neg"; operand: Expr } + | { kind: "add"; left: Expr; right: Expr } + | { kind: "sub"; left: Expr; right: Expr } + | { kind: "mul"; left: Expr; right: Expr } + | { kind: "max"; items: Expr[] } + | { else: Expr; kind: "if"; test: Expr; then: Expr } + | { body: Expr; kind: "let"; name: string; value: Expr }; + +let seed = 12345; +function random(): number { + seed = (Math.imul(seed, 1103515245) + 12345) >>> 0; + return seed / 4294967296; +} + +const targets = ["menu", "search", "button", "list", "footer"]; +function eventText(i: number): string { + const target = targets[i % targets.length]!; + switch (i % 8) { + case 0: + return `{"kind":"click","target":"${target}","x":${i % 640},"y":${i % 480}}`; + case 1: + return `{"key":"k${i % 26}","kind":"key","repeat":${i % 3 === 0}}`; + case 2: + return `{"kind":"scroll","offset":${(i * 7) % 2000}}`; + case 3: + return `{"height":${400 + (i % 300)},"kind":"resize","width":${600 + (i % 500)}}`; + case 4: + return `{"kind":"focus","target":"${target}"}`; + case 5: + return `{"kind":"blur","target":"${target}"}`; + case 6: + return `{"kind":"load","ms":${i % 900},"url":"/page/${i % 50}"}`; + default: + return `{"code":${400 + (i % 100)},"kind":"error","message":"failed ${i % 13}"}`; + } +} + +const names = ["a", "b", "c", "d"]; +function exprText(depth: number): string { + const pick = depth <= 0 ? Math.floor(random() * 2) : Math.floor(random() * 9); + switch (pick) { + case 0: + return `{"kind":"num","value":${Math.floor(random() * 100)}}`; + case 1: + return `{"kind":"var","name":"${names[Math.floor(random() * names.length)]!}"}`; + case 2: + return `{"kind":"neg","operand":${exprText(depth - 1)}}`; + case 3: + case 4: + case 5: { + const kind = pick === 3 ? "add" : pick === 4 ? "sub" : "mul"; + return `{"kind":"${kind}","left":${exprText(depth - 1)},"right":${exprText(depth - 1)}}`; + } + case 6: + return `{"items":[${exprText(depth - 1)},${exprText(depth - 2)},${exprText(depth - 2)}],"kind":"max"}`; + case 7: + return `{"else":${exprText(depth - 2)},"kind":"if","test":${exprText(depth - 1)},"then":${exprText(depth - 2)}}`; + default: + return `{"body":${exprText(depth - 1)},"kind":"let","name":"${names[Math.floor(random() * names.length)]!}","value":${exprText(depth - 2)}}`; + } +} + +function evaluate(e: Expr, env: Map): number { + switch (e.kind) { + case "num": + return e.value; + case "var": + return env.get(e.name) ?? 1; + case "neg": + return -evaluate(e.operand, env); + case "add": + return (evaluate(e.left, env) + evaluate(e.right, env)) % 1000003; + case "sub": + return (evaluate(e.left, env) - evaluate(e.right, env)) % 1000003; + case "mul": + return (evaluate(e.left, env) * evaluate(e.right, env)) % 1000003; + case "max": { + let best = -Infinity; + for (const item of e.items) best = Math.max(best, evaluate(item, env)); + return best; + } + case "if": + return evaluate(e.test, env) > 0 ? evaluate(e.then, env) : evaluate(e.else, env); + case "let": { + const inner = new Map(env); + inner.set(e.name, evaluate(e.value, env)); + return evaluate(e.body, inner); + } + } +} + +const rounds = Number(process.argv[2] ?? "6"); +const eventsText = `[${Array.from({ length: 60000 }, (_, i) => eventText(i)).join(",")}]`; +const treesText = `[${Array.from({ length: 400 }, () => exprText(9)).join(",")}]`; +let clicks = 0; +let keys = 0; +let scrolled = 0; +let area = 0; +let loadMs = 0; +let errors = 0; +let focus = 0; +let total = 0; +for (let round = 0; round < rounds; round++) { + const events = JSON.parse(eventsText) as Event[]; + for (const event of events) { + switch (event.kind) { + case "click": + clicks += event.x + event.y; + break; + case "key": + if (event.repeat) keys++; + break; + case "scroll": + scrolled += event.offset; + break; + case "resize": + area += (event.width * event.height) % 1000; + break; + case "focus": + case "blur": + focus += event.target.length; + break; + case "load": + loadMs += event.ms; + break; + case "error": + errors += event.code % 7; + break; + } + } + const trees = JSON.parse(treesText) as Expr[]; + const env = new Map([ + ["a", 3], + ["b", 5], + ]); + for (const tree of trees) total = (total + evaluate(tree, env)) % 1000003; +} +console.log(clicks, keys, scrolled, area, loadMs, errors, focus, total); diff --git a/benchmarks/runtime/workloads/keyed-option-writes.json b/benchmarks/runtime/workloads/keyed-option-writes.json new file mode 100644 index 000000000..802911e7d --- /dev/null +++ b/benchmarks/runtime/workloads/keyed-option-writes.json @@ -0,0 +1,6 @@ +{ + "entry": "workloads/keyed-option-writes.ts", + "args": [ + "4" + ] +} diff --git a/benchmarks/runtime/workloads/keyed-option-writes.ts b/benchmarks/runtime/workloads/keyed-option-writes.ts new file mode 100644 index 000000000..95cdbb86b --- /dev/null +++ b/benchmarks/runtime/workloads/keyed-option-writes.ts @@ -0,0 +1,76 @@ +// Class methods whose names are also keys of option records written through +// element accesses: `enabled[name] = true` over a union of literal names and +// `totals["count"] = n`. The writes store booleans and numbers into plain +// records, so the methods keep their native calls. +type Step = "scale" | "offset" | "clamp"; + +const enabled: Record = { scale: false, offset: false, clamp: false }; +function enable(name: Step): void { + enabled[name] = true; +} + +const totals: Record = {}; + +class Transform { + private readonly factor: number; + private readonly shift: number; + constructor(factor: number, shift: number) { + this.factor = factor; + this.shift = shift; + } + scale(x: number): number { + return enabled.scale ? x * this.factor : x; + } + offset(x: number): number { + return enabled.offset ? x + this.shift : x; + } + clamp(x: number): number { + if (!enabled.clamp) return x; + return x < 0 ? 0 : x > 1000 ? 1000 : x; + } +} + +class Counter { + private seen = 0; + private sum = 0; + add(value: number): void { + this.seen++; + this.sum += value; + } + count(): number { + return this.seen; + } + mean(): number { + return this.seen === 0 ? 0 : this.sum / this.seen; + } +} + +const scale = Number(process.argv[2] ?? "1"); +const steps: Step[] = ["scale", "offset", "clamp"]; +for (const step of steps) enable(step); + +const transforms: Transform[] = []; +for (let i = 0; i < 16; i++) transforms.push(new Transform(1 + (i % 5) * 0.25, (i % 7) - 3)); +const counters: Counter[] = []; +for (let i = 0; i < 8; i++) counters.push(new Counter()); + +let seed = 12345; +let checksum = 0; +const rounds = Math.floor(2_000_000 * scale); +for (let i = 0; i < rounds; i++) { + seed = (Math.imul(seed, 1103515245) + 12345) >>> 0; + const t = transforms[seed & 15]!; + const value = t.clamp(t.offset(t.scale((seed >>> 8) % 1200))); + const counter = counters[(seed >>> 4) & 7]!; + counter.add(value); + checksum += value; + if ((i & 65535) === 0) totals["count"] = counter.count(); +} + +let mean = 0; +let seen = 0; +for (const counter of counters) { + mean += counter.mean(); + seen += counter.count(); +} +console.log("seen", seen, "count", totals["count"], "mean", (mean / counters.length).toFixed(4), "checksum", checksum); diff --git a/docs/content/docs/how-it-works.mdx b/docs/content/docs/how-it-works.mdx index e017267d7..a6520b46e 100644 --- a/docs/content/docs/how-it-works.mdx +++ b/docs/content/docs/how-it-works.mdx @@ -58,7 +58,7 @@ The LLVM helper divides an executable's LLVM IR into partitions by its size. Up -The count depends only on the size of the LLVM text, so a program has the same partitions, and builds the same executable, on every host. A function's partition follows from its name, and a small internal function goes with the one function that calls it, so an edit leaves the other functions in their partitions. +The count depends only on the size of the LLVM text, so a program has the same partitions, and builds the same executable, on every host. In programs with 32 partitions or more, functions from the same source file are placed in the same partition, so that their calls to each other stay inside it; a file larger than 1 MiB of LLVM text or than an average partition is divided into parts by function name. Other functions, and the functions of smaller programs, are placed by their name, and a small internal function goes with the one function that calls it. An edit leaves almost every other function in its partition. These builds are not divided into partitions and compile on one thread: `--emit=asm`, `--emit=obj`, and `--print=native-link-info` output, library builds (`--lib`; `dev` library builds are compiled in pieces, one after another), and `--sanitize` builds, which compile the LLVM IR with the C compiler. diff --git a/docs/content/limitations/builtin-apis/url.mdx b/docs/content/limitations/builtin-apis/url.mdx index c9eee85b3..dddf2a8f7 100644 --- a/docs/content/limitations/builtin-apis/url.mdx +++ b/docs/content/limitations/builtin-apis/url.mdx @@ -1,5 +1,4 @@ -- **URL:** Hosts of special URLs that hold non-ASCII text, or percent-escapes that decode to it, throw `ERR_INVALID_URL`, and `xn--` labels are kept without Punycode validation. - Absolute inputs and relative inputs with runtime string or URL bases, including optional bases, compile. +- **URL:** Absolute inputs and relative inputs with runtime string or URL bases, including optional bases, compile. `URL.canParse`, `URL.parse`, `toString()`, and `toJSON()` compile; the factories use the same parser as construction, returning `false` or `null` for rejected inputs while preserving conversion errors. `href`, `protocol`, `username`, `password`, `host`, `hostname`, `port`, `pathname`, `search`, and `hash` support assignment in native code. Invalid component writes are ignored except for invalid `href` values, which throw; credentials are percent-encoded and ports accept valid leading digits. diff --git a/native/llvm-codegen/src/partition.cpp b/native/llvm-codegen/src/partition.cpp index f4add0e6d..93b9af4a3 100644 --- a/native/llvm-codegen/src/partition.cpp +++ b/native/llvm-codegen/src/partition.cpp @@ -11,6 +11,7 @@ #include #include +#include using namespace llvm; @@ -36,6 +37,9 @@ struct Symbol { std::string PromotedHeader; std::vector Refs; std::vector Metadata; + // The source-module group the emitter wrote before this definition, or + // empty. + StringRef Group; unsigned Partition = NoPartition; }; @@ -70,6 +74,83 @@ uint32_t stableHash(StringRef Name) { return Hash; } +// Places the functions whose root carries a source-module group by group: a +// module's functions call each other most, so keeping them together leaves +// fewer calls for the thin link to import across partitions. A group larger +// than the mean partition, or than MaxChunkBytes, is divided into a power of +// two of chunks by the hash of each root's name, so the division changes only +// when the group's size doubles. Chunks are placed largest first, each in the +// lighter of two partitions chosen by hashes of its key, after the functions +// placed by name. The placement depends only on the module text and the +// partition count. Modules with fewer than MinGroupedPartitions partitions +// keep the placement by name: with few partitions per core, the balance of +// the largest partitions decides the build time more than the imports. +void placeGroups(std::vector &Symbols, ArrayRef Roots, + unsigned Count) { + constexpr unsigned MinGroupedPartitions = 32; + if (Count < MinGroupedPartitions) + return; + auto Bytes = [](const Symbol &S) -> uint64_t { + return S.Header.size() + S.Body.size(); + }; + std::vector Load(Count, 0); + uint64_t Total = 0; + StringMap GroupBytes; + for (unsigned I = 0; I < Symbols.size(); ++I) { + const Symbol &S = Symbols[I]; + if (S.Kind != SymbolKind::Function) + continue; + Total += Bytes(S); + StringRef Group = Symbols[Roots[I]].Group; + if (Group.empty()) + Load[S.Partition] += Bytes(S); + else + GroupBytes[Group] += Bytes(S); + } + if (GroupBytes.empty()) + return; + // A chunk is at most MaxChunkBytes, so the code a partition optimizes + // together, and the inlining within it, does not grow when a program has + // fewer partitions. + constexpr uint64_t MaxChunkBytes = 1024 * 1024; + uint64_t Chunk = + std::min(MaxChunkBytes, std::max(1, Total / Count)); + std::vector Keys(Symbols.size()); + StringMap KeyBytes; + for (unsigned I = 0; I < Symbols.size(); ++I) { + const Symbol &S = Symbols[I]; + StringRef Group = Symbols[Roots[I]].Group; + if (S.Kind != SymbolKind::Function || Group.empty()) + continue; + uint64_t Chunks = 1; + while (Chunks * Chunk < GroupBytes[Group]) + Chunks *= 2; + Keys[I] = Chunks == 1 + ? Group.str() + : (Group + "#" + + Twine(stableHash(Symbols[Roots[I]].Name) % Chunks)) + .str(); + KeyBytes[Keys[I]] += Bytes(S); + } + std::vector> Order; + for (const auto &Entry : KeyBytes) + Order.push_back({Entry.getKey(), Entry.getValue()}); + llvm::sort(Order, [](const auto &A, const auto &B) { + return A.second != B.second ? A.second > B.second : A.first < B.first; + }); + StringMap Where; + for (const auto &[Key, Size] : Order) { + unsigned First = stableHash(Key) % Count; + unsigned Second = stableHash((Key + "/2").str()) % Count; + unsigned Chosen = Load[Second] < Load[First] ? Second : First; + Load[Chosen] += Size; + Where[Key] = Chosen; + } + for (unsigned I = 0; I < Symbols.size(); ++I) + if (!Keys[I].empty()) + Symbols[I].Partition = Where[Keys[I]]; +} + bool startsWithLinkage(StringRef Rest) { for (StringRef Linkage : {"linkonce", "weak", "common", "appending", "extern_weak", @@ -326,6 +407,11 @@ std::optional partitionProgramText(StringRef Text, return true; }; + // The emitter writes "; scriptc-group " before the definitions of each + // source function; the group applies until the next such line, and an + // empty key ends it. + constexpr StringRef GroupLine = "; scriptc-group"; + StringRef Group; size_t Pos = 0; while (Pos < Text.size()) { size_t End = Text.find('\n', Pos); @@ -335,10 +421,15 @@ std::optional partitionProgramText(StringRef Text, size_t Next = End + 1; if (Line.empty() || Line.trim().empty()) { // Blank separator. + } else if (Line.starts_with(GroupLine) && + (Line.size() == GroupLine.size() || + Line[GroupLine.size()] == ' ')) { + Group = Line.drop_front(GroupLine.size()).trim(); } else if (Line.starts_with("define ")) { Symbol S; S.Kind = SymbolKind::Function; S.Header = Line; + S.Group = Group; if (!parseFunctionHeader(Line, S)) return std::nullopt; // The emitter indents every body line, so the first line that is @@ -466,6 +557,7 @@ std::optional partitionProgramText(StringRef Text, Referrer[Ref] = I; } } + std::vector Roots(Symbols.size()); for (unsigned I = 0; I < Symbols.size(); ++I) { if (Symbols[I].Kind != SymbolKind::Function) continue; @@ -479,8 +571,10 @@ std::optional partitionProgramText(StringRef Text, break; Root = Parent; } + Roots[I] = Root; Symbols[I].Partition = stableHash(Symbols[Root].Name) % Count; } + placeGroups(Symbols, Roots, Count); for (const Symbol &S : Symbols) { if (S.Kind != SymbolKind::Function) continue; diff --git a/native/llvm-codegen/src/partition.h b/native/llvm-codegen/src/partition.h index 838919443..35ef4c543 100644 --- a/native/llvm-codegen/src/partition.h +++ b/native/llvm-codegen/src/partition.h @@ -20,11 +20,15 @@ struct ProgramPartitions { }; // Divides the LLVM assembly scriptc emits into Count modules without parsing -// it, so the partitions can be parsed concurrently. Functions are placed by a -// stable hash of their name; an internal function referenced by exactly one -// other function stays with that function unless its body is too large to -// inline, and a global stays with the first function that references it. Returns nullopt for any construct outside the -// emitter's narrow top-level grammar; callers then split a parsed module. +// it, so the partitions can be parsed concurrently. In modules of 32 or +// more partitions, functions under a "; scriptc-group " line are placed +// with the other functions of that source-module group, which is divided +// into chunks when it is larger than the mean partition or 1 MiB; other +// functions are placed by a stable hash of their name. An internal function referenced by exactly one other function stays +// with that function unless its body is too large to inline, and a global +// stays with the first function that references it. Returns nullopt for any +// construct outside the emitter's narrow top-level grammar; callers then +// split a parsed module. std::optional partitionProgramText(llvm::StringRef Text, unsigned Count); diff --git a/packages/compiler/native/cli.profdata.gz b/packages/compiler/native/cli.profdata.gz index b5f1e079f..31423fa04 100644 Binary files a/packages/compiler/native/cli.profdata.gz and b/packages/compiler/native/cli.profdata.gz differ diff --git a/packages/compiler/src/backend/llvm/dyn.ts b/packages/compiler/src/backend/llvm/dyn.ts index 6d02daa8b..da31ffdbb 100644 --- a/packages/compiler/src/backend/llvm/dyn.ts +++ b/packages/compiler/src/backend/llvm/dyn.ts @@ -193,6 +193,11 @@ function builtinStreamKind(className: string): number | null { return kinds[className] ?? null; } +/** Unions whose discriminant has fewer string literals keep one literal test + * per arm (dispatchValues): one runtime lookup costs about what a few + * literal tests do. */ +const DISCRIMINANT_DISPATCH_VALUES = 4; + export class LlDyn { private readonly dynMatchers = new Map(); private readonly dynBuilders = new Map(); @@ -413,6 +418,109 @@ export class LlDyn { return order; } + /** The string literals of `def`'s discriminant when its arms are chosen + * by reading the discriminant once (unionArmsByDiscriminant), or null + * when the union keeps one literal test per arm: some literal is not a + * string, or there are fewer than DISCRIMINANT_DISPATCH_VALUES. */ + private dispatchValues(def: IrUnionDef): string[] | null { + const values: string[] = []; + for (const guard of def.discriminant?.cases ?? []) { + if (guard.values.length === 0) return null; + for (const value of guard.values) { + if (typeof value !== "string") return null; + if (!values.includes(value)) values.push(value); + } + } + return values.length < DISCRIMINANT_DISPATCH_VALUES ? null : values; + } + + /** Branches to `hits[tag]` for the first arm in `order` that `%d` matches + * (unionArmMatch), or to `none` when no arm does, for + * a union whose discriminant literals are `values` (dispatchValues). The + * discriminant is read once (scr_dyn_discriminant_case, against the + * constant `table`), and only the arms whose literal it holds and the arms + * without a literal are tested, in `order`: every other arm's literal test + * fails. An arm's failed check can run a getter that changes the + * discriminant, so it is read again after each one; when its case + * changed, the tests continue arm by arm after the failed arm, as they + * would have run from the first arm. A value the runtime cannot classify + * that way (a typed capsule, an accessor, an inherited discriminant) takes + * the tests arm by arm. */ + private unionArmsByDiscriminant( + B: BlockBuilder, + def: IrUnionDef, + values: readonly string[], + order: readonly number[], + hits: readonly string[], + none: string, + preserveRefs: boolean, + table: string, + ): void { + const discriminant = def.discriminant!; + const literalsOf = (tag: number): readonly (string | number | boolean)[] | null => + discriminant.cases.find((candidate) => candidate.tag === tag)?.values ?? null; + const host = this.host; + this.defs.push( + `${table} = private unnamed_addr constant [${values.length} x ptr] [${values.map((value) => `ptr ${host.internLiteral(value)}`).join(", ")}]`, + ``, + ); + host.declare(`declare i32 @scr_dyn_discriminant_case(ptr, ptr, ptr, ${host.sizeType})`); + const classify = (): string => { + const index = B.tmp(); + B.line( + `${index} = call i32 @scr_dyn_discriminant_case(ptr %d, ptr ${host.internLiteral(discriminant.field)}, ptr ${table}, ${host.sizeType} ${values.length})`, + ); + return index; + }; + // The arm-by-arm tests, entered at any arm (`from[i]` tests order[i] + // and the arms after it). + const tests = B.newLabel("du.tests"); + const from = order.map((_, i) => (i === 0 ? tests : B.newLabel("du.t"))); + const after = (position: number): string => + position + 1 < order.length ? from[position + 1]! : none; + // Tests the arms a value in case `k` (-1: no literal) can match. + const tryCase = (k: number): void => { + const value = k < 0 ? null : values[k]!; + const positions = order + .map((_, position) => position) + .filter((position) => { + const literals = literalsOf(order[position]!); + return literals === null || (value !== null && literals.includes(value)); + }); + if (positions.length === 0) { + B.br(none); + return; + } + positions.forEach((position, index) => { + const tag = order[position]!; + const failed = B.newLabel("du.c.f"); + this.unionArmStructure(B, def, tag, hits[tag]!, failed, preserveRefs); + B.startBlock(failed); + const next = index === positions.length - 1 ? none : B.newLabel("du.c.n"); + const same = B.tmp(); + B.line(`${same} = icmp eq i32 ${classify()}, ${k}`); + B.condBr(same, next, after(position)); + if (next !== none) B.startBlock(next); + }); + }; + const index = classify(); + const unmatched = B.newLabel("du.none"); + const cases = values.map(() => B.newLabel("du.case")); + B.terminate( + `switch i32 ${index}, label %${tests} [ i32 -1, label %${unmatched} ${cases.map((label, k) => `i32 ${k}, label %${label}`).join(" ")} ]`, + ); + values.forEach((_, k) => { + B.startBlock(cases[k]!); + tryCase(k); + }); + B.startBlock(unmatched); + tryCase(-1); + order.forEach((tag, position) => { + B.startBlock(from[position]!); + this.unionArmMatch(B, def, tag, hits[tag]!, after(position), preserveRefs); + }); + } + /** Branch to a variant only after its literal discriminator and complete * structure match. Use the same path for predicates and builders. */ private unionArmMatch( @@ -450,6 +558,18 @@ export class LlDyn { }); B.startBlock(structure); } + this.unionArmStructure(B, def, tag, yes, no, preserveRefs); + } + + /** The part of unionArmMatch past the arm's literal test. */ + private unionArmStructure( + B: BlockBuilder, + def: IrUnionDef, + tag: number, + yes: string, + no: string, + preserveRefs: boolean, + ): void { const matches = B.tmp(); B.line( `${matches} = call zeroext i1 @${this.dynMatchHelper(def.arms[tag]!, preserveRefs)}(ptr %d)`, @@ -1278,11 +1398,26 @@ export class LlDyn { ); // Arms in canonical order; any full match answers true. const yes = B.newLabel("dm.y"); - for (let tag = 0; tag < def.arms.length; tag++) { - const ln = B.newLabel("dm.n"); - this.unionArmMatch(B, def, tag, yes, ln, preserveRefs); - B.startBlock(ln); - } + const values = this.dispatchValues(def); + if (values !== null) { + const no = B.newLabel("dm.n"); + this.unionArmsByDiscriminant( + B, + def, + values, + def.arms.map((_, tag) => tag), + def.arms.map(() => yes), + no, + preserveRefs, + `@${name}.cases`, + ); + B.startBlock(no); + } else + for (let tag = 0; tag < def.arms.length; tag++) { + const ln = B.newLabel("dm.n"); + this.unionArmMatch(B, def, tag, yes, ln, preserveRefs); + B.startBlock(ln); + } B.terminate(`ret i1 false`); B.startBlock(yes); B.terminate(`ret i1 true`); @@ -2462,12 +2597,8 @@ export class LlDyn { ); // Arms in unionArmOrder, first FULL match wins. The matched arm's // builder can no longer fail. - for (const i of this.unionArmOrder(def)) { + const build = (i: number): void => { const arm = def.arms[i]!; - const lHit = B.newLabel("dcu.h"); - const lNext = B.newLabel("dcu.n"); - this.unionArmMatch(B, def, i, lHit, lNext, preserveRefs); - B.startBlock(lHit); if (arm.kind === "undefinedT" || arm.kind === "nullT") { // A matched unit arm builds nothing: THE interned immortal // instance (rc == SIZE_MAX — no retain owed). @@ -2506,8 +2637,36 @@ export class LlDyn { ); B.terminate(`ret ptr ${u}`); } - B.startBlock(lNext); - } + }; + const order = this.unionArmOrder(def); + const values = this.dispatchValues(def); + if (values !== null) { + const hits = def.arms.map(() => B.newLabel("dcu.h")); + const unmatched = B.newLabel("dcu.n"); + this.unionArmsByDiscriminant( + B, + def, + values, + order, + hits, + unmatched, + preserveRefs, + `@${name}.cases`, + ); + for (const i of order) { + B.startBlock(hits[i]!); + build(i); + } + B.startBlock(unmatched); + } else + for (const i of order) { + const lHit = B.newLabel("dcu.h"); + const lNext = B.newLabel("dcu.n"); + this.unionArmMatch(B, def, i, lHit, lNext, preserveRefs); + B.startBlock(lHit); + build(i); + B.startBlock(lNext); + } if (preserveRefs) { B.line( `call void @scr_dyn_check_fail(ptr %path, ptr ${host.cstr("reference with its original layout")}, ptr %d)`, diff --git a/packages/compiler/src/backend/llvm/emit-function.ts b/packages/compiler/src/backend/llvm/emit-function.ts index 8f306035a..e015c2d72 100644 --- a/packages/compiler/src/backend/llvm/emit-function.ts +++ b/packages/compiler/src/backend/llvm/emit-function.ts @@ -159,6 +159,7 @@ export function emitFunction(emitter: LlEmitter, fn: IrFunction): string { emitter.optionalArrayReads, emitter.referenceWrites, emitter.fnByName, + emitter.nodeIndex, ) : new Set(); emitter.localArrayReads = findLocalArrayReads( @@ -324,7 +325,7 @@ export function emitFunction(emitter: LlEmitter, fn: IrFunction): string { const paramIds = new Set(fn.params.map((p) => p.localId)); if (emitter.debug === null) { - for (const id of findIntegerViews(numericFn)) { + for (const id of findIntegerViews(numericFn, emitter.nodeIndex)) { // An i32 slot is its own view. if (emitter.int32Locals.has(id)) continue; const slot = B.slot(); @@ -1307,8 +1308,10 @@ export function emitArgPackAndTrampolinePrologue( * code. */ export function parkingFunctions(emitter: LlEmitter): Set | null { if (emitter.wasi || emitter.mod.workers === true) return null; - for (const decl of emitter.decls) if (/@scr_test_sub\b/.test(decl)) return null; - for (const text of emitter.fnDefText.values()) if (/@scr_test_sub\b/.test(text)) return null; + const testSub = (text: string): boolean => + text.includes("@scr_test_sub") && /@scr_test_sub\b/.test(text); + for (const decl of emitter.decls) if (testSub(decl)) return null; + for (const text of emitter.fnDefText.values()) if (testSub(text)) return null; // Reverse mention edges, then propagate "may park" to every mentioner. const mentionedBy = new Map>(); const parks = new Set(); diff --git a/packages/compiler/src/backend/llvm/emit-module.ts b/packages/compiler/src/backend/llvm/emit-module.ts index 5aa18bf9f..5b3389155 100644 --- a/packages/compiler/src/backend/llvm/emit-module.ts +++ b/packages/compiler/src/backend/llvm/emit-module.ts @@ -65,6 +65,7 @@ import { STRING_READ_HELPERS, stringScanDecl } from "./string-reads.js"; import { emitLibraryIdentityLines } from "../library-identity-markers.js"; import { InternalCompilerError } from "../../errors.js"; import { targetCpuAttribute } from "../target-cpu.js"; +import type { IrFunction } from "../../ir/ir.js"; /** The function bodies emitted in `chunks` contiguous ranges, each by its * own emitter (LlEmitter.chunkEmitter), then merged in function order: the @@ -206,6 +207,26 @@ function functionAttributes(emitter: LlEmitter, hasNoInlineRecordClone: boolean) ]; } +/** The group line written before each function's text. Functions from one + * source file share a group, which the LLVM helper keeps in the same + * partitions so that their calls to each other stay inside a partition + * (native/llvm-codegen/src/partition.h). A group is numbered by the first + * function of its file in the module's function order, which does not depend + * on the host. Functions without a source file get no group. */ +function partitionGroupLines(functions: readonly IrFunction[]): string[] { + const groups = new Map(); + return functions.map((fn) => { + const file = fn.loc.file; + if (file === "") return PARTITION_GROUP_END; + let group = groups.get(file); + if (group === undefined) groups.set(file, (group = groups.size)); + return `${PARTITION_GROUP_END} ${group}`; + }); +} + +/** Ends the group of the definitions that follow. */ +const PARTITION_GROUP_END = "; scriptc-group"; + export function emitModuleParts(emitter: LlEmitter): string[] { // Function bodies first (the literal/unit/fn-value tables fill as they // emit), then the file assembles around them — the runtime ABI’s order. @@ -344,6 +365,8 @@ export function emitModuleParts(emitter: LlEmitter): string[] { // Fetch-referencing programs register native fetch before main. // Runtime selection compiles scr_fetch.c on the same predicate. const usesFetch = runtimeFeatures.fetch; + // URL programs give the host parser the IDNA unit's ToASCII. + const usesUrlIdna = runtimeFeatures.urlIdna; const snapshotsTlsCa = runtimeFeatures.tls || runtimeFeatures.tlsCa; // The process verdict has the same precedence as the C reference // emitter: node:test owns the final status when present; otherwise an @@ -396,6 +419,7 @@ export function emitModuleParts(emitter: LlEmitter): string[] { if (usesHttp2) emitter.declare(`declare void @scr_http2_dyn_install()`); if (usesHttp) emitter.declare(`declare void @scr_http_dyn_install()`); if (usesFetch) emitter.declare(`declare void @scr_fetch_install()`); + if (usesUrlIdna) emitter.declare(`declare void @scr_url_idna_install()`); if (usesNodeTest) { emitter.declare(`declare i32 @scr_test_exit_code()`); @@ -912,9 +936,13 @@ export function emitModuleParts(emitter: LlEmitter): string[] { for (const line of wrappers) out.push(line); for (const line of asyncDefs) out.push(line); for (const line of emitter.resolveThunkDefs) out.push(line); - out.push(fnDefs[0] ?? ""); - for (let i = 1; i < fnDefs.length; i++) out.push("", fnDefs[i]!); - out.push(""); + const groups = partitionGroupLines(emitter.mod.functions); + for (let i = 0; i < fnDefs.length; i++) { + if (i > 0) out.push(""); + out.push(groups[i] ?? PARTITION_GROUP_END, fnDefs[i]!); + } + if (fnDefs.length === 0) out.push(""); + out.push(PARTITION_GROUP_END, ""); // WebAssembly modules translated into this program, and the registry. for (const line of emitWasmModules(emitter.mod, flushedDecls, emitter.wasmBoundsChecks)) out.push(line); @@ -1027,6 +1055,7 @@ export function emitModuleParts(emitter: LlEmitter): string[] { ...(emitter.ffiHasForeignCallback ? [` call void @scr_ffi_install()`] : []), ...(snapshotsTlsCa ? [` call void @scr_tls_ca_install()`] : []), ...(usesFetch ? [` call void @scr_fetch_install()`] : []), + ...(usesUrlIdna ? [` call void @scr_url_idna_install()`] : []), ...(usesNet ? [` call void @scr_net_install()`, ` call void @scr_net_dyn_install()`] : []), ...(usesNetLookup ? [` call void @scr_net_lookup_install()`] : []), ...(usesHttp ? [` call void @scr_http_dyn_install()`] : []), diff --git a/packages/compiler/src/backend/llvm/emit-statements.ts b/packages/compiler/src/backend/llvm/emit-statements.ts index 55e31219c..d356bfa3d 100644 --- a/packages/compiler/src/backend/llvm/emit-statements.ts +++ b/packages/compiler/src/backend/llvm/emit-statements.ts @@ -2437,7 +2437,7 @@ function planProvenLoop( loop: Extract, ): ProvenLoopPlan | null { if ( - loop.kind !== "for" || + loop.kind === "doWhile" || !emitter.speculativeLoopsEnabled || emitter.speculation !== null || emitter.speculationResume !== null || diff --git a/packages/compiler/src/backend/llvm/emitter.ts b/packages/compiler/src/backend/llvm/emitter.ts index 6760f3b74..6729da197 100644 --- a/packages/compiler/src/backend/llvm/emitter.ts +++ b/packages/compiler/src/backend/llvm/emitter.ts @@ -97,7 +97,12 @@ import { unshareNodes } from "../../ir/unshare.js"; import { runPairedWhileLoopsAsFor } from "../../ir/paired-while-loops.js"; import { copyStableCaptures } from "../../ir/capture-values.js"; import { resolveSelfCaptures } from "../../ir/self-captures.js"; -import { analyzeInt32Slots, dynamicFieldStores, type Int32Slots } from "../../ir/int32-slots.js"; +import { + analyzeInt32Slots, + dynamicFieldStores, + type DynamicFieldStores, + type Int32Slots, +} from "../../ir/int32-slots.js"; import { analyzeFinalFields, type FinalFields } from "../../ir/final-fields.js"; import { foldConstantGlobals } from "../../ir/constant-globals.js"; import { dropUnwrittenPropertyBags } from "../../ir/property-bags.js"; @@ -171,7 +176,7 @@ import { type MapEntryFusion } from "./map-entry-fusion.js"; import { emitBorrowedFieldSequence, guardedValue } from "./borrowed-receivers.js"; import { emitBorrowedInputs } from "./borrowed-inputs.js"; import { findWalkBorrows, type WalkBorrowHost } from "./walk-borrows.js"; -import { IrNodeIndex } from "../../ir/node-index.js"; +import { eachNodeOfKinds, IrNodeIndex, type IrNodeKind } from "../../ir/node-index.js"; import { lastUseRead } from "./local-extents.js"; import { findReboundParameters } from "./rebound-parameters.js"; import { ReferenceEffects } from "./reference-effects.js"; @@ -506,14 +511,24 @@ function versionParameters(mod: IrModule): IrModule { return versionLoopBoundParameters(mod, mangleFunction); } -/** The int32 arrays the int32 slot analysis keeps in `mod` (proven and - * checked), under the inputs LlEmitter gives it (selectInt32Arrays). */ -function provenInt32Arrays(mod: IrModule): ReadonlySet { - const slots = analyzeInt32Slots(mod, { - dynamicStores: dynamicFieldStores(mod), - initializerBindings: findInitializerBindings(mod), - }); - return new Set([...slots.int32Arrays, ...slots.checkedInt32Arrays]); +/** The int32 slot analysis of one module and the inputs it read: the + * constructor's own analyses of that module, which LlEmitter takes over + * instead of repeating them when it emits the same module. */ +export interface Int32SlotAnalysis { + readonly mod: IrModule; + readonly nodeIndex: IrNodeIndex; + readonly initializerBindings: ReturnType; + readonly dynamicStores: DynamicFieldStores; + readonly slots: Int32Slots; +} + +/** The int32 slot analysis of `mod`, with the inputs LlEmitter gives it. */ +function analyzeModuleInt32Slots(mod: IrModule): Int32SlotAnalysis { + const nodeIndex = new IrNodeIndex(mod); + const initializerBindings = findInitializerBindings(mod, nodeIndex); + const dynamicStores = dynamicFieldStores(mod, undefined, undefined, nodeIndex); + const slots = analyzeInt32Slots(mod, { dynamicStores, initializerBindings, nodes: nodeIndex }); + return { mod, nodeIndex, initializerBindings, dynamicStores, slots }; } /** The functions whose calls the backend may emit as calls of their @@ -530,8 +545,16 @@ function moduleEmitter(input: IrModule, options: LlvmTargetOptions): LlEmitter { // Facts are keyed by node identity: emit the same module whether it was // lowered in memory or read back from its serialized form. const mod = unshareNodes(input); + // Each round of selectInt32Arrays analyzes the module it marks; the + // emitter of the module it returns reuses that round's analysis. + const rounds: { last: Int32SlotAnalysis | null } = { last: null }; + const proven = (marked: IrModule): ReadonlySet => { + const analysis = analyzeModuleInt32Slots(marked); + rounds.last = analysis; + return new Set([...analysis.slots.int32Arrays, ...analysis.slots.checkedInt32Arrays]); + }; // Keep source storage intact for debugger inspection in dev builds. - return new LlEmitter( + const emitted = options.debugSources === undefined ? selectInt32Arrays( versionParameters( @@ -559,12 +582,12 @@ function moduleEmitter(input: IrModule, options: LlvmTargetOptions): LlEmitter { ), ), ), - provenInt32Arrays, + proven, resultEntryTargets, ) - : mod, - options, - ); + : mod; + const prepared = rounds.last; + return new LlEmitter(emitted, options, null, prepared?.mod === emitted ? prepared : null); } export function emitLlvmModule(input: IrModule, options: LlvmTargetOptions = {}): string { @@ -598,6 +621,21 @@ export function llStrBytes(text: string): string { const NO_BORROWED: ReadonlySet = new Set(); +/** The nodes that write a binding by its id (numberGlobalWriters). */ +const GLOBAL_WRITE_KINDS: readonly IrNodeKind[] = ["assign", "varDecl", "assignExpr", "incDec"]; +/** Runtime library calls (readsLoopMetrics). */ +const LIB_CALL_KINDS: readonly IrNodeKind[] = ["libCall"]; + +/** [A-Za-z0-9_] */ +function isIdentifierByte(code: number): boolean { + return ( + (code >= 48 && code <= 57) || + (code >= 65 && code <= 90) || + (code >= 97 && code <= 122) || + code === 95 + ); +} + /** Runtime entry points (and await-call entries) that may park the running * execution context, as they appear in emitted function text. */ export const SUSPENDS = @@ -1562,6 +1600,7 @@ export class LlEmitter { readonly mod: IrModule, options: LlvmTargetOptions, journal: EmitJournal | null = null, + int32Analysis: Int32SlotAnalysis | null = null, ) { this.targetOptions = options; this.journal = journal; @@ -1572,7 +1611,7 @@ export class LlEmitter { options.debugSources === undefined); // One walk of every body, which the analyses below read instead of // walking the bodies themselves. - this.nodeIndex = new IrNodeIndex(mod); + this.nodeIndex = int32Analysis?.nodeIndex ?? new IrNodeIndex(mod); this.stackTraces = observesStacks(mod, this.nodeIndex); this.nullableUnions = new NullableUnions( mod, @@ -1591,14 +1630,18 @@ export class LlEmitter { this.nullableUnions.armTypes(), ); this.constantNumericTables = findConstantNumericTables(mod); - this.initializerBindings = findInitializerBindings(mod, this.nodeIndex); + this.initializerBindings = + int32Analysis?.initializerBindings ?? findInitializerBindings(mod, this.nodeIndex); // One walk finds the dynamic field stores every analysis below consults. - const dynamicStores = dynamicFieldStores(mod, undefined, undefined, this.nodeIndex); - this.int32Slots = analyzeInt32Slots(mod, { - dynamicStores, - initializerBindings: this.initializerBindings, - nodes: this.nodeIndex, - }); + const dynamicStores = + int32Analysis?.dynamicStores ?? dynamicFieldStores(mod, undefined, undefined, this.nodeIndex); + this.int32Slots = + int32Analysis?.slots ?? + analyzeInt32Slots(mod, { + dynamicStores, + initializerBindings: this.initializerBindings, + nodes: this.nodeIndex, + }); this.liveViewsReadOnly = mod.lib === undefined && !dynamicStores.all && dynamicStores.names.size === 0; this.int32Abis = this.debug === null ? findInt32Abis(mod, this.int32Slots) : new Map(); @@ -1804,9 +1847,9 @@ export class LlEmitter { ) : NO_INTEGER_CLONES; this.stackCallbacks = new StackCallbacks(this.fnByName); - this.receiverReaders = new AmbientReceiverReaders(this.fnByName); + this.receiverReaders = new AmbientReceiverReaders(this.fnByName, this.nodeIndex); this.stackCaptures = new StackCaptures(this.fnByName, this.callLifetimes, this.stackCallbacks); - this.lazyCaptures = new LazyCaptures(this.fnByName); + this.lazyCaptures = new LazyCaptures(this.fnByName, this.nodeIndex); const traced = computeTraced(mod, dynamicStores, cycleMutability); this.tracedShapes = traced.shapes; this.tracedUnions = traced.unions; @@ -2235,9 +2278,23 @@ export class LlEmitter { this.mangledFunctions.set(mangleBorrowedFunction(fn.name), fn.name); } } - for (const match of text.matchAll(/@(sc_b?f_[A-Za-z0-9_]+)/g)) { - const callee = this.mangledFunctions.get(match[1]!); + // The names `@(sc_b?f_[A-Za-z0-9_]+)` matches, found with indexOf. + for (let at = text.indexOf("@sc_"); at >= 0;) { + let start = at + 4; + if (text.charCodeAt(start) === 98 /* b */) start++; + if (text.charCodeAt(start) !== 102 /* f */ || text.charCodeAt(start + 1) !== 95 /* _ */) { + at = text.indexOf("@sc_", at + 1); + continue; + } + let end = start + 2; + while (end < text.length && isIdentifierByte(text.charCodeAt(end))) end++; + if (end === start + 2) { + at = text.indexOf("@sc_", at + 1); + continue; + } + const callee = this.mangledFunctions.get(text.slice(at + 1, end)); if (callee !== undefined) yield callee; + at = text.indexOf("@sc_", end); } } @@ -3569,15 +3626,14 @@ export class LlEmitter { set.add(fn.name); writers.set(localId, set); }; - everyStmtList(fn.body, { - stmt: (s) => { - if (s.kind === "assign" || s.kind === "varDecl") note(s.localId); - return true; - }, - expr: (e) => { - if (e.kind === "assignExpr" || e.kind === "incDec") note(e.localId); - return true; - }, + eachNodeOfKinds(fn, this.nodeIndex.scan(GLOBAL_WRITE_KINDS), (node) => { + if ( + node.kind === "assign" || + node.kind === "varDecl" || + node.kind === "assignExpr" || + node.kind === "incDec" + ) + note(node.localId); }); } this.globalWriters = writers; @@ -5357,13 +5413,21 @@ export class LlEmitter { private loopMetricsRead: boolean | undefined; readsLoopMetrics(): boolean { if (this.loopMetricsRead !== undefined) return this.loopMetricsRead; - const visitor = { - expr: (e: IrExpr): boolean => - !(e.kind === "libCall" && (e.fn === "perf.loopStart" || e.fn === "perf.idleTime")), - stmt: (): boolean => true, - }; - this.loopMetricsRead = !this.mod.functions.every((fn) => everyStmtList(fn.body, visitor)); - return this.loopMetricsRead; + let found = false; + const scan = this.nodeIndex.scan(LIB_CALL_KINDS); + for (const fn of this.mod.functions) { + eachNodeOfKinds(fn, scan, (node) => { + if ( + node.kind === "libCall" && + (node.fn === "perf.loopStart" || node.fn === "perf.idleTime") + ) + found = true; + return !found; + }); + if (found) break; + } + this.loopMetricsRead = found; + return found; } private lastUseMovesAllowed(): boolean { @@ -5731,6 +5795,10 @@ export class LlEmitter { }); return; } + if (e.kind === "dynInvoke" && e.calleeNameValue === undefined) { + emitDynamicExpr(this, e, true); + return; + } this.emitExpr(e); } diff --git a/packages/compiler/src/backend/llvm/expr-dynamic.ts b/packages/compiler/src/backend/llvm/expr-dynamic.ts index 26905a7ef..b7a5c7736 100644 --- a/packages/compiler/src/backend/llvm/expr-dynamic.ts +++ b/packages/compiler/src/backend/llvm/expr-dynamic.ts @@ -23,6 +23,7 @@ import { isUnitType, typeEquals, typeKey, + VOID, } from "../../ir/ir.js"; import { BYTES_ELEM_NUM, closureIdentityEqual } from "./common.js"; import { DYN_KIND } from "./dyn.js"; @@ -57,6 +58,7 @@ export function emitDynamicExpr( | "caughtNarrow" | "caughtToDyn" >, + discarded = false, ): LlValue { const B = host.B; switch (e.kind) { @@ -257,6 +259,18 @@ export function emitDynamicExpr( // Each call passes its method name's id slot, so only the first call // looks the name up. const invoke = DYN_BUFFER_METHODS.get(e.method) ?? "scr_dyn_invoke_prepared_id"; + // A statement-position push drops the new length, so its entry + // need not box it. + if (discarded && e.method === "push" && invoke === "scr_dyn_invoke_prepared_id") { + host.declare( + `declare void @scr_dyn_invoke_discard_id(ptr, ptr, ptr, ptr, ptr, ${host.sizeType}, ptr)`, + ); + B.line( + `call void @scr_dyn_invoke_discard_id(ptr ${recv.name}, ptr ${prepared}, ptr ${host.cstr(e.method)}, ptr ${host.methodIdSlot(e.method)}, ptr ${argsPtr}, ${host.sizeType} ${args.length}, ptr ${host.cstr(e.calleeName)})`, + ); + host.emitPendingCheck(); + return { name: "", type: VOID }; + } host.declare(`declare ptr @${invoke}(ptr, ptr, ptr, ptr, ptr, ${host.sizeType}, ptr)`); let name = host.cstr(e.calleeName); if (e.calleeNameValue !== undefined) { diff --git a/packages/compiler/src/backend/llvm/integer-views.ts b/packages/compiler/src/backend/llvm/integer-views.ts index c38770160..579f010a2 100644 --- a/packages/compiler/src/backend/llvm/integer-views.ts +++ b/packages/compiler/src/backend/llvm/integer-views.ts @@ -1,6 +1,9 @@ import type { IrExpr, IrFunction, IrStmt } from "../../ir/ir.js"; import { everyExprChild, everyStmtChild } from "../../ir/traverse.js"; import { byteNumberAccess } from "../../ir/byte-numbers.js"; +import { eachNodeOfKinds, type IrNodeIndex, type IrNodeKind } from "../../ir/node-index.js"; + +const LOOP_KINDS: readonly IrNodeKind[] = ["for", "forOf", "while", "doWhile"]; /** Keep a computed ToUint32 view beside an ordinary number when integer * consumers in loops can reuse it. The number remains authoritative for floating @@ -14,9 +17,16 @@ import { byteNumberAccess } from "../../ir/byte-numbers.js"; * compare converts the stored number back, and LLVM moves that number's * conversion into the join of a NaN-coded typed-array read (`phi [sitofp x, * 0.0]`), where nothing folds it back to the integer. */ -export function findIntegerViews(fn: IrFunction): Set { +export function findIntegerViews(fn: IrFunction, nodes?: IrNodeIndex): Set { const wanted = new Set(); if (fn.async || fn.generator) return wanted; + // Every view is wanted inside a loop: a body without one wants none. + let loops = false; + eachNodeOfKinds(fn, nodes?.scan(LOOP_KINDS), (node, statement) => { + if (statement && LOOP_KINDS.includes(node.kind)) loops = true; + return !loops; + }); + if (!loops) return wanted; const initialized = new Set(fn.params.map((p) => p.localId)); const excluded = new Set( [...(fn.captures ?? []), ...(fn.classCaptures ?? [])].map((c) => c.localId), diff --git a/packages/compiler/src/backend/llvm/lazy-captures.ts b/packages/compiler/src/backend/llvm/lazy-captures.ts index 2014869af..85acff2ef 100644 --- a/packages/compiler/src/backend/llvm/lazy-captures.ts +++ b/packages/compiler/src/backend/llvm/lazy-captures.ts @@ -1,5 +1,8 @@ import type { IrExpr, IrFunction, IrStmt } from "../../ir/ir.js"; import { everyStmtList } from "../../ir/traverse.js"; +import { eachNodeOfKinds, type IrNodeIndex, type IrNodeKind } from "../../ir/node-index.js"; + +const CLOSURE_KINDS: readonly IrNodeKind[] = ["closure"]; /** The local a node rebinds. Writes through the bound reference, such as a * field store or an array push, leave the binding unchanged. Catch bindings @@ -37,7 +40,11 @@ export class LazyCaptures { private readonly pending = new Set(); private sites: Map | null = null; - constructor(private readonly functions: ReadonlyMap) {} + constructor( + private readonly functions: ReadonlyMap, + /** The module's node index, read instead of walking the bodies it holds. */ + private readonly nodes?: IrNodeIndex, + ) {} /** The captures of closure body `fnName` that hold one value for the * closure's whole life: at every creation site the captured binding is @@ -103,16 +110,15 @@ export class LazyCaptures { private creationSites(): Map { if (this.sites !== null) return this.sites; const sites = new Map(); + const scan = this.nodes?.scan(CLOSURE_KINDS); for (const fn of this.functions.values()) { - const visit = (node: IrStmt | IrExpr): boolean => { + eachNodeOfKinds(fn, scan, (node) => { if (node.kind === "closure") { let list = sites.get(node.fnName); if (!list) sites.set(node.fnName, (list = [])); list.push({ owner: fn.name, captures: node.captures }); } - return true; - }; - everyStmtList(fn.body, { stmt: visit, expr: visit }); + }); } this.sites = sites; return sites; diff --git a/packages/compiler/src/backend/llvm/loop-array-borrows.ts b/packages/compiler/src/backend/llvm/loop-array-borrows.ts index 9335894d9..cc9e03a5c 100644 --- a/packages/compiler/src/backend/llvm/loop-array-borrows.ts +++ b/packages/compiler/src/backend/llvm/loop-array-borrows.ts @@ -1,5 +1,6 @@ import { F64, isRefCounted, type IrExpr, type IrFunction, type IrStmt } from "../../ir/ir.js"; import { everyStmtList, transformStmtList } from "../../ir/traverse.js"; +import { eachNodeOfKinds, type IrNodeIndex, type IrNodeKind } from "../../ir/node-index.js"; import type { CallLifetimes } from "./call-lifetimes.js"; import type { ReferenceEffects } from "./reference-effects.js"; import type { OptionalArrayReads } from "./local-array-reads.js"; @@ -57,6 +58,8 @@ export function dynHoleRead(value: IrExpr): (IrExpr & { kind: "arrayGet" }) | nu return same(cond.left.arr, then.arr) && same(cond.left.index, then.index) ? then : null; } +const LOOP_KINDS: readonly IrNodeKind[] = ["forOf", "for", "while", "doWhile"]; + /** Limit a borrow to a loop whose complete iteration preserves heap edges. * Initialization and unrelated work elsewhere in the function may mutate. * The array remains owned by its unchanged local or global binding; the @@ -72,6 +75,8 @@ export function findLoopArrayBorrows( optionalReads?: OptionalArrayReads, writes?: ReferenceWrites | null, functions?: ReadonlyMap, + /** The module's node index, read instead of walking for the loops. */ + nodes?: IrNodeIndex, ): ReadonlySet { const result = new Set(); if (fn.async || fn.generator) return result; @@ -88,9 +93,10 @@ export function findLoopArrayBorrows( (!local.boxed && !local.tdz && (bindings?.has(local.id) || parameters.has(local.id))) ); }; - everyStmtList(fn.body, { - expr: () => true, - stmt: (loop) => { + eachNodeOfKinds(fn, nodes?.scan(LOOP_KINDS), (node, statement) => { + if (!statement) return true; + const loop = node as IrStmt; + { if (loop.kind === "forOf") { const local = locals.get(loop.localId); if ( @@ -187,7 +193,7 @@ export function findLoopArrayBorrows( }, }); return true; - }, + } }); return result; } diff --git a/packages/compiler/src/backend/llvm/shapes.test.ts b/packages/compiler/src/backend/llvm/shapes.test.ts index 7af16ab83..61c1a074d 100644 --- a/packages/compiler/src/backend/llvm/shapes.test.ts +++ b/packages/compiler/src/backend/llvm/shapes.test.ts @@ -110,10 +110,9 @@ describe("LLVM runtime RC symbols", () => { expect(releaseSym(host, { kind: "union", unionId: "u0" } as never)).toBe( "@sc_rc_release_union", ); - // Array retains inline; array releases stay calls. const array = { kind: "array", elem: { kind: "string" } } as never; expect(retainSym(host, array)).toBe("@sc_rc_retain_arr"); - expect(releaseSym(host, array)).toBe("@scr_arr_release"); + expect(releaseSym(host, array)).toBe("@sc_rc_release_arr"); expect(retainSym(host, { kind: "dyn" })).toBe("@sc_rc_retain_dyn"); expect(releaseSym(host, { kind: "dyn" })).toBe("@sc_rc_release_dyn"); // Unmirrored families and caught snapshots keep their runtime entry points. @@ -125,6 +124,7 @@ describe("LLVM runtime RC symbols", () => { expect(declarations).toContain("declare void @scr_str_release(ptr)"); expect(declarations).toContain("declare void @scr_cyc_on_release(ptr)"); expect(inlineRcDecls(host).sort()).toEqual([ + "declare void @scr_arr_release(ptr)", "declare void @scr_cyc_on_release(ptr)", "declare void @scr_dyn_release(ptr)", "declare void @scr_dyn_release_survivor(ptr)", @@ -280,6 +280,22 @@ describe("inline release budgets", () => { ); }); + test("each family's budget applies to its own sites in a function that has both", () => { + const strings = body("mixed", INLINE_STR_RELEASES_PER_FUNCTION + 1, "str"); + // One function: string sites over their budget, dyn sites within theirs. + const mixed = strings.replace( + " ret void", + ` call void @sc_rc_release_dyn(ptr %v)\n call void @sc_rc_release_other(ptr %v)\n ret void`, + ); + const other = body("other", 2, "str"); + const limited = limitInlineReleases([mixed, other], "release").join("\n"); + const out = limited.slice(0, limited.indexOf("@other(")); + expect(out).not.toContain("@sc_rc_release_str("); + expect(out).toContain("call void @sc_rc_release_dyn(ptr %v)"); + expect(out).toContain("call void @sc_rc_release_other(ptr %v)"); + expect(limited).toContain(other); + }); + test("a function over the string budget calls the runtime string release at every site", () => { const small = body("small", INLINE_STR_RELEASES_PER_FUNCTION, "str"); const large = body("large", INLINE_STR_RELEASES_PER_FUNCTION + 1, "str"); diff --git a/packages/compiler/src/backend/llvm/shapes.ts b/packages/compiler/src/backend/llvm/shapes.ts index b3f5084d3..a966ff305 100644 --- a/packages/compiler/src/backend/llvm/shapes.ts +++ b/packages/compiler/src/backend/llvm/shapes.ts @@ -550,17 +550,16 @@ interface InlineRcFamily { * scr_dyn_release_survivor for kinds whose candidacy is per value, * including arrays and objects of intact documents (ScrDynRegion). * - * Releases stay plain calls for: - * - arr: array releases are the most numerous release sites and the - * least hot (at most 1% self time across the benchmark suite), and - * inlining them roughly doubled the code growth for no measured speedup. - * Array retains are inlined. */ + * The array release inlines the decrement of an array that stays alive + * and is not headered (`a->cyc`): those are most array releases in + * type-checker workloads, where scr_arr_release's own prologue and checks + * took about half of its time. */ export const INLINE_RC_FAMILIES: Readonly> = { str: { cycle: "none", retainNull: false, release: "scr_str_release" }, bytes: { cycle: "none", retainNull: false, release: "scr_bytes_release" }, regex: { cycle: "none", retainNull: false, release: "scr_regex_release" }, bigint: { cycle: "none", retainNull: true, release: "scr_bigint_release" }, - arr: { cycle: "arr", retainNull: false, release: null }, + arr: { cycle: "arr", retainNull: false, release: "scr_arr_release" }, map: { cycle: "map", retainNull: false, release: "scr_map_release" }, union: { cycle: "always", retainNull: false, release: "scr_union_release" }, closure: { cycle: "always", retainNull: false, release: "scr_closure_release" }, @@ -692,34 +691,32 @@ export function limitInlineReleases( parts: readonly string[], mode: InlineRcMode | undefined, ): string[] { - const dyn = limitInlineReleaseFamily(parts, "dyn", INLINE_DYN_RELEASES_PER_FUNCTION); - return mode === "release" - ? limitInlineReleaseFamily(dyn, "str", INLINE_STR_RELEASES_PER_FUNCTION) - : dyn; -} - -function limitInlineReleaseFamily( - parts: readonly string[], - family: "dyn" | "str", - max: number, -): string[] { - const call = `call void @${inlineRcName(family, "release")}(`; - const runtime = `call void @${INLINE_RC_FAMILIES[family]!.release}(`; + const families: readonly ("dyn" | "str")[] = mode === "release" ? ["dyn", "str"] : ["dyn"]; + const limits = families.map((family) => + family === "dyn" ? INLINE_DYN_RELEASES_PER_FUNCTION : INLINE_STR_RELEASES_PER_FUNCTION, + ); + const calls = families.map((family) => `${inlineRcName(family, "release")}(`); + const runtimes = families.map((family) => `call void @${INLINE_RC_FAMILIES[family]!.release}(`); + const head = "call void @"; + const prefix = `${head}${inlineRcName("", "release")}`; // The function each release site belongs to: the last `define` line - // before it, counted over the joined text. + // before it, counted over the joined text. One walk over each part finds + // the sites of every family. const sites = ( part: string, - each: (offset: number, fn: number) => void, + each: (offset: number, family: number, fn: number) => void, first: number, ): number => { let fn = first; let define = part.startsWith("define ") ? 0 : part.indexOf("\ndefine "); - for (let at = part.indexOf(call); at >= 0; at = part.indexOf(call, at + call.length)) { + for (let at = part.indexOf(prefix); at >= 0; at = part.indexOf(prefix, at + prefix.length)) { + const family = calls.findIndex((call) => part.startsWith(call, at + head.length)); + if (family < 0) continue; while (define >= 0 && define < at) { fn++; define = part.indexOf("\ndefine ", define + 1); } - each(at, fn); + each(at, family, fn); } while (define >= 0) { fn++; @@ -727,29 +724,47 @@ function limitInlineReleaseFamily( } return fn; }; - const counts = new Map(); + const counts = families.map(() => new Map()); + // The functions each part's sites can belong to. + const spans: number[] = []; let fn = 0; - for (const part of parts) fn = sites(part, (_, f) => counts.set(f, (counts.get(f) ?? 0) + 1), fn); - const over = new Set([...counts].filter(([, n]) => n > max).map(([f]) => f)); - if (over.size === 0) return [...parts]; - fn = 0; - return parts.map((part) => { - const cuts: number[] = []; + for (const part of parts) { + spans.push(fn); fn = sites( part, - (at, f) => { - if (over.has(f)) cuts.push(at); - }, + (_, family, f) => counts[family]!.set(f, (counts[family]!.get(f) ?? 0) + 1), fn, ); + spans.push(fn); + } + const over = counts.map( + (count, family) => new Set([...count].filter(([, n]) => n > limits[family]!).map(([f]) => f)), + ); + if (over.every((set) => set.size === 0)) return [...parts]; + const anyOver = new Set(over.flatMap((set) => [...set])); + const touches = (from: number, to: number): boolean => { + for (let f = from; f <= to; f++) if (anyOver.has(f)) return true; + return false; + }; + return parts.map((part, index) => { + const from = spans[2 * index]!; + if (!touches(from, spans[2 * index + 1]!)) return part; + const cuts: { at: number; family: number }[] = []; + sites( + part, + (at, family, f) => { + if (over[family]!.has(f)) cuts.push({ at, family }); + }, + from, + ); if (cuts.length === 0) return part; let out = ""; - let from = 0; - for (const at of cuts) { - out += part.slice(from, at) + runtime; - from = at + call.length; + let next = 0; + for (const cut of cuts) { + out += part.slice(next, cut.at) + runtimes[cut.family]!; + next = cut.at + head.length + calls[cut.family]!.length; } - return out + part.slice(from); + return out + part.slice(next); }); } diff --git a/packages/compiler/src/backend/may-throw.ts b/packages/compiler/src/backend/may-throw.ts index 0cfab9244..693d20b3b 100644 --- a/packages/compiler/src/backend/may-throw.ts +++ b/packages/compiler/src/backend/may-throw.ts @@ -93,7 +93,7 @@ export function computeMayThrow( const classes = new Map((mod.classes ?? []).map((cls) => [cls.name, cls])); const constructorClasses = new Map(); const classValueCallers = new Map>(); - const tdzGlobals = (mod.globals ?? []).filter((g) => g.tdz).map((g) => g.id); + const tdzGlobals = new Set((mod.globals ?? []).filter((g) => g.tdz).map((g) => g.id)); for (const cls of mod.classes ?? []) { constructorClasses.set(`%${cls.name}.constructor`, cls.name); for (const m of cls.methods ?? []) { @@ -113,11 +113,10 @@ export function computeMayThrow( if (explain) f.calleeLocs = new Map(); // TDZ reads and non-initializing writes can throw ReferenceError. // Capture locals carry the same flag as the declaring binding. - const tdzIds = new Set([...tdzGlobals, ...fn.locals.filter((l) => l.tdz).map((l) => l.id)]); - const mutableTdzIds = new Set([ - ...tdzGlobals, - ...fn.locals.filter((l) => l.tdz && l.mutable).map((l) => l.id), - ]); + // Every TDZ global counts for both reads and writes; the module's set is + // shared by every function. + const tdzLocals = new Set(fn.locals.filter((l) => l.tdz).map((l) => l.id)); + const mutableTdzLocals = new Set(fn.locals.filter((l) => l.tdz && l.mutable).map((l) => l.id)); // Traverse typed executable nodes without copying the IR into unknown. const visit = (rec: IrExpr | IrStmt): boolean => { const threw = f.throws; @@ -133,13 +132,17 @@ export function computeMayThrow( case "varRef": case "incDec": case "assignExpr": - if (tdzIds.has(rec.localId)) f.throws = true; + if (tdzGlobals.has(rec.localId) || tdzLocals.has(rec.localId)) f.throws = true; break; case "assign": // A declaration is allowed to fill an empty box. Legacy const // TDZ stores also initialize; only mutable subsequent stores // introduce the new write-side exception edge. - if (rec.initializes !== true && mutableTdzIds.has(rec.localId)) f.throws = true; + if ( + rec.initializes !== true && + (tdzGlobals.has(rec.localId) || mutableTdzLocals.has(rec.localId)) + ) + f.throws = true; break; case "dynCheck": case "caughtCheck": diff --git a/packages/compiler/src/backend/vendor-archives.ts b/packages/compiler/src/backend/vendor-archives.ts index dbcaf2d36..ef63aa4e4 100644 --- a/packages/compiler/src/backend/vendor-archives.ts +++ b/packages/compiler/src/backend/vendor-archives.ts @@ -14,6 +14,10 @@ import type { CcDriver } from "./native/driver.js"; const execFileAsync = promisify(execFile); +/** The section flags the libregexp/libunicode objects build with (the + * runtime pack's executable_section_elimination flags). */ +const LRE_SECTION_FLAGS = ["-ffunction-sections", "-fdata-sections"]; + import { QJS_COMMIT, MBEDTLS_VERSION, @@ -191,13 +195,13 @@ export function createVendorArchives(context: VendorArchiveContext) { (sanitize ? "asan" : "plain") + (driver.argv.length === 1 && driver.argv[0] === "clang" ? "" : "-zigcc") + `-${vendorCacheTargetFlavor(driver)}-${buildIdentity}`; - const cacheDir = join(cacheRoot, `${QJS_COMMIT.slice(0, 12)}-lre-${flavor}`); + const cacheDir = join(cacheRoot, `${QJS_COMMIT.slice(0, 12)}-lre-sections-${flavor}`); return LRE_SOURCES.map((f) => join(cacheDir, f.replace(/\.c$/, ".o"))); } /** The libregexp objects for one flavor, compiled lazily on the first * regex-using build (~1s) and cached under - * /vendor/-lre---/*.o — with the same atomic-rename + * /vendor/-lre-sections---/*.o — with the same atomic-rename * publish (parallel first builds race safely; losers discard their work). * Plain is -Os to keep regex-using binaries small; asan matches the final link so * the sanitized lane instruments the matcher too. Cross targets get their @@ -228,6 +232,10 @@ export function createVendorArchives(context: VendorArchiveContext) { vendor: vendorEngineDir(), sources: LRE_SOURCES, objects: lreObjectPaths(sanitize, driver, buildIdentity, cacheRoot), + // Section elimination keeps only the libunicode tables a program + // reaches: the URL host parser's NFC links the normalization tables + // without the regex property ones. + flags: LRE_SECTION_FLAGS, }); } @@ -296,6 +304,7 @@ export function createVendorArchives(context: VendorArchiveContext) { vendor: string; sources: readonly string[]; objects: string[]; + flags?: readonly string[]; }): Promise { const { sanitize, driver, cacheRoot, flavor, name, vendor, sources, objects } = options; if ((await Promise.all(objects.map(validVendorArtifact))).every(Boolean)) return objects; @@ -312,6 +321,7 @@ export function createVendorArchives(context: VendorArchiveContext) { "-std=c11", ...driver.targetArgs, ...(sanitize ? ["-O1", "-fsanitize=address"] : ["-Os"]), + ...(options.flags ?? []), "-I", vendor, "-c", diff --git a/packages/compiler/src/contracts/int32-elements.ts b/packages/compiler/src/contracts/int32-elements.ts index ca61a8e49..67446f2d1 100644 --- a/packages/compiler/src/contracts/int32-elements.ts +++ b/packages/compiler/src/contracts/int32-elements.ts @@ -709,7 +709,8 @@ export class Int32ElementContracts { const findChainsIn = (st: IrStmt): boolean => everyStmtChild(st, findChains, findChainsIn); for (const st of fn.body) findChainsIn(st); // Parameters whose callers are not all visible hold opaque arrays. - if (this.valueEscaped.has(fn.name) || this.runtimeVirtual(fn.name)) + const runtimeVirtual = this.runtimeVirtual; + if (this.valueEscaped.has(fn.name) || runtimeVirtual(fn.name)) for (const p of fn.params) if (this.holds(p.type)) { const why = this.valueEscaped.has(fn.name) diff --git a/packages/compiler/src/frontend/input-tracker.test.ts b/packages/compiler/src/frontend/input-tracker.test.ts index b4f3957f5..d6f9a91b7 100644 --- a/packages/compiler/src/frontend/input-tracker.test.ts +++ b/packages/compiler/src/frontend/input-tracker.test.ts @@ -270,3 +270,49 @@ test("synchronous tracking restores parents after throws and propagates unstable trackedFileExists(join(dir, "outside.ts")); expect(parent.snapshot()).toEqual(before); }); + +test("a settled file is checked by its identity, and any write invalidates it", async () => { + const dir = await mkdtemp(join(tmpdir(), "scriptc-inputs-")); + scratch.push(dir); + const source = join(dir, "entry.ts"); + await writeFile(source, "export const answer = 1;\n"); + const written = fs.statSync(source); + // The file was last changed long enough before the read. + vi.useFakeTimers({ toFake: ["Date"] }); + vi.setSystemTime(Date.now() + 5000); + const tracker = new FrontendInputTracker(); + try { + tracker.run(() => trackedReadFile(source)); + } finally { + vi.useRealTimers(); + } + const snapshot = tracker.snapshot(); + const probe = snapshot.probes.find((entry) => entry.op === "file"); + expect(probe?.op === "file" && probe.identity !== undefined).toBe(true); + expect(validFrontendInputSnapshot(snapshot)).toBe(true); + const reads = vi.mocked(fs.readFileSync); + reads.mockClear(); + expect(frontendInputsStillMatch(snapshot)).toBe(true); + expect(reads.mock.calls.filter(([path]) => path === source)).toHaveLength(0); + // A same-size rewrite that restores the modification time still changes + // the change time. + await writeFile(source, "export const answer = 2;\n"); + fs.utimesSync(source, written.atime, written.mtime); + expect(frontendInputsStillMatch(snapshot)).toBe(false); +}); + +test("a file changed shortly before the read is hashed on every check", async () => { + const dir = await mkdtemp(join(tmpdir(), "scriptc-inputs-")); + scratch.push(dir); + const source = join(dir, "entry.ts"); + await writeFile(source, "export const answer = 1;\n"); + const tracker = new FrontendInputTracker(); + tracker.run(() => trackedReadFile(source)); + const snapshot = tracker.snapshot(); + const probe = snapshot.probes.find((entry) => entry.op === "file"); + expect(probe?.op === "file" && probe.identity).toBeUndefined(); + const reads = vi.mocked(fs.readFileSync); + reads.mockClear(); + expect(frontendInputsStillMatch(snapshot)).toBe(true); + expect(reads.mock.calls.filter(([path]) => path === source)).toHaveLength(1); +}); diff --git a/packages/compiler/src/frontend/input-tracker.ts b/packages/compiler/src/frontend/input-tracker.ts index d27ab6285..c23a87313 100644 --- a/packages/compiler/src/frontend/input-tracker.ts +++ b/packages/compiler/src/frontend/input-tracker.ts @@ -12,7 +12,8 @@ import { observeSourceText } from "./ts7/ast-node.js"; * behind an old resolution answer. */ export type FrontendInputProbe = - | { op: "file"; path: string; digest: string } + /** `identity` (fileIdentity) lets a later check skip hashing the file. */ + | { op: "file"; path: string; digest: string; identity?: string } | { op: "read-error"; path: string } | { op: "kind"; path: string; kind: "file" | "directory" | "other" | "missing" } | { op: "entries"; path: string; files: string[]; directories: string[] } @@ -53,6 +54,36 @@ function frontendSourceDigest(text: string): string { return createHash("sha256").update(text).digest("hex"); } +/** How long a file must have been unchanged before a read for its identity + * to stand for its content: a later write then gets a newer change time + * even on file systems whose timestamps tick once a second or every two. */ +const IDENTITY_SETTLE_MS = 2000; + +/** The device, inode, size, modification time and change time of `path`. A + * write changes the change time, which no tool can set back, so an equal + * identity means the content is the one hashed with it. */ +function fileIdentity(path: string): string | null { + try { + const info = statSync(path); + return `${info.dev}:${info.ino}:${info.size}:${info.mtimeMs}:${info.ctimeMs}`; + } catch { + return null; + } +} + +/** The text of `path` with its identity when the identity was the same before + * and after the read and the file had been unchanged for IDENTITY_SETTLE_MS. */ +function readWithIdentity(path: string): { text: string; identity: string | null } { + const readAt = Date.now(); + const before = statSync(path); + const text = readFileSync(path, "utf8"); + const after = fileIdentity(path); + const identity = `${before.dev}:${before.ino}:${before.size}:${before.mtimeMs}:${before.ctimeMs}`; + const settled = + before.mtimeMs <= readAt - IDENTITY_SETTLE_MS && before.ctimeMs <= readAt - IDENTITY_SETTLE_MS; + return { text, identity: settled && after === identity ? identity : null }; +} + function pathKind(path: string): Extract["kind"] { try { const info = statSync(path); @@ -199,8 +230,11 @@ export function trackedReadFile(path: string): string | null { const virtual = virtualSources.get(path); if (virtual !== undefined) return virtual; try { - const text = readFileSync(path, "utf8"); - record({ op: "file", path, digest: frontendSourceDigest(text) }); + const { text, identity } = readWithIdentity(path); + const digest = frontendSourceDigest(text); + record( + identity === null ? { op: "file", path, digest } : { op: "file", path, digest, identity }, + ); return text; } catch { record({ op: "read-error", path }); @@ -337,6 +371,9 @@ export function frontendInputsStillMatch( } switch (probe.op) { case "file": { + // An unchanged identity stands for the content hashed with it. + if (probe.identity !== undefined && fileIdentity(probe.path) === probe.identity) + return true; try { return frontendSourceDigest(readFileSync(probe.path, "utf8")) === probe.digest; } catch { @@ -443,6 +480,11 @@ export function frontendInputsSemanticallyMatch( probes.push(probe); continue; } + if (probe.identity !== undefined && fileIdentity(probe.path) === probe.identity) { + currentSources.set(probe.path, previous); + probes.push(probe); + continue; + } let current: string; try { current = readFileSync(probe.path, "utf8"); @@ -460,7 +502,7 @@ export function frontendInputsSemanticallyMatch( if (rawText.has(probe.path) || !isSemanticEquivalent(probe.path, previous, current)) return null; changed.push({ path: probe.path, previous, current }); - probes.push({ ...probe, digest: currentDigest }); + probes.push({ op: "file", path: probe.path, digest: currentDigest }); } const adjusted: FrontendInputSnapshot = { ...snapshot, probes }; if (!frontendInputsStillMatch(adjusted, exclusions)) return null; @@ -488,7 +530,11 @@ export function validFrontendInputSnapshot(snapshot: unknown): snapshot is Front if (typeof value.path !== "string" || typeof value.op !== "string") return false; switch (value.op) { case "file": - return typeof value.digest === "string" && /^[0-9a-f]{64}$/.test(value.digest); + return ( + typeof value.digest === "string" && + /^[0-9a-f]{64}$/.test(value.digest) && + (value.identity === undefined || typeof value.identity === "string") + ); case "read-error": return true; case "kind": diff --git a/packages/compiler/src/frontend/lowering/builtins/webassembly.ts b/packages/compiler/src/frontend/lowering/builtins/webassembly.ts index 8e5c3a048..f33ebe9fe 100644 --- a/packages/compiler/src/frontend/lowering/builtins/webassembly.ts +++ b/packages/compiler/src/frontend/lowering/builtins/webassembly.ts @@ -403,7 +403,10 @@ export function lowerWebAssemblyValueOf( recv: IrExpr, loc: IrExpr["loc"], ): IrExpr | null { - const ctx = { checker: lowerer.checker, isStdlibFile: lowerer.isStdlibFile }; + const ctx = { + checker: lowerer.checker, + isStdlibFile: (sf: ts.SourceFile) => lowerer.isStdlibFile(sf), + }; if (!isWebAssemblyType(lowerer.typeOf(receiver), ctx)) return null; lowerer.usesWebAssembly = true; return { kind: "libCall", fn: "wasm.valueOf", args: [recv], type: DYN, loc }; diff --git a/packages/compiler/src/frontend/lowering/class-method-mutations.ts b/packages/compiler/src/frontend/lowering/class-method-mutations.ts index b57729181..8f12d15de 100644 --- a/packages/compiler/src/frontend/lowering/class-method-mutations.ts +++ b/packages/compiler/src/frontend/lowering/class-method-mutations.ts @@ -36,9 +36,20 @@ export function collectClassMethodMutations( const computedWrites: { key: ts.Expression | undefined; receiver: ts.Expression | undefined; + /** The written element access, when the write has one. */ + target?: ts.ElementAccessExpression; site: ts.Node; }[] = []; const propertyWrites: { target: ts.PropertyAccessExpression; site: ts.Node }[] = []; + /** Writes and deletes of a method name through an element access with a + * known key (`o["m"] = f`, `o[k] = f` with `k: "m" | "n"`), judged like + * the named writes in propertyWrites. */ + const keyedWrites: { + name: string; + receiver: ts.Expression; + target: ts.ElementAccessExpression; + site: ts.Node; + }[] = []; const reflection: ts.PropertyAccessExpression[] = []; const globals: ts.Identifier[] = []; const jsAssignments: (ts.PropertyAccessExpression | ts.ElementAccessExpression)[] = []; @@ -274,13 +285,32 @@ export function collectClassMethodMutations( sameBinding(receiver, keys.arguments.nodes[0]!) ); }; - const markKey = (key: ts.Expression, site: ts.Node, receiver?: ts.Expression): void => { + /** A write of the known key `name`: through an element access, a write + * to a method name is judged like a named write (keyedWrites). */ + const markKeyed = ( + name: string, + site: ts.Node, + receiver?: ts.Expression, + target?: ts.ElementAccessExpression, + ): void => { + if (receiver && target && methods.has(name) && name !== "prototype" && name !== "__proto__") { + receiverFields.add(name); + keyedWrites.push({ name, receiver, target, site }); + } else markName(name, site); + }; + const markKey = ( + key: ts.Expression, + site: ts.Node, + receiver?: ts.Expression, + target?: ts.ElementAccessExpression, + ): void => { const value = unwrap(key); - if (ts.isStringLiteralLike(value)) markName(value.text, site); - else if (ts.isNumericLiteral(value)) markName(String(Number(value.text)), site); + if (ts.isStringLiteralLike(value)) markKeyed(value.text, site, receiver, target); + else if (ts.isNumericLiteral(value)) + markKeyed(String(Number(value.text)), site, receiver, target); // The key keeps its assertions: `slots[at as number]` and `slots[at!]` // write a number key whatever the type of `at`. - else computedWrites.push({ key, receiver, site }); + else computedWrites.push(target ? { key, receiver, target, site } : { key, receiver, site }); }; const markFields = (value: ts.Expression, site: ts.Node, receiver?: ts.Expression): void => { value = unwrap(value); @@ -307,7 +337,7 @@ export function collectClassMethodMutations( propertyWrites.push({ target, site }); } else markName(name, site, arrow); } else if (ts.isElementAccessExpression(target)) - markKey(target.argumentExpression, site, target.expression); + markKey(target.argumentExpression, site, target.expression, target); else if (ts.isArrayLiteralExpression(target)) { for (const element of target.elements.nodes) { if (ts.isSpreadElement(element)) markTarget(element.expression, site); @@ -623,7 +653,7 @@ export function collectClassMethodMutations( if (lowerer.prototypeMethodAccesses.has(name)) lowerer.noteMethodSite(name, site); else lowerer.noteMethodAccess(name, site); } - for (const { key, receiver, site } of computedWrites) { + for (const { key, receiver, target, site } of computedWrites) { // A class method declared with a well-known symbol key is not a write; // any other write with one can replace that method. const symbolName = key ? wellKnownMethodName(unwrap(key)) : null; @@ -640,7 +670,7 @@ export function collectClassMethodMutations( const type = key ? lowerer.typeOf(key) : undefined; const names = type && (literalValues(type) ?? (key ? literalLoopKeys(unwrap(key)) : null)); if (names) { - for (const name of names) markName(String(name), site); + for (const name of names) markKeyed(String(name), site, receiver, target); continue; } if (key && receiver && preservesExistingMethod(key, receiver, site)) continue; @@ -662,6 +692,19 @@ export function collectClassMethodMutations( markUnknownKey(site); } } + lowerer.checker.prefetchClassCollection( + keyedWrites.flatMap((write) => [write.target, write.receiver]), + [], + ); + for (const { name, receiver, target, site } of keyedWrites) { + if ( + !dictionaryView(receiver) && + (functionFree(lowerer.typeOf(target)) || !methodOwnerCanInhabit(receiver, name)) + ) + continue; + if (lowerer.prototypeMethodAccesses.has(name)) lowerer.noteMethodSite(name, site); + else lowerer.noteMethodAccess(name, site); + } if (numericKey) { for (const name of [...methods, ...arrowFields]) { if (String(Number(name)) === name) markName(name, numericKey); diff --git a/packages/compiler/src/frontend/lowering/containers/array-pipelines.ts b/packages/compiler/src/frontend/lowering/containers/array-pipelines.ts index 2f839a20b..ff7da5fd7 100644 --- a/packages/compiler/src/frontend/lowering/containers/array-pipelines.ts +++ b/packages/compiler/src/frontend/lowering/containers/array-pipelines.ts @@ -1019,7 +1019,9 @@ class FusedLoop { private callee(arg: IrExpr & { type: FuncType }): Callee { const known = this.callees.find((c) => c.arg === arg); if (known) return known; - const target = this.inlineTarget(arg); + // Called as a plain function: the loop is not the target's receiver. + const inlineTarget = this.inlineTarget; + const target = inlineTarget(arg); let callee: Callee; if (target && arg.kind === "closure") { const captures = new Map(); diff --git a/packages/compiler/src/frontend/lowering/containers/iterator-helpers.ts b/packages/compiler/src/frontend/lowering/containers/iterator-helpers.ts index ba5d0684d..f6580c70e 100644 --- a/packages/compiler/src/frontend/lowering/containers/iterator-helpers.ts +++ b/packages/compiler/src/frontend/lowering/containers/iterator-helpers.ts @@ -1182,10 +1182,10 @@ function bindSource( /** Where chain code allocates its locals: a lifted helper function's own * frame, or hidden locals of the function being lowered (a fused for-of). */ -interface ChainFrame { - local(base: string, type: IrType, mutable: boolean): string; +abstract class ChainFrame { + abstract local(base: string, type: IrType, mutable: boolean): string; /** `id = id + 1` over a number local. */ - bump(id: string): IrStmt; + abstract bump(id: string): IrStmt; } /** `id = id + 1` over a number local. */ @@ -1206,11 +1206,13 @@ function bumpLocal(id: string, loc: SrcLoc): IrStmt { } /** Hidden locals of the function being lowered. */ -class HostFrame implements ChainFrame { +class HostFrame extends ChainFrame { constructor( private readonly lowerer: Lowerer, private readonly loc: SrcLoc, - ) {} + ) { + super(); + } local(base: string, type: IrType, mutable: boolean): string { const local = this.lowerer.declareHiddenLocal(`%${base}`, type); local.mutable = mutable; @@ -1223,11 +1225,13 @@ class HostFrame implements ChainFrame { /** The locals and parameters of a lifted helper function under * construction. */ -export class LiftedFrame implements ChainFrame { +export class LiftedFrame extends ChainFrame { readonly locals: IrLocal[] = []; readonly params: IrParam[] = []; private n = 0; - constructor(private readonly loc: SrcLoc) {} + constructor(private readonly loc: SrcLoc) { + super(); + } local(base: string, type: IrType, mutable: boolean): string { const id = `${base}.${this.n++}`; this.locals.push({ id, name: base, type, mutable }); diff --git a/packages/compiler/src/frontend/lowering/expressions/relational.ts b/packages/compiler/src/frontend/lowering/expressions/relational.ts index 1651bf0d9..393425cdd 100644 --- a/packages/compiler/src/frontend/lowering/expressions/relational.ts +++ b/packages/compiler/src/frontend/lowering/expressions/relational.ts @@ -96,6 +96,8 @@ const NO_CLASS_INSTANCE_KINDS: ReadonlySet = new Set([ "nullT", ]); +const PRIMITIVE_WRAPPERS = ["Boolean", "Number", "String", "BigInt", "Symbol"] as const; + /** `x instanceof C` for a program-declared class C. When x's static * class and C are both in extends-hierarchies the test is dynamic — the * O(1) preorder-interval check against the vtable (`instanceOf` node). @@ -187,6 +189,20 @@ export function lowerInstanceOf(lowerer: Lowerer, expr: ts.BinaryExpression, loc loc, }; } + // The primitive wrapper constructors: a compiled program never creates a + // boxed primitive (`new Boolean(...)` and friends have no lowering), and + // a primitive is never an instance, so the test is false for every value + // (typebox's IsBoolean/IsNumber/IsString guards). The operand still runs. + if (PRIMITIVE_WRAPPERS.some((name) => lowerer.isStdlibGlobal(expr.right, name))) { + const value = lowerer.lowerExpr(expr.left); + return { + kind: "seqExpr", + stmts: [{ kind: "exprStmt", expr: value, loc }], + result: { kind: "boolLit", value: false, type: BOOL, loc }, + type: BOOL, + loc, + }; + } if (lowerer.isStdlibGlobal(expr.right, "FinalizationRegistry")) { const value = lowerer.coerceInto(expr.left, lowerer.lowerExpr(expr.left), DYN); return { kind: "libCall", fn: "finalizationRegistry.is", args: [value], type: BOOL, loc }; diff --git a/packages/compiler/src/frontend/lowering/generic-functions.ts b/packages/compiler/src/frontend/lowering/generic-functions.ts index 222cb89f9..7896da578 100644 --- a/packages/compiler/src/frontend/lowering/generic-functions.ts +++ b/packages/compiler/src/frontend/lowering/generic-functions.ts @@ -645,8 +645,7 @@ function extendsParameter( info: GenericFnInfo, ): boolean { const params = new Set(); - const fn = info.decl as { parameters?: { nodes: readonly ts.ParameterDeclaration[] } }; - for (const param of fn.parameters?.nodes ?? []) { + for (const param of info.decl.parameters.nodes) { const symbol = ts.isIdentifier(param.name) ? lowerer.checker.getSymbolAtLocation(param.name) : undefined; @@ -672,8 +671,7 @@ function declaredInside(node: ts.Node, owner: ts.Node): boolean { * return gives anything else (or nothing). Nested functions and classes keep * their own returns. */ function returnedLocalClass(lowerer: Lowerer, info: GenericFnInfo): ts.ClassLikeDeclaration | null { - const fn = info.decl as ts.Node & { body?: ts.Node }; - const body = fn.body; + const body = info.decl.body; if (!body) return null; const classOf = (node: ts.Expression): ts.ClassLikeDeclaration | null => { let expr = node; diff --git a/packages/compiler/src/frontend/lowering/instantiation-path.test.ts b/packages/compiler/src/frontend/lowering/instantiation-path.test.ts index a31617106..267b402bc 100644 --- a/packages/compiler/src/frontend/lowering/instantiation-path.test.ts +++ b/packages/compiler/src/frontend/lowering/instantiation-path.test.ts @@ -1,19 +1,23 @@ import { expect, test } from "vitest"; +import type * as ts from "../ts7/adapter.js"; import { extendInstantiationPath, MAX_INSTANTIATION_RECURSION } from "./instantiation-path.js"; +/** A stand-in declaration: paths compare declarations by identity only. */ +const node = (): ts.Node => ({}) as ts.Node; + test("independent declarations and roots do not consume a recursive budget", () => { - const declaration = {}; + const declaration = node(); for (let i = 0; i < MAX_INSTANTIATION_RECURSION * 2; i++) expect(extendInstantiationPath(null, declaration)).not.toBeNull(); let path = extendInstantiationPath(null, declaration)!; for (let i = 0; i < MAX_INSTANTIATION_RECURSION * 2; i++) - path = extendInstantiationPath(path, {})!; + path = extendInstantiationPath(path, node())!; expect(extendInstantiationPath(path, declaration)).not.toBeNull(); }); test("bounds fresh recursive demands across intervening declarations", () => { - const first = {}; - const second = {}; + const first = node(); + const second = node(); let path = extendInstantiationPath(null, first)!; for (let i = 1; i < MAX_INSTANTIATION_RECURSION; i++) { path = extendInstantiationPath(path, second)!; diff --git a/packages/compiler/src/frontend/lowering/instantiation-path.ts b/packages/compiler/src/frontend/lowering/instantiation-path.ts index d79760988..b4a8375f6 100644 --- a/packages/compiler/src/frontend/lowering/instantiation-path.ts +++ b/packages/compiler/src/frontend/lowering/instantiation-path.ts @@ -1,7 +1,9 @@ +import type * as ts from "../ts7/adapter.js"; + /** Queued bodies preserve their demand ancestry. Independent concrete uses * do not consume a recursion budget; only fresh instances on one path do. */ export interface InstantiationPath { - declaration: object; + declaration: ts.Node; parent: InstantiationPath | null; } @@ -9,7 +11,7 @@ export const MAX_INSTANTIATION_RECURSION = 100; export function extendInstantiationPath( parent: InstantiationPath | null, - declaration: object, + declaration: ts.Node, ): InstantiationPath | null { let repetitions = 0; for (let path = parent; path; path = path.parent) { diff --git a/packages/compiler/src/frontend/lowering/lower-modules.ts b/packages/compiler/src/frontend/lowering/lower-modules.ts index 75f18fe37..19f7243ed 100644 --- a/packages/compiler/src/frontend/lowering/lower-modules.ts +++ b/packages/compiler/src/frontend/lowering/lower-modules.ts @@ -1183,19 +1183,22 @@ function jsDynHoldableInitializer(lowerer: Lowerer, init: ts.Expression | undefi } if (e.kind === ts.SyntaxKind.NullKeyword) return true; if (ts.isIdentifier(e) && e.text === "undefined") return true; - if (ts.isConditionalExpression(e)) { - return ( - jsDynHoldableInitializer(lowerer, e.whenTrue) && - jsDynHoldableInitializer(lowerer, e.whenFalse) - ); - } + // The arms of a conditional or logical expression: checked values, or + // strings, numbers and booleans, which box into the checked-dynamic + // result (`const APP_NAME = config.name || "pi"`). + const arm = (x: ts.Expression): boolean => { + if (jsDynHoldableInitializer(lowerer, x)) return true; + const kind = lowerer.mapTypeOf(lowerer.typeOf(x))?.kind; + return kind === "string" || kind === "f64" || kind === "bool"; + }; + if (ts.isConditionalExpression(e)) return arm(e.whenTrue) && arm(e.whenFalse); if ( ts.isBinaryExpression(e) && (e.operatorToken.kind === ts.SyntaxKind.QuestionQuestionToken || e.operatorToken.kind === ts.SyntaxKind.BarBarToken || e.operatorToken.kind === ts.SyntaxKind.AmpersandAmpersandToken) ) { - return jsDynHoldableInitializer(lowerer, e.left) && jsDynHoldableInitializer(lowerer, e.right); + return arm(e.left) && arm(e.right); } // A member of a checked module value is itself checked, including a // callable whose inferred overloads cannot describe one native ABI. diff --git a/packages/compiler/src/frontend/lowering/lowerer.ts b/packages/compiler/src/frontend/lowering/lowerer.ts index d4d9d59c1..0b8d448e4 100644 --- a/packages/compiler/src/frontend/lowering/lowerer.ts +++ b/packages/compiler/src/frontend/lowering/lowerer.ts @@ -3756,7 +3756,7 @@ export class Lowerer { positionOf: (node) => this.positions.of(node), shapes: this.shapes, unions: this.unions, - classNamer: this.classNamer, + classNamer: (decl) => this.classNamer(decl), resolveTypeParam: this.typeParamResolver, resolveTypeParamTs: this.typeParamTsResolver, isErasing: () => this.erasingDepth > 0, @@ -3837,8 +3837,8 @@ export class Lowerer { }, mixinIntersectionInstance: (widened) => mixinIntersectionInstanceType(this, widened), classImplementedProtocol: (type) => this.classImplementedProtocol(type), - isStdlibFile: this.isStdlibFile, - isNpmFile: this.isNpmFile, + isStdlibFile: (sf) => this.isStdlibFile(sf), + isNpmFile: (sf) => this.isNpmFile(sf), // The declaration-typed boundary's generated types (decl-projection.ts) // describe data, like an explicitly mapped external type surface. isExternalTypeFile: (sf) => @@ -8102,16 +8102,24 @@ export class Lowerer { lower: () => IrFunction | null; } >(); - const bodyRoots = (...roots: (ts.Node | undefined | null)[]): ts.Node[] => - roots.filter((root) => root !== undefined && root !== null); - const functionRoots = (decl: ts.FunctionLikeDeclaration): ts.Node[] => - bodyRoots(...decl.parameters.nodes.map((param) => param.initializer), decl.body); - const classCtorRoots = (info: ClassInfo): ts.Node[] => - bodyRoots( - ...(info.ctor?.parameters?.nodes ?? []).map((param) => param.initializer), - info.ctor?.body, - ...info.fieldOrder.map((field) => field.initializer), - ); + // Roots are collected with plain pushes: spread arguments into a rest + // parameter pass the nodes as dynamic values. + const addRoot = (roots: ts.Node[], root: ts.Node | undefined | null): void => { + if (root !== undefined && root !== null) roots.push(root); + }; + const functionRoots = (decl: ts.FunctionLikeDeclaration): ts.Node[] => { + const roots: ts.Node[] = []; + for (const param of decl.parameters.nodes) addRoot(roots, param.initializer); + addRoot(roots, decl.body); + return roots; + }; + const classCtorRoots = (info: ClassInfo): ts.Node[] => { + const roots: ts.Node[] = []; + for (const param of info.ctor?.parameters?.nodes ?? []) addRoot(roots, param.initializer); + addRoot(roots, info.ctor?.body); + for (const field of info.fieldOrder) addRoot(roots, field.initializer); + return roots; + }; const classMemberRoots = (info: ClassInfo): ts.Node[] => { const roots = classCtorRoots(info); for (const { member } of this.classMethodMembers(info)) roots.push(...functionRoots(member)); @@ -8288,7 +8296,8 @@ export class Lowerer { register(`%${cName}.set:${prop}`, [], () => this.throwingSetterFn(info, prop)); } }; - for (const name of collectionEdges) this.onEdge(name); + const onEdge = this.onEdge; + for (const name of collectionEdges) onEdge(name); // Generic instances queued by the bodies above lower here (an instance // body fires edges of its own and can queue further instances of // either kind — function instances and class instantiations drain to @@ -8423,7 +8432,8 @@ export class Lowerer { // profile-mapped exports are called from outside the graph, so they // seed the worklist beside the init bodies. Unknown names are inert // (the export-map resolution reports them as SC4002 later). - for (const root of extraRoots ?? []) this.onEdge?.(root); + const rootEdge = this.onEdge; + for (const root of extraRoots ?? []) rootEdge?.(root); const shuffle = lowerShuffle(); const drainUnits = (): void => { while (queue.length > 0) { @@ -8541,7 +8551,9 @@ export class Lowerer { /** Reachability hook (see emitReachable): fires when lowering resolves a * reference to a lowerable body. */ noteEdge(name: string): void { - if (this.onEdge) this.onEdge(name); + // Called as a plain function: the instance is not its receiver. + const onEdge = this.onEdge; + if (onEdge) onEdge(name); } /** A subclass constructor's super() edge: the base constructor runs, but @@ -11952,12 +11964,15 @@ export class Lowerer { /** Method-like members with their collected method-map names. Body * lowering omits abstract declarations; ABI analysis must include them - * because an abstract-typed call selects the same virtual slot. */ - *classMethodMembers( + * because an abstract-typed call selects the same virtual slot. An array, + * not a generator: a generator's yielded values are dynamic values, which + * would make every syntax node one. */ + classMethodMembers( info: ClassInfo, includeAbstract = false, - ): Generator<{ mName: string; member: ts.MethodDeclaration | ts.AccessorDeclaration }> { - if (!info.decl) return; // builtin error classes: runtime-provided bodies + ): { mName: string; member: ts.MethodDeclaration | ts.AccessorDeclaration }[] { + const members: { mName: string; member: ts.MethodDeclaration | ts.AccessorDeclaration }[] = []; + if (!info.decl) return members; // builtin error classes: runtime-provided bodies for (const member of info.decl.members.nodes) { const fnLike = ts.isMethodDeclaration(member) || ts.isGetAccessor(member) || ts.isSetAccessor(member) @@ -11981,8 +11996,9 @@ export class Lowerer { : `${ts.isGetAccessor(fnLike) ? "get" : "set"}:${baseName}`; const signature = info.methods.get(mName); if (!signature || (!fnLike.body && !(includeAbstract && signature.abstract))) continue; - yield { mName, member: fnLike }; + members.push({ mName, member: fnLike }); } + return members; } lowerClassCtor(info: ClassInfo): IrFunction { diff --git a/packages/compiler/src/frontend/lowering/site-dependencies.test.ts b/packages/compiler/src/frontend/lowering/site-dependencies.test.ts index ca5affcbe..0580a52ee 100644 --- a/packages/compiler/src/frontend/lowering/site-dependencies.test.ts +++ b/packages/compiler/src/frontend/lowering/site-dependencies.test.ts @@ -1,9 +1,13 @@ import { expect, test } from "vitest"; +import type * as ts from "../ts7/adapter.js"; import { SiteDependencies } from "./site-dependencies.js"; +/** A stand-in node fact: facts are compared by identity. */ +const node = (): ts.Node => ({}) as ts.Node; + test("a changed fact makes the sites that read it stale", () => { const dependencies = new SiteDependencies(3); - const a = {}; + const a = node(); const b = "key"; dependencies.begin(0); dependencies.read(a); @@ -20,8 +24,8 @@ test("a changed fact makes the sites that read it stale", () => { test("a visit that changes a fact makes every reader of its facts stale", () => { const dependencies = new SiteDependencies(3); - const shared = {}; - const own = {}; + const shared = node(); + const own = node(); dependencies.begin(0); dependencies.read(shared); dependencies.end(false); @@ -37,7 +41,7 @@ test("a visit that changes a fact makes every reader of its facts stale", () => test("reads outside a visit are not recorded", () => { const dependencies = new SiteDependencies(1); - const fact = {}; + const fact = node(); dependencies.read(fact); dependencies.changed(fact); expect(dependencies.isStale(0)).toBe(false); @@ -45,7 +49,7 @@ test("reads outside a visit are not recorded", () => { test("a site visited again is listed once per fact", () => { const dependencies = new SiteDependencies(3); - const fact = {}; + const fact = node(); for (const site of [0, 2, 1, 2, 0]) { dependencies.begin(site); dependencies.read(fact); diff --git a/packages/compiler/src/frontend/lowering/site-dependencies.ts b/packages/compiler/src/frontend/lowering/site-dependencies.ts index da8531e48..868531353 100644 --- a/packages/compiler/src/frontend/lowering/site-dependencies.ts +++ b/packages/compiler/src/frontend/lowering/site-dependencies.ts @@ -1,3 +1,5 @@ +import type * as ts from "../ts7/adapter.js"; + /** The facts each site of a fixed-point scan read on its last visit, so that * a pass visits again only the sites whose facts changed since. * @@ -13,14 +15,19 @@ * that with a pass over every site once a pass changes nothing, and resumes * when that pass changes a fact: a fact this tracking misses costs passes, * not a different result. */ +/** A fact a site reads: a symbol, a syntax node, or the name of a derived + * fact. A typed union keeps nodes and symbols native values; as `unknown` + * every syntax node would become a dynamic value. */ +export type SiteFact = ts.Node | ts.Symbol | string; + export class SiteDependencies { /** Each fact's entry: the visit that last read it, so that a visit lists a * fact once, followed by the sites that read it, ascending. */ - private readonly readers = new Map(); + private readonly readers = new Map(); private readonly stale: Uint8Array; private site = -1; private visit = 0; - private facts: unknown[] = []; + private facts: SiteFact[] = []; constructor(siteCount: number) { this.stale = new Uint8Array(siteCount); @@ -46,7 +53,7 @@ export class SiteDependencies { } /** The site being visited read `fact`. */ - read(fact: unknown): void { + read(fact: SiteFact): void { if (this.site < 0) return; const entry = this.readers.get(fact); if (entry === undefined) { @@ -72,7 +79,7 @@ export class SiteDependencies { } /** `fact` changed: every site that read it is visited again. */ - changed(fact: unknown): void { + changed(fact: SiteFact): void { const entry = this.readers.get(fact); if (entry !== undefined) for (let i = 1; i < entry.length; i++) this.stale[entry[i]!] = 1; } diff --git a/packages/compiler/src/frontend/lowering/string-index-bounds.ts b/packages/compiler/src/frontend/lowering/string-index-bounds.ts index af1b00ec6..6418aed76 100644 --- a/packages/compiler/src/frontend/lowering/string-index-bounds.ts +++ b/packages/compiler/src/frontend/lowering/string-index-bounds.ts @@ -182,11 +182,28 @@ export class StringIndexBounds { private readonly identifiers = new Map>(); constructor( - private readonly symbolAt: (node: ts.Node) => ts.Symbol | null, - private readonly declarationOf: (symbol: ts.Symbol) => ts.Node | undefined, - private readonly isString: (node: ts.Expression) => boolean, + private readonly lookupSymbol: (node: ts.Node) => ts.Symbol | null, + private readonly lookupDeclaration: (symbol: ts.Symbol) => ts.Node | undefined, + private readonly stringTyped: (node: ts.Expression) => boolean, ) {} + // The lookups are called as plain functions: the instance is not their + // receiver. + private symbolAt(node: ts.Node): ts.Symbol | null { + const lookup = this.lookupSymbol; + return lookup(node); + } + + private declarationOf(symbol: ts.Symbol): ts.Node | undefined { + const lookup = this.lookupDeclaration; + return lookup(symbol); + } + + private isString(node: ts.Expression): boolean { + const test = this.stringTyped; + return test(node); + } + /** True when `read` (an element access on a string receiver) is proven * to name an existing code unit. */ inBounds(read: ts.ElementAccessExpression): boolean { diff --git a/packages/compiler/src/frontend/npm-static.ts b/packages/compiler/src/frontend/npm-static.ts index eed02c95c..b00e14aa0 100644 --- a/packages/compiler/src/frontend/npm-static.ts +++ b/packages/compiler/src/frontend/npm-static.ts @@ -254,13 +254,18 @@ function aliasedPackageName(pkgDir: string): string | null { export function npmStaticTransformPkgJson(pkg: Record): void { delete pkg["types"]; delete pkg["typings"]; - const transform = (v: unknown): void => { - if (v === null || typeof v !== "object") return; - if (Array.isArray(v)) { - for (const item of v) transform(item); - return; - } - const obj = v as Record; + const transform = (v: unknown): unknown => { + if (v === null || typeof v !== "object") return v; + if (Array.isArray(v)) return v.map(transform); + // The result is a new object, built in key order; the first entry of a + // key wins. + const entries: [string, unknown][] = []; + const seen = new Set(); + const add = (key: string, value: unknown): void => { + if (seen.has(key)) return; + seen.add(key); + entries.push([key, value]); + }; // 0. The "node" condition HOISTS: Node's runtime resolution matches it // (yaml's `{ types, node: "./dist/index.js", default: "./browser/ // index.js" }` runs the node dist; the browser dist is a DIFFERENT @@ -270,24 +275,18 @@ export function npmStaticTransformPkgJson(pkg: Record): void { // under import+require (either matcher finds it there), an object's // entries splice inline — earlier real keys keep winning, exactly // Node's first-match-in-key-order rule. - if (obj["node"] !== undefined) { - const entries = Object.entries(obj); - for (const k of Object.keys(obj)) delete obj[k]; - for (const [k, val] of entries) { - if (k !== "node") { - if (!(k in obj)) obj[k] = val; - continue; - } - const inner: Record = - typeof val === "string" - ? { import: val, require: val } - : val !== null && typeof val === "object" && !Array.isArray(val) - ? (val as Record) - : {}; - for (const [ik, iv] of Object.entries(inner)) { - if (!(ik in obj)) obj[ik] = iv; - } + for (const [key, value] of Object.entries(v as Record)) { + if (key !== "node") { + add(key, value); + continue; } + const inner: Record = + typeof value === "string" + ? { import: value, require: value } + : value !== null && typeof value === "object" && !Array.isArray(value) + ? (value as Record) + : {}; + for (const [innerKey, innerValue] of Object.entries(inner)) add(innerKey, innerValue); } // 1. The types claims go AFTER the hoist: a node object's own // "types" key (uuid's `node: { types: "./dist/index.d.ts", default: @@ -295,15 +294,17 @@ export function npmStaticTransformPkgJson(pkg: Record): void { // in place tsgo follows it, maps the hidden declaration to its // same-stem JS (dist/index.js, the browser build) and disagrees with // this resolver's runtime answer. - delete obj["types"]; - delete obj["typings"]; - for (const value of Object.values(obj)) transform(value); + return Object.fromEntries( + entries + .filter(([key]) => key !== "types" && key !== "typings") + .map(([key, value]): [string, unknown] => [key, transform(value)]), + ); }; - if (pkg["exports"] !== undefined) transform(pkg["exports"]); + if (pkg["exports"] !== undefined) pkg["exports"] = transform(pkg["exports"]); // The "imports" field maps "#" specifiers with the same condition objects // (chalk 5's `"#supports-color": { node, default }`), so Node's "node" // target hoists there too. - if (pkg["imports"] !== undefined) transform(pkg["imports"]); + if (pkg["imports"] !== undefined) pkg["imports"] = transform(pkg["imports"]); } /** npmStaticTransformPkgJson over a package.json TEXT (the fs shadow's diff --git a/packages/compiler/src/frontend/services.ts b/packages/compiler/src/frontend/services.ts index 19728d6ce..14b366bfa 100644 --- a/packages/compiler/src/frontend/services.ts +++ b/packages/compiler/src/frontend/services.ts @@ -52,8 +52,10 @@ export class FrontendServices { } private sourceParser(): Ts7SourceParser { this.ensureOpen(); + // Called as a plain function: the instance is not its receiver. + const createApi = this.createApi; return (this.parser ??= new Ts7SourceParser( - (options) => this.createApi({ ...options, collectTiming: false }), + (options) => createApi({ ...options, collectTiming: false }), this.cwd, )); } @@ -62,9 +64,10 @@ export class FrontendServices { } evaluateComptime(source: string, timeoutMs: number): unknown { this.ensureOpen(); - if (this.comptimeEvaluator === undefined) + const evaluate = this.comptimeEvaluator; + if (evaluate === undefined) throw new Error("this compiler host does not provide compile-time evaluation"); - return this.comptimeEvaluator(source, timeoutMs); + return evaluate(source, timeoutMs); } createProgramHost(options?: Ts7HostOptions): Ts7Host { this.ensureOpen(); diff --git a/packages/compiler/src/frontend/token-equivalence.test.ts b/packages/compiler/src/frontend/token-equivalence.test.ts index 880a07221..7bbe08929 100644 --- a/packages/compiler/src/frontend/token-equivalence.test.ts +++ b/packages/compiler/src/frontend/token-equivalence.test.ts @@ -1,8 +1,10 @@ import { expect, test } from "vitest"; -import { tokenEquivalentSource } from "./token-equivalence.js"; +import { tokenEquivalentSource, tokenPositionEquivalentSource } from "./token-equivalence.js"; const equivalent = (previous: string, current: string, path = "/src/a.ts") => tokenEquivalentSource(path, previous, current); +const positionEquivalent = (previous: string, current: string, path = "/src/a.ts") => + tokenPositionEquivalentSource(path, previous, current); test("whitespace and plain comments are trivia", () => { expect(equivalent("const a = 1;\n", "const a = 1; // one\n")).toBe(true); @@ -105,3 +107,50 @@ test("text the scanner cannot classify is never equivalent", () => { expect(equivalent("const a = ; // a\n", "const a = ; // b\n", "/src/a.tsx")).toBe(false); expect(equivalent('const s = "open\n', 'const s = "open // \n')).toBe(false); }); + +test("a / after a non-null assertion divides", () => { + expect( + equivalent("const q = data[i]! / 7; // ratio\n", "const q = data[i]! / 7; // the ratio\n"), + ).toBe(true); + expect( + equivalent("const q = value! / 7; // ratio\n", "const q = value! / 7; // the ratio\n"), + ).toBe(true); + expect(equivalent("const q = this! / 7; // ratio\n", "const q = this! / 7; // the ratio\n")).toBe( + true, + ); + // A prefix `!` negates a regular expression, whose `//` is code. + expect( + equivalent("const t = !/[//]/.test(s); c = 1;\n", "const t = !/[//]/.test(s); c = 2;\n"), + ).toBe(false); + expect(equivalent("return !/[//]/.test(s); // one\n", "return !/[//]/.test(s); // two\n")).toBe( + true, + ); + // After a line break, `!` starts a new expression. + expect(equivalent("a\n!/[//]/.test(s); c = 1;\n", "a\n!/[//]/.test(s); c = 2;\n")).toBe(false); + // `f()!` and `if (a) !` read the same up to the `/`. + expect(equivalent("const q = f()! / 2; // one\n", "const q = f()! / 2; // two\n")).toBe(false); +}); + +test("position equivalence also keeps every token's line and column", () => { + const source = "const a = 1; // one\nconst b = a; /* two */ f(b);\n"; + // A comment's text within its line, after the line's last token. + expect( + positionEquivalent(source, "const a = 1; // the first\nconst b = a; /* two */ f(b);\n"), + ).toBe(true); + // Comments and whitespace after the last token, and trailing whitespace. + expect(positionEquivalent(source, source + "// appended\n\n")).toBe(true); + expect( + positionEquivalent(source, "const a = 1; // one\nconst b = a; /* two */ f(b); \n"), + ).toBe(true); + expect(positionEquivalent(source, source.replaceAll("\n", "\r\n"))).toBe(true); + // A longer comment before a token moves its column; a new line moves the + // lines below it. Both stay token-equivalent. + const column = "const a = 1; // one\nconst b = a; /* the second */ f(b);\n"; + const line = "// header\nconst a = 1; // one\nconst b = a; /* two */ f(b);\n"; + expect(positionEquivalent(source, column)).toBe(false); + expect(positionEquivalent(source, line)).toBe(false); + expect(equivalent(source, column)).toBe(true); + expect(equivalent(source, line)).toBe(true); + // A token edit is never equivalent. + expect(positionEquivalent(source, source.replace("1", "2"))).toBe(false); +}); diff --git a/packages/compiler/src/frontend/token-equivalence.ts b/packages/compiler/src/frontend/token-equivalence.ts index 7a0f6cc79..9f6fd4613 100644 --- a/packages/compiler/src/frontend/token-equivalence.ts +++ b/packages/compiler/src/frontend/token-equivalence.ts @@ -17,7 +17,10 @@ * before it. After `)` it divides unless the parenthesis closes an `if`, * `while`, `for`, `for await`, or `with` header. After `}`, `++` or * `--` the scanner cannot tell which, so it gives up, as it does for - * escaped identifiers, which can spell keywords. + * escaped identifiers, which can spell keywords. After `!` in + * TypeScript it divides when the `!` follows an operand on the same line + * (a non-null assertion, `x! / 2`) and gives up when that operand is a + * `)`, `}`, `++` or `--`. * - A `/` with a line break (or a comment that spans lines) since the * previous token gives up: whether automatic semicolon insertion ends the * statement before it decides how it reads. @@ -35,6 +38,8 @@ interface Token { text: string; comment: boolean; lineBreakBefore: boolean; + /** The offset of the token's first code unit. */ + start: number; } /** Keywords after which `/` starts a regular expression. */ @@ -135,14 +140,20 @@ function scan(path: string, source: string): Token[] | null { // A line break, or a comment spanning lines, since the last token other // than a comment. let breakSinceCode = false; - // The previous token and the one before it, and whether a `)` closed a - // statement header. - const last: { token: string | null; before: string | null } = { token: null, before: null }; + // The previous token and the one before it, whether a line break came + // before the previous token, and whether a `)` closed a statement header. + const last: { token: string | null; before: string | null; broke: boolean } = { + token: null, + before: null, + broke: false, + }; let closedHeader = false; + // Every token ends at `index` when it is pushed. const push = (text: string, comment = false): void => { - tokens.push({ text, comment, lineBreakBefore }); + tokens.push({ text, comment, lineBreakBefore, start: index - text.length }); lineBreakBefore = false; if (!comment) { + last.broke = breakSinceCode; breakSinceCode = false; last.before = last.token; last.token = text; @@ -291,7 +302,19 @@ function scan(path: string, source: string): Token[] | null { else if (before.startsWith("'") || before.startsWith('"') || before.endsWith("`")) regex = false; else if (before.startsWith("/") && before.length > 1 && before !== "/=") regex = false; - else regex = true; + else if (before === "!" && typeScript && !last.broke && last.before !== null) { + // `x! / 2` divides: a `!` right after an operand on the same line is + // a non-null assertion. `!/re/.test(s)` negates a regular expression. + const operand = last.before; + if (operand === ")" || operand === "}" || operand === "++" || operand === "--") return null; + regex = !( + operand === "]" || + operand.startsWith("'") || + operand.startsWith('"') || + operand.endsWith("`") || + (/^[A-Za-z0-9_$]/.test(operand) && !EXPRESSION_KEYWORDS.has(operand)) + ); + } else regex = true; if (regex) { const start = index; index++; @@ -373,9 +396,33 @@ function scan(path: string, source: string): Token[] | null { return tokens; } -/** Whether `current` differs from `previous` only in whitespace and in - * comments that cannot affect compilation. */ -export function tokenEquivalentSource(path: string, previous: string, current: string): boolean { +/** The line and column of each token's start, with the line terminators + * source locations count (backend/source-locations.ts). */ +function tokenPositions(source: string, tokens: readonly Token[]): number[] { + const positions: number[] = []; + let line = 0; + let lineStart = 0; + let offset = 0; + for (const token of tokens) { + for (; offset < token.start; offset++) { + const code = source.charCodeAt(offset); + if (code === 13 && source.charCodeAt(offset + 1) === 10) offset++; + if (code === 10 || code === 13 || code === 0x2028 || code === 0x2029) { + line++; + lineStart = offset + 1; + } + } + positions.push(line, token.start - lineStart); + } + return positions; +} + +function equivalentSources( + path: string, + previous: string, + current: string, + positions: boolean, +): boolean { if (previous === current) return true; const left = scan(path, previous); if (left === null) return false; @@ -387,5 +434,27 @@ export function tokenEquivalentSource(path: string, previous: string, current: s if (a.text !== b.text || a.comment !== b.comment) return false; if (index > 0 && a.lineBreakBefore !== b.lineBreakBefore) return false; } - return true; + if (!positions) return true; + const before = tokenPositions(previous, left); + const after = tokenPositions(current, right); + return before.every((value, index) => value === after[index]); +} + +/** Whether `current` differs from `previous` only in whitespace and in + * comments that cannot affect compilation. */ +export function tokenEquivalentSource(path: string, previous: string, current: string): boolean { + return equivalentSources(path, previous, current, false); +} + +/** Whether `current` is token-equivalent to `previous` (tokenEquivalentSource) + * and every token keeps its line and column: the edit leaves the source + * positions debug information records unchanged (a comment's text within + * its line, comments and whitespace after the last token, trailing + * whitespace). */ +export function tokenPositionEquivalentSource( + path: string, + previous: string, + current: string, +): boolean { + return equivalentSources(path, previous, current, true); } diff --git a/packages/compiler/src/frontend/ts7/lib-diagnostics.test.ts b/packages/compiler/src/frontend/ts7/lib-diagnostics.test.ts index f041a6cb5..e87db5f1a 100644 --- a/packages/compiler/src/frontend/ts7/lib-diagnostics.test.ts +++ b/packages/compiler/src/frontend/ts7/lib-diagnostics.test.ts @@ -219,3 +219,53 @@ test("a supplied file that changes is a different entry", () => { expect(store.entries.size).toBe(2); writeFileSync(fallback, fallbackText); }); + +/** The program's files, recording which ones were read. */ +function recording(texts: Record): ((name: string) => SourceFile | undefined) & { + read: string[]; +} { + const inner = files(texts); + const read: string[] = []; + return Object.assign((name: string) => (read.push(name), inner(name)), { read }); +} + +test("a program of modules stores its answer without reading the supplied files", () => { + const store = memoryStore(); + setLibDiagnosticsStore(store, "checker"); + const sources = recording({ [main]: "export const n: number = 'x';\n" }); + reusedLibDiagnostics( + sessionProgram({ semantic: new Map([[main, []]]) }), + names, + supplied, + sources, + ); + expect(sources.read).toEqual([main]); + const [text] = [...store.entries.values()]; + expect(JSON.parse(text!).globals).toBeNull(); + // A later program of modules reuses it, again without reading them. + const again = recording({ [main]: "export const n: number = 'x';\n" }); + const program = sessionProgram({ semantic: new Map([[main, []]]) }); + expect(reusedLibDiagnostics(program, names, supplied, again)).not.toBeNull(); + expect(program.asked).toEqual([`checked ${main}`]); + expect(again.read).toEqual([main]); +}); + +test("a script that uses an answer stored without global names collects and stores them", () => { + const store = memoryStore(); + setLibDiagnosticsStore(store, "checker"); + const semantic = new Map([[main, []]]); + reusedLibDiagnostics(sessionProgram({ semantic }), names, supplied, moduleFile); + const script = recording({ [main]: "const total = 1;\n" }); + const program = sessionProgram({ semantic }); + expect(reusedLibDiagnostics(program, names, supplied, script)).not.toBeNull(); + expect(program.asked).toEqual([`checked ${main}`]); + expect(script.read).toEqual([main, lib, fallback]); + const [text] = [...store.entries.values()]; + expect(new Set(JSON.parse(text!).globals)).toEqual(new Set(["NaN", "Array", "process"])); + // A script that declares one of them runs the full check, and the entry keeps its names. + const clash = sessionProgram({ semantic }); + reusedLibDiagnostics(clash, names, supplied, files({ [main]: "declare var NaN: number;\n" })); + expect(clash.asked[0]).toBe("getGlobalDiagnostics"); + expect(store.entries.size).toBe(1); + expect(JSON.parse([...store.entries.values()][0]!).globals).not.toBeNull(); +}); diff --git a/packages/compiler/src/frontend/ts7/lib-diagnostics.ts b/packages/compiler/src/frontend/ts7/lib-diagnostics.ts index 827b337bd..30c6e81f0 100644 --- a/packages/compiler/src/frontend/ts7/lib-diagnostics.ts +++ b/packages/compiler/src/frontend/ts7/lib-diagnostics.ts @@ -10,7 +10,14 @@ * (getDiagnosticsOfCheckedFiles), and the semantic diagnostics merge in the * order the full check sorts them, by file name. An answer is stored only * when the full check reports no global diagnostics: the global diagnostics - * of checking the other files are then the full check's. */ + * of checking the other files are then the full check's. + * + * The global names the supplied files declare matter only to programs with + * scripts, and collecting them reads every supplied file's syntax tree, + * which a program of modules does not otherwise read (about 4 MB at the + * peak of a small build). An answer stored by such a program leaves them + * out; the first program with scripts that uses the answer collects them + * and stores them with it. */ import { createHash } from "node:crypto"; import { statSync } from "node:fs"; import type { Node, SourceFile } from "./ast-types.js"; @@ -25,7 +32,7 @@ export interface LibDiagnosticsStore { write(key: string, text: string): void; } -const SCHEMA = "scriptc.lib-diagnostics.v1"; +const SCHEMA = "scriptc.lib-diagnostics.v2"; let configured: { store: LibDiagnosticsStore; checker: string } | null = null; @@ -89,7 +96,14 @@ export function reusedLibDiagnostics( const key = entryKey(checker, options, libs); if (key === null) return null; const stored = readEntry(store, key, libs); - if (stored !== null && !collides(stored.globals)) { + // An answer stored without the supplied files' global names: a program + // with scripts collects them and stores them with the answer. + let globals = stored?.globals ?? null; + if (stored !== null && globals === null && scripts.length > 0) { + globals = suppliedGlobalNames(libs, sourceFile); + if (globals !== null) writeEntry(store, key, globals, Object.fromEntries(stored.files)); + } + if (stored !== null && (scripts.length === 0 || (globals !== null && !collides(globals)))) { const checked = program.checkedFileDiagnosticData(own); if (checked === null) return null; const byFile = new Map(stored.files); @@ -104,20 +118,35 @@ export function reusedLibDiagnostics( const semantic = program.diagnosticData("getSemanticDiagnostics"); // A server that cannot check only some files could never use the answer. if (global.length === 0 && stored === null && program.checkedFileDiagnosticData([]) !== null) { - const globals = suppliedGlobalNames(libs, sourceFile); - if (globals !== null && !collides(globals)) { + const names = scripts.length === 0 ? null : suppliedGlobalNames(libs, sourceFile); + if (scripts.length === 0 || (names !== null && !collides(names))) { const files: Record = {}; for (const name of libs) files[name] = program.diagnosticData("getSemanticDiagnostics", name); - try { - store.write(key, JSON.stringify({ schema: SCHEMA, globals: [...globals], files })); - } catch { - // Best effort: a failed write is a later miss. - } + writeEntry(store, key, names, files); } } return { semantic, global }; } +/** Stores an answer: the supplied files' diagnostics, and their global names + * when they were collected (null otherwise). Best effort: a failed write is + * a later miss. */ +function writeEntry( + store: LibDiagnosticsStore, + key: string, + globals: ReadonlySet | null, + files: Record, +): void { + try { + store.write( + key, + JSON.stringify({ schema: SCHEMA, globals: globals === null ? null : [...globals], files }), + ); + } catch { + // A failed write is a later miss. + } +} + /** Whether a file cannot add to or change the global declarations the * supplied files declare: a module with no `declare global`, no module * augmentation and no UMD `export as namespace` declaration. */ @@ -276,7 +305,7 @@ function readEntry( store: LibDiagnosticsStore, key: string, libs: readonly string[], -): { files: Map; globals: Set } | null { +): { files: Map; globals: Set | null } | null { let text: string | null; try { text = store.read(key); @@ -290,7 +319,11 @@ function readEntry( } catch { return null; } - if (entry.schema !== SCHEMA || entry.files === undefined || !Array.isArray(entry.globals)) + if ( + entry.schema !== SCHEMA || + entry.files === undefined || + (entry.globals !== null && !Array.isArray(entry.globals)) + ) return null; const files = new Map(); for (const name of libs) { @@ -298,7 +331,10 @@ function readEntry( if (!Array.isArray(list)) return null; files.set(name, list); } - return { files, globals: new Set(entry.globals.filter((name) => typeof name === "string")) }; + const globals = Array.isArray(entry.globals) + ? new Set(entry.globals.filter((name): name is string => typeof name === "string")) + : null; + return { files, globals }; } /** Each file's diagnostics (already in the checker's order) in the order the diff --git a/packages/compiler/src/frontend/ts7/semantic-model.ts b/packages/compiler/src/frontend/ts7/semantic-model.ts index 405f4775b..04fa8e637 100644 --- a/packages/compiler/src/frontend/ts7/semantic-model.ts +++ b/packages/compiler/src/frontend/ts7/semantic-model.ts @@ -64,7 +64,9 @@ export interface SemanticQuery { export class SemanticProgram { constructor(private readonly sourceFile: (path: string) => AstNode | undefined) {} getSourceFile(path: string): AstNode | undefined { - return this.sourceFile(path); + // Called as a plain function: the instance is not its receiver. + const sourceFile = this.sourceFile; + return sourceFile(path); } } diff --git a/packages/compiler/src/frontend/ts7/source-parser.ts b/packages/compiler/src/frontend/ts7/source-parser.ts index f9029774c..bbb6cae7a 100644 --- a/packages/compiler/src/frontend/ts7/source-parser.ts +++ b/packages/compiler/src/frontend/ts7/source-parser.ts @@ -42,7 +42,9 @@ export class Ts7SourceParser { } private connect(): Ts7Api { - return this.createApi({ + // Called as a plain function: the instance is not its receiver. + const createApi = this.createApi; + return createApi({ cwd: this.cwd, fs: { readFile: (path) => this.files.get(tsgoPath(path)) ?? null, diff --git a/packages/compiler/src/ir/callback-specialization.ts b/packages/compiler/src/ir/callback-specialization.ts index 6038ccaa5..4764754af 100644 --- a/packages/compiler/src/ir/callback-specialization.ts +++ b/packages/compiler/src/ir/callback-specialization.ts @@ -130,10 +130,7 @@ export function specializeCallbackCalls( const { written, closureWritten } = rebindings(mod); const copies = new Map(); const candidates = new Map(); - let totalNodes = 0; for (const fn of mod.functions) { - const nodes = countNodes(fn); - totalNodes += nodes; if ( !(fn.ownsPrototype || fn.name.startsWith("%arr.")) || fn.name === mod.entry || @@ -142,10 +139,11 @@ export function specializeCallbackCalls( fn.captures !== undefined || fn.classCaptures?.length || fn.fastRead !== undefined || - fn.speculativeDispatch || - nodes > MAX_FUNCTION_NODES + fn.speculativeDispatch ) continue; + const nodes = countNodes(fn); + if (nodes > MAX_FUNCTION_NODES) continue; const rebound = written.get(fn.name)!; const slots: number[] = []; fn.params.forEach((param, index) => { @@ -165,19 +163,30 @@ export function specializeCallbackCalls( const definitions = new Map>(); const globalIds = new Set(mod.globals?.map((global) => global.id)); const globalWrites = new Map(); + // The sum of countNodes over every function, counted by the walk below. + let totalNodes = 0; for (const fn of mod.functions) { + let nodes = 0; const values = new Map(); const counts = new Map(); const record = (id: string, value: IrExpr | null, weight: number): void => { counts.set(id, (counts.get(id) ?? 0) + weight); - if (value) values.set(id, [...(values.get(id) ?? []), value]); + if (value) { + const list = values.get(id); + if (list === undefined) values.set(id, [value]); + else list.push(value); + } }; everyStmtList(fn.body, { stmt: (stmt) => { + nodes++; if (stmt.kind === "varDecl") record(stmt.localId, stmt.init, stmt.init ? 1 : 0); if (stmt.kind === "assign") { - if (globalIds.has(stmt.localId)) - globalWrites.set(stmt.localId, [...(globalWrites.get(stmt.localId) ?? []), stmt.value]); + if (globalIds.has(stmt.localId)) { + const list = globalWrites.get(stmt.localId); + if (list === undefined) globalWrites.set(stmt.localId, [stmt.value]); + else list.push(stmt.value); + } record(stmt.localId, stmt.value, 1); } if (stmt.kind === "forOf") record(stmt.localId, null, 2); @@ -186,6 +195,7 @@ export function specializeCallbackCalls( return true; }, expr: (expr) => { + nodes++; if (expr.kind === "assignExpr" || expr.kind === "incDec") { record(expr.localId, null, 2); if (globalIds.has(expr.localId)) globalWrites.set(expr.localId, []); @@ -208,6 +218,7 @@ export function specializeCallbackCalls( ) single.set(id, list[0]!); definitions.set(fn.name, single); + totalNodes += Math.min(nodes, MAX_FUNCTION_NODES + 1); } const globalDefinitions = new Map(); for (const global of mod.globals ?? []) diff --git a/packages/compiler/src/ir/capture-values.ts b/packages/compiler/src/ir/capture-values.ts index 779ca1b76..bcc587f1b 100644 --- a/packages/compiler/src/ir/capture-values.ts +++ b/packages/compiler/src/ir/capture-values.ts @@ -538,7 +538,7 @@ export class CaptureFlow { facts.callees.add(e.callee.localId); break; } - return everyExprChild(e, expr, (s) => stmt(s, false)); + return everyExprChild(e, expr, nested); }; // `split` mirrors rewrite(): only statement lists reached through // ordinary nesting from the body are expanded. Switch cases share one @@ -599,8 +599,10 @@ export class CaptureFlow { list(s.body, split) ); } - return everyStmtChild(s, expr, (child) => stmt(child, false)); + return everyStmtChild(s, expr, nested); }; + // Statements nested in expressions and in other statements' heads. + const nested = (s: IrStmt): boolean => stmt(s, false); list(fn.body, true); return facts; } diff --git a/packages/compiler/src/ir/constant-globals.ts b/packages/compiler/src/ir/constant-globals.ts index f18764223..76b086f40 100644 --- a/packages/compiler/src/ir/constant-globals.ts +++ b/packages/compiler/src/ir/constant-globals.ts @@ -36,7 +36,11 @@ export function foldConstantGlobals(mod: IrModule): IrModule { if (value && (value.kind === "numLit" || value.kind === "boolLit")) values.set(id, value); else values.delete(id); }; - for (const fn of mod.functions) + // Per function, the candidates it reads: the rewrite below visits only the + // functions that read a constant. + const readsByFunction = new Map>(); + for (const fn of mod.functions) { + let read: Set | undefined; everyStmtList(fn.body, { // Every node other than a read that names the binding is a write // (assignments, loop and catch bindings, compound updates); only a @@ -48,16 +52,21 @@ export function foldConstantGlobals(mod: IrModule): IrModule { return true; }, expr: (e) => { - if (e.kind !== "varRef" && "localId" in e && typeof e.localId === "string") - write(e.localId, null); + if (e.kind === "varRef") { + if (candidates.has(e.localId)) { + if (read === undefined) read = new Set(); + read.add(e.localId); + } + } else if ("localId" in e && typeof e.localId === "string") write(e.localId, null); return true; }, }); + if (read !== undefined) readsByFunction.set(fn, read); + } const constants = new Map(); for (const [id, value] of values) if (writes.get(id) === 1) constants.set(id, value); if (constants.size === 0) return mod; - const reads = (e: IrExpr): boolean => !(e.kind === "varRef" && constants.has(e.localId)); const rewriteExpr = (e: IrExpr): IrExpr => { if (e.kind === "varRef") { const value = constants.get(e.localId); @@ -68,7 +77,8 @@ export function foldConstantGlobals(mod: IrModule): IrModule { const rewriteStmt = (s: IrStmt): IrStmt => mapStmtChildren(s, rewriteExpr, rewriteStmt); let changed = false; const functions = mod.functions.map((fn): IrFunction => { - if (everyStmtList(fn.body, { stmt: () => true, expr: reads })) return fn; + const read = readsByFunction.get(fn); + if (read === undefined || ![...read].some((id) => constants.has(id))) return fn; changed = true; return { ...fn, body: fn.body.map(rewriteStmt) }; }); diff --git a/packages/compiler/src/ir/int32-slots.test.ts b/packages/compiler/src/ir/int32-slots.test.ts index 0739bbb4f..b94dee341 100644 --- a/packages/compiler/src/ir/int32-slots.test.ts +++ b/packages/compiler/src/ir/int32-slots.test.ts @@ -523,6 +523,12 @@ test("counting-down recursion proves its parameter from the call-site ranges", ( ); const rising = analyzeInt32Slots(mod([], [up, fn("user", [], [stmt(call("up", [num(-18)]))])])); expect(rising.facts("up").params?.get("d")).toEqual({ min: -18, max: 0 }); + // The same with the caller first, so the fixpoint analyzes it after the + // parameter dropped: the narrowing still sees its argument. + const callerFirst = analyzeInt32Slots( + mod([], [fn("user", [], [stmt(call("up", [num(-18)]))]), up]), + ); + expect(callerFirst.facts("up").params?.get("d")).toEqual({ min: -18, max: 0 }); // A negative start counts down past the smallest int32. const negative = analyzeInt32Slots(mod([], [count, even, odd, user([18, -1])])); diff --git a/packages/compiler/src/ir/int32-slots.ts b/packages/compiler/src/ir/int32-slots.ts index 06a173c1a..1616a2a77 100644 --- a/packages/compiler/src/ir/int32-slots.ts +++ b/packages/compiler/src/ir/int32-slots.ts @@ -1208,8 +1208,11 @@ export function analyzeInt32Slots(mod: IrModule, inputs: Int32SlotInputs = {}): // Each function's parameter and result writes under its latest analysis: // at the fixpoint, the ranges of every write under the final facts. const lastWrites = new Map>(); - while (queue.length) { - const fn = functions.get(queue.pop()!)!; + // First in, first out: a function whose slots change while others wait + // goes after them, so a caller of many callees is analyzed again once + // for all of their changes, not once for each. + for (let head = 0; head < queue.length; head++) { + const fn = functions.get(queue[head]!)!; queued.delete(fn.name); const params = slots.facts(fn.name).params; if (params) for (const id of params.keys()) reads(`param:${fn.name}\0${id}`, fn.name); @@ -1663,8 +1666,14 @@ function narrowSlotRanges( // The ranges every function was last analyzed under. let analyzed: Map = new Map(proven); let verified: Map = new Map(proven); - // Functions whose recorded writes predate their current facts. + // Functions whose recorded writes predate their current facts: those + // whose parameters rangedParameters proved, and their callers, which + // wrote nothing into those parameters while they had no fact. const dirty = new Set(changed); + for (const name of changed) { + const callers = readersOf(`ret:${name}`); + if (callers) for (const caller of callers) dirty.add(caller); + } // The narrowing operator: a bound moves at most once from its proven // value (a revived result's from int32), so recursion that shifts its // argument (`f(d - 1)`) cannot walk a bound down one step per round. diff --git a/packages/compiler/src/ir/integer-loop-facts.ts b/packages/compiler/src/ir/integer-loop-facts.ts index 8807212dc..d66153b19 100644 --- a/packages/compiler/src/ir/integer-loop-facts.ts +++ b/packages/compiler/src/ir/integer-loop-facts.ts @@ -152,7 +152,7 @@ export function boundedIntegerLoopFacts( e.receiver.kind === "varRef" && stable(e.receiver.localId) ) - return { min: 0, max: Number.MAX_SAFE_INTEGER }; + return BYTES_LENGTH_RANGE; // Every evaluation of an array length lies in its range, so the // condition bounds each pass even when the body resizes the array. if (isArrayLength(e)) return ARRAY_LENGTH_RANGE; @@ -225,6 +225,12 @@ export function boundedIntegerLoopFacts( return result; } +/** Every length, byteLength and byteOffset of a typed array, DataView or + * Buffer: each counts elements or bytes of one allocation, and every + * supported target addresses less than 2^48 bytes. Counters and offsets + * derived from them then stay exact integers far below 2^53. */ +export const BYTES_LENGTH_RANGE: IntegerRange = { min: 0, max: 2 ** 48 }; + /** Every length an array can have: SCR_ARR_MAX_LENGTH (scr_array.c). */ export const ARRAY_LENGTH_RANGE: IntegerRange = { min: 0, max: 4294967295 }; @@ -353,6 +359,13 @@ export function guardedLoopCounterFacts( stable(e.receiver.localId) ) return { min: 0, max: STRING_MAX_UNITS }; + if ( + e.kind === "bytesIntrinsic" && + (e.method === "length" || e.method === "byteLength") && + e.receiver.kind === "varRef" && + stable(e.receiver.localId) + ) + return BYTES_LENGTH_RANGE; // An array length can change on any pass, but never leaves its range, // which is all a ceiling needs. if (isArrayLength(e)) return ARRAY_LENGTH_RANGE; @@ -402,9 +415,84 @@ export function guardedLoopCounterFacts( if (Number.isSafeInteger(start.min) && Number.isSafeInteger(max)) result.set(id, { min: start.min, max }); } + // Companion counters: a binding the loop changes only by positive literal + // increments that no condition bounds (`out[k++] = ...` beside + // `while (i < n)`). A bounded counter that every pass advances before it + // can continue limits the number of passes, and each pass adds at most + // the companion's increments in the body. Increments inside nested loops + // can repeat within a pass, so those companions are not bounded. + let passes = Infinity; + for (const [id, range] of result) { + const hi = ceiling.get(id)!; + const step = guaranteedStep(loop.body, id); + if (step <= 0) continue; + passes = Math.min(passes, Math.floor((hi - range.min) / step) + 1); + } + if (!Number.isSafeInteger(passes) || passes < 0) return result; + const repeated = new Set(); + for (const inner of nested) for (const id of writtenLocals(inner.body)) repeated.add(id); + for (const [id, total] of budget) { + const local = locals.get(id), + start = initial.get(id); + if (result.has(id) || refused.has(id) || repeated.has(id) || ceiling.has(id)) continue; + if (!local || local.boxed || local.tdz || !start) continue; + const max = start.max + passes * total; + if (Number.isSafeInteger(start.min) && Number.isSafeInteger(max)) + result.set(id, { min: start.min, max }); + } return result; } +/** The increments of counter `id` that every pass of `body` makes before it + * can start another pass: those of the leading statements up to the first + * one that can continue the loop or holds a nested region (0 when there is + * none). A pass that leaves the loop earlier starts no further pass, so the + * count bounds the passes that continue. */ +function guaranteedStep(body: readonly IrStmt[], id: string): number { + let step = 0; + for (const s of body) { + if (s.kind === "assign" && s.localId === id) { + const delta = increment(id, s.value); + if (delta === null || delta <= 0) return 0; + step += delta; + continue; + } + if (s.kind === "exprStmt" && s.expr.kind === "incDec" && s.expr.localId === id) { + if (s.expr.op !== "+") return 0; + step += 1; + continue; + } + if (mayContinue(s)) break; + } + return step; +} + +/** Whether `s` can start another pass of the enclosing loop before its + * following statements run: it holds a `continue`, a loop or another + * region whose control flow is not walked here. */ +function mayContinue(s: IrStmt): boolean { + let found = false; + const stmt = (x: IrStmt): boolean => { + if ( + x.kind === "continue" || + x.kind === "for" || + x.kind === "while" || + x.kind === "doWhile" || + x.kind === "forOf" || + x.kind === "tryCatch" || + x.kind === "switch" || + (x.kind === "block" && !!x.labels?.length) + ) { + found = true; + return false; + } + return everyStmtChild(x, expr, stmt); + }; + const expr = (e: IrExpr): boolean => everyExprChild(e, expr, stmt); + stmt(s); + return found; +} + /** The largest value counter `id` takes in one pass of `body`, which starts * at most at `entry`. The caller has checked that the body changes the * counter only by positive literal increments outside switch, exception and diff --git a/packages/compiler/src/ir/integer-ranges.test.ts b/packages/compiler/src/ir/integer-ranges.test.ts index 5ae04d28f..318ff2c09 100644 --- a/packages/compiler/src/ir/integer-ranges.test.ts +++ b/packages/compiler/src/ir/integer-ranges.test.ts @@ -828,6 +828,72 @@ test("a while condition on a string's length bounds a counter, and guarded incre expect(run((inside) => [observe(inside), bump, rebound]).after).toBeNull(); }); +test("a while counter every pass advances bounds the companions the loop increments", () => { + // pos = 0; k = 0; while (pos < text.length) { pos++; observe(k); if (unknown === 1) k++; k++; } observe(k) + const text: IrExpr = { loc, kind: "varRef", localId: "text", type: STRING }; + const length: IrExpr = { + loc, + kind: "strIntrinsic", + method: "length", + receiver: text, + args: [], + type: F64, + }; + const set = (localId: string, value: IrExpr): IrStmt => ({ loc, kind: "assign", localId, value }); + const bump = (localId: string): IrStmt => ({ + loc, + kind: "exprStmt", + expr: { loc, kind: "incDec", op: "+", prefix: false, localId, type: F64 }, + }); + const maybe = (then: IrStmt[]): IrStmt => ({ + loc, + kind: "if", + cond: condition("===", ref("unknown"), num(1)), + then, + else_: null, + }); + const run = (body: (inside: IrExpr) => IrStmt[]) => { + const inside = ref("k"), + after = ref("k"); + const loop: IrStmt = { + loc, + kind: "while", + cond: condition("<", ref("pos"), length), + body: body(inside), + }; + const f = fn([set("pos", num(0)), set("k", num(0)), loop, observe(after)]); + f.locals.push( + { id: "pos", name: "pos", type: F64, mutable: true }, + { id: "k", name: "k", type: F64, mutable: true }, + { id: "text", name: "text", type: STRING, mutable: false }, + ); + const facts = analyzeIntegerRangeFacts(f); + return { + inside: facts.ranges.get(inside), + after: facts.ranges.get(after), + loop: facts.loopRanges.get(loop)?.get("k"), + }; + }; + // At most 2^31 - 1 passes, each adding at most 2. + const bounded = { min: 0, max: 4294967294 }; + expect(run((inside) => [bump("pos"), observe(inside), maybe([bump("k")]), bump("k")])).toEqual({ + inside: bounded, + after: bounded, + loop: bounded, + }); + // A pass that can continue before advancing the counter, or an increment a + // nested loop repeats, bounds nothing. + const skip: IrStmt = maybe([{ loc, kind: "continue" }]); + expect(run((inside) => [skip, bump("pos"), observe(inside), bump("k")]).after).toBeNull(); + const repeat: IrStmt = { + loc, + kind: "while", + cond: condition("===", ref("unknown"), num(1)), + body: [bump("k")], + }; + expect(run((inside) => [bump("pos"), observe(inside), repeat]).after).toBeNull(); +}); + test("a nested while loop may advance a counter its own condition bounds", () => { // while (pos < text.length) { observe(pos); while (pos < text.length && unknown === 1) pos++; pos += 2; } const text: IrExpr = { loc, kind: "varRef", localId: "text", type: STRING }; @@ -1323,7 +1389,7 @@ test("a for loop bounds cursors it resets and copies, and their writes keep the } }); -test("a cursor of a loop over a typed array keeps its writes exact up to 2^53 - 1", () => { +test("a cursor of a loop over a typed array keeps its writes within the largest length", () => { // for (x = 0; x < bytes.length; x++) changes = changes + 1; const bytes: IrExpr = { loc, @@ -1355,7 +1421,7 @@ test("a cursor of a loop over a typed array keeps its writes exact up to 2^53 - { id: "changes", name: "changes", type: F64, mutable: true }, { id: "bytes", name: "bytes", type: { kind: "bytes", elem: "u8" }, mutable: false }, ); - // The sum of the header range passes 2^53 - 1; the write takes the loop - // range instead. - expect(analyzeIntegerRanges(f).get(counted)).toEqual({ min: 0, max: Number.MAX_SAFE_INTEGER }); + // The sum of the header range passes the largest typed array length; the + // write takes the loop range instead. + expect(analyzeIntegerRanges(f).get(counted)).toEqual({ min: 1, max: 2 ** 48 }); }); diff --git a/packages/compiler/src/ir/integer-ranges.ts b/packages/compiler/src/ir/integer-ranges.ts index 683ceab75..1909040c4 100644 --- a/packages/compiler/src/ir/integer-ranges.ts +++ b/packages/compiler/src/ir/integer-ranges.ts @@ -2,6 +2,7 @@ import type { IrExpr, IrFunction, IrStmt } from "./ir.js"; import { everyExprChild, everyStmtChild, everyStmtList } from "./traverse.js"; import { byteNumberAccess, byteNumberRange } from "./byte-numbers.js"; import { + BYTES_LENGTH_RANGE, boundedIntegerLoopFacts, guardedLoopCounterFacts, integerPathFallsThrough, @@ -821,7 +822,7 @@ export function analyzeIntegerRangeFacts( const numeric = byteNumberAccess(e); if (numeric && !numeric.write) range = byteNumberRange(numeric); else if (e.method === "length" || e.method === "byteLength" || e.method === "byteOffset") - range = { min: 0, max: Number.MAX_SAFE_INTEGER }; + range = BYTES_LENGTH_RANGE; else if (e.method === "get" && e.invalidNaN !== true && e.receiver.type.kind === "bytes") { const elem = e.receiver.type.elem; if (elem === "u8" || elem === "u8c") range = { min: 0, max: 255 }; @@ -1075,6 +1076,19 @@ export function analyzeIntegerRangeFacts( // inside and after the loop. const guarded = s.kind === "while" ? guardedLoopCounterFacts(s, loopLocals, facts) : noInvariants; + // The same bounds hold on entry and at every write in the loop, + // which is what a proven integer loop needs (loopRanges). A loop + // analyzed again keeps the bindings every analysis bounds. + if (s.kind === "while") { + const previous = loopRanges.get(s); + if (!previous) loopRanges.set(s, new Map(guarded)); + else + for (const [id, range] of previous) { + const other = guarded.get(id); + if (other) previous.set(id, union(range, other)); + else previous.delete(id); + } + } const loop = new Map(facts); havocWith(loop, [s], loop, guarded); for (const [id, range] of guarded) set(loop, id, range); diff --git a/packages/compiler/src/ir/js-ic.test.ts b/packages/compiler/src/ir/js-ic.test.ts index f71a1e9db..f5391b83e 100644 --- a/packages/compiler/src/ir/js-ic.test.ts +++ b/packages/compiler/src/ir/js-ic.test.ts @@ -126,6 +126,140 @@ test("property helpers whose class branches only test capsules are cache sites", expect(sites.site("%dyn.class.property.1")).toEqual({ kind: "set", key: "y" }); }); +test("bags converted between dyn and jsval are still the receiver", () => { + // A program that declares a class returns its bags as jsvals, and the + // helpers read them through js.toDyn. + const toDyn = (value: IrExpr): IrExpr => ({ + kind: "libCall", + fn: "js.toDyn", + args: [value], + type: DYN, + loc, + }); + const fromDyn = (value: IrExpr): IrExpr => ({ + kind: "libCall", + fn: "js.fromDyn", + args: [value], + type: { kind: "jsval" }, + loc, + }); + const jsvalBag = helper( + "%dyn.class.readPropertiesOrEmpty.js", + ["v"], + [capsuleBranch("v"), ret(fromDyn(ref("v")))], + ); + const staticBag = helper("%dyn.class.staticProperties", ["v"], [ret(ref("v"))]); + const getter = getHelper( + "%dyn.class.property.0", + "x", + [capsuleBranch("p.0")], + call(staticBag.name, toDyn(call(jsvalBag.name, ref("p.0")))), + ); + const setter = setHelper("%dyn.class.property.1", "y"); + const write = setter.body[0] as Extract; + const keySet = write.expr as Extract; + keySet.args[0] = toDyn(call(jsvalBag.name, ref("p.0"))); + // A converted value that is not the receiver stays a helper call. + const foreign = getHelper( + "%dyn.class.property.2", + "x", + [], + toDyn(fromDyn({ kind: "dynObjLit", type: DYN, loc } as IrExpr)), + ); + const sites = sitesOf(jsvalBag, staticBag, getter, setter, foreign); + expect(sites.site("%dyn.class.property.0")).toEqual({ kind: "get", key: "x" }); + expect(sites.site("%dyn.class.property.1")).toEqual({ kind: "set", key: "y" }); + expect(sites.site("%dyn.class.property.2")).toBeNull(); +}); + +test("a bag helper may first rebind its parameter to a bag of it", () => { + // Programs with subclasses normalize the value before the class branches. + const normalize = helper("%dyn.class.normalize", ["v"], [capsuleBranch("v"), ret(ref("v"))]); + const rebind = (value: IrExpr): IrStmt => ({ kind: "assign", localId: "v", value, loc }); + const normalized = helper( + "%dyn.class.readPropertiesOrEmpty.normalized", + ["v"], + [rebind(call(normalize.name, ref("v"))), capsuleBranch("v"), ret(ref("v"))], + ); + const replaced = helper( + "%dyn.class.readPropertiesOrEmpty.replaced", + ["v"], + [rebind({ kind: "dynObjLit", type: DYN, loc } as IrExpr), ret(ref("v"))], + ); + const sites = sitesOf( + normalize, + normalized, + replaced, + getHelper("%dyn.class.property.0", "x", [], call(normalized.name, ref("p.0"))), + getHelper("%dyn.class.property.1", "x", [], call(replaced.name, ref("p.0"))), + // The property helper itself normalizes its receiver first, too. + getHelper("%dyn.class.property.2", "x", [ + { kind: "assign", localId: "p.0", value: call(normalize.name, ref("p.0")), loc }, + capsuleBranch("p.0"), + ]), + ); + expect(sites.site("%dyn.class.property.0")).toEqual({ kind: "get", key: "x" }); + expect(sites.site("%dyn.class.property.1")).toBeNull(); + expect(sites.site("%dyn.class.property.2")).toEqual({ kind: "get", key: "x" }); +}); + +test("keyed helpers may test capsules between the key's coercion and the read", () => { + const read = (object: IrExpr, key: string): IrExpr => ({ + kind: "libCall", + fn: "js.fromDyn", + args: [{ kind: "dynKeyGet", value: object, key: ref(key), type: DYN, loc }], + type: { kind: "jsval" }, + loc, + }); + const guard = (test: "symbol" | "nullish", tested: string, object: IrExpr): IrStmt => ({ + kind: "if", + cond: { kind: "dynTest", test, value: ref(tested), type: BOOL, loc }, + then: [ret(read(object, "p.key"))], + else_: null, + loc, + }); + const keyed = (name: string, between: IrStmt[]): IrFunction => + helper( + name, + ["p.0", "p.key"], + [ + capsuleBranch("p.0"), + guard("symbol", "p.key", call(identityBag.name, ref("p.0"))), + guard("nullish", "p.0", ref("p.0")), + { + kind: "varDecl", + localId: "key.string", + init: { + kind: "libCall", + fn: "dyn.toStringCoerce", + args: [ref("p.key")], + type: STRING, + loc, + }, + loc, + }, + ...between, + ret(read(call(identityBag.name, ref("p.0")), "key.string")), + ], + ["key.string"], + ); + const errorBranch: IrStmt = { + kind: "if", + cond: { kind: "dynTest", test: "error", value: ref("p.0"), type: BOOL, loc }, + then: [ret(ref("p.0"))], + else_: null, + loc, + }; + const sites = sitesOf( + keyed("%dyn.class.computed.0", []), + keyed("%dyn.class.computed.1", [capsuleBranch("p.0"), capsuleBranch("p.0", "dyn.classIs")]), + keyed("%dyn.class.computed.2", [errorBranch]), + ); + expect(sites.site("%dyn.class.computed.0")).toEqual({ kind: "keyed" }); + expect(sites.site("%dyn.class.computed.1")).toEqual({ kind: "keyed" }); + expect(sites.site("%dyn.class.computed.2")).toBeNull(); +}); + test("any other prefix or tail keeps the helper call", () => { const errorBranch: IrStmt = { kind: "if", @@ -237,3 +371,57 @@ test("bag helpers that exit early for values no class branch matches are still t expect(sites.site("%dyn.class.property.2")).toBeNull(); expect(sites.site("%dyn.class.property.3")).toBeNull(); }); + +test("helpers that answer plain values first with the same operation are cache sites", () => { + const negated = (test: "classCapsule" | "funcKind"): IrExpr => ({ + kind: "dynTest", + test, + negated: true, + value: ref("p.0"), + type: BOOL, + loc, + }); + const exit = (body: IrStmt[], test: IrExpr = negated("classCapsule")): IrStmt => ({ + kind: "if", + cond: test, + then: [{ kind: "if", cond: negated("funcKind"), then: body, else_: null, loc }], + else_: null, + loc, + }); + const store = (key: string): IrStmt[] => [ + { + kind: "exprStmt", + expr: { + kind: "libCall", + fn: "dyn.keySet", + args: [ref("p.0"), str(key), ref("p.1")], + type: VOID, + loc, + }, + loc, + }, + ret(null), + ]; + const load = (key: string): IrStmt[] => [ + ret({ kind: "dynKeyGet", value: ref("p.0"), key: str(key), type: DYN, loc }), + ]; + const sites = sitesOf( + setHelper("%dyn.class.property.0", "x", [exit(store("x")), capsuleBranch("p.0")]), + getHelper("%dyn.class.property.1", "x", [exit(load("x")), capsuleBranch("p.0")]), + // An exit for another key, or one plain values do not take. + setHelper("%dyn.class.property.2", "x", [exit(store("y"))]), + setHelper("%dyn.class.property.3", "x", [ + exit(store("x"), { + kind: "dynTest", + test: "classCapsule", + value: ref("p.0"), + type: BOOL, + loc, + }), + ]), + ); + expect(sites.site("%dyn.class.property.0")).toEqual({ kind: "set", key: "x" }); + expect(sites.site("%dyn.class.property.1")).toEqual({ kind: "get", key: "x" }); + expect(sites.site("%dyn.class.property.2")).toBeNull(); + expect(sites.site("%dyn.class.property.3")).toBeNull(); +}); diff --git a/packages/compiler/src/ir/js-ic.ts b/packages/compiler/src/ir/js-ic.ts index 752316db3..48019636a 100644 --- a/packages/compiler/src/ir/js-ic.ts +++ b/packages/compiler/src/ir/js-ic.ts @@ -61,13 +61,68 @@ export class JsIcSites { if (fn.params.some((p) => p.type.kind !== "dyn")) return null; const receiver = fn.params[0]?.localId; if (receiver === undefined) return null; + const second = fn.params[1]?.localId; + const exits: JsIcSite[] = []; let i = 0; - while (i < fn.body.length && this.capsuleBranch(fn.body[i]!, receiver)) i++; + for (; i < fn.body.length; i++) { + const s = fn.body[i]!; + if (this.capsuleBranch(s, receiver) || this.rebinding(s, receiver)) continue; + const exit = this.plainExit(s, receiver, second); + if (exit === null) break; + exits.push(exit); + } const tail = fn.body.slice(i); - if (fn.params.length === 1) return this.getTail(tail, receiver); - if (fn.params.length !== 2) return null; - const second = fn.params[1]!.localId; - return this.setTail(tail, receiver, second) ?? this.keyedTail(tail, receiver, second); + let site: JsIcSite | null; + if (fn.params.length === 1) site = this.getTail(tail, receiver); + else if (fn.params.length !== 2 || second === undefined) return null; + else site = this.setTail(tail, receiver, second) ?? this.keyedTail(tail, receiver, second); + if (site === null || site.kind === "keyed") return exits.length === 0 ? site : null; + // An exit answers plain values with the operation the site caches. + return exits.every((exit) => exit.kind === site.kind && exit.key === site.key) ? site : null; + } + + /** `if (!classCapsule(receiver)) { [if (!funcKind(receiver))] OP }`, no + * else: an exit that runs the helper's generic operation directly on a + * value no class table matches. OP is `return [js.fromDyn](dynKeyGet(R, + * K))` or `dyn.keySet(R, K, value); return`, with R the receiver (or a + * bag of it). */ + private plainExit(s: IrStmt, receiver: string, value: string | undefined): JsIcSite | null { + const negated = (e: IrExpr, test: "classCapsule" | "funcKind"): boolean => + e.kind === "dynTest" && e.test === test && e.negated === true && isVar(e.value, receiver); + if (s.kind !== "if" || s.else_ !== null || !negated(s.cond, "classCapsule")) return null; + let body = s.then; + const inner = body[0]; + if ( + body.length === 1 && + inner?.kind === "if" && + inner.else_ === null && + negated(inner.cond, "funcKind") + ) + body = inner.then; + const [first, ret] = body; + if (body.length === 1 && first?.kind === "return") { + const read = unwrapFromDyn(first.value); + if ( + read?.kind !== "dynKeyGet" || + read.key.kind !== "strLit" || + !this.bagOf(read.value, receiver) + ) + return null; + return { kind: "get", key: read.key.value }; + } + if ( + body.length !== 2 || + first?.kind !== "exprStmt" || + ret?.kind !== "return" || + ret.value !== null + ) + return null; + const call = first.expr; + if (call.kind !== "libCall" || call.fn !== "dyn.keySet" || call.args.length !== 3) return null; + const [bag, key, stored] = call.args; + if (!bag || !this.bagOf(bag, receiver) || key?.kind !== "strLit") return null; + if (value === undefined || !isVar(stored, value)) return null; + return { kind: "set", key: key.value }; } /** `varDecl bag = BAG(receiver); return [js.fromDyn](bag === receiver ? @@ -109,10 +164,14 @@ export class JsIcSites { /** `if (typeof key === "symbol") return dynKeyGet(BAG(receiver), key); * if (receiver == null) return dynKeyGet(receiver, key); * const s = String(key); return dynKeyGet(BAG(receiver), s)`, each read - * possibly under js.fromDyn. */ + * possibly under js.fromDyn. Capsule branches may precede the final read: + * a program that declares classes looks the string key up in each class's + * key table there. */ private keyedTail(tail: IrStmt[], receiver: string, key: string): JsIcSite | null { - const [symbol, nullish, decl, ret] = tail; - if (tail.length !== 4) return null; + const [symbol, nullish, decl] = tail; + const ret = tail[tail.length - 1]; + if (tail.length < 4) return null; + if (!tail.slice(3, -1).every((s) => this.capsuleBranch(s, receiver))) return null; if (!this.guard(symbol, "symbol", key, (read) => this.bagOf(read, receiver), key)) return null; if (!this.guard(nullish, "nullish", receiver, (read) => isVar(read, receiver), key)) return null; @@ -159,9 +218,12 @@ export class JsIcSites { } /** Whether `e` is the receiver itself for every ordinary object or array: - * the receiver, or a chain of property-bag helpers applied to it. */ + * the receiver, or a chain of property-bag helpers applied to it. A + * conversion between the dyn and jsval forms keeps the object it holds. */ private bagOf(e: IrExpr, receiver: string): boolean { if (isVar(e, receiver)) return true; + if (e.kind === "libCall" && (e.fn === "js.toDyn" || e.fn === "js.fromDyn")) + return e.args.length === 1 && this.bagOf(e.args[0]!, receiver); return ( e.kind === "call" && e.args.length === 1 && @@ -182,6 +244,13 @@ export class JsIcSites { return ret.kind === "return" && ret.value !== null && this.bagOf(ret.value, param); } + /** `param = `, such as the subclass normalization + * `value = %dyn.class.normalize(value)`, which leaves an ordinary object + * or array in `param`. */ + private rebinding(s: IrStmt, param: string): boolean { + return s.kind === "assign" && s.localId === param && this.bagOf(s.value, param); + } + /** A property-bag helper that returns its argument unless a capsule * branch is taken. */ private identityHelper(name: string): boolean { @@ -204,7 +273,10 @@ export class JsIcSites { identity = fn.body .slice(0, -1) - .every((s) => this.capsuleBranch(s, param) || this.kindExit(s, param)) && + .every( + (s) => + this.capsuleBranch(s, param) || this.kindExit(s, param) || this.rebinding(s, param), + ) && last?.kind === "return" && last.value !== null && this.bagOf(last.value, param); diff --git a/packages/compiler/src/ir/receiver-readers.ts b/packages/compiler/src/ir/receiver-readers.ts index 3d0e6f35d..f5270aa7d 100644 --- a/packages/compiler/src/ir/receiver-readers.ts +++ b/packages/compiler/src/ir/receiver-readers.ts @@ -1,6 +1,6 @@ import { isBorrowSafeMath } from "./analysis.js"; import type { IrExpr, IrFunction, IrLibFn, IrStmt } from "./ir.js"; -import { everyStmtList } from "./traverse.js"; +import { eachNodeOfKinds, type IrNodeIndex } from "./node-index.js"; /** Expression kinds that run no user code of their own and never consult * the ambient receiver stack. Calls are admitted here and resolved through @@ -85,6 +85,25 @@ const RECEIVER_FREE_STMTS = new Set([ "runtimeFence", ]); +/** The kinds whose nodes the readers walk answers true for without looking + * at them: every receiver-free statement, and the receiver-free expressions + * other than calls, library calls, string conversions and data intrinsics. */ +const TRIVIALLY_RECEIVER_FREE = new Set([ + ...RECEIVER_FREE_STMTS, + ...[...RECEIVER_FREE_EXPRS].filter( + (kind) => + kind !== "call" && + kind !== "new" && + kind !== "virtualCall" && + kind !== "libCall" && + kind !== "toString" && + kind !== "arrIntrinsic" && + kind !== "strIntrinsic" && + kind !== "mapIntrinsic" && + kind !== "setIntrinsic", + ), +]); + /** Runtime library calls beyond scalar Math that only compute or build and * park a native error: the f64[] folds behind Math.max/min(...xs), * `new Error(message)` (scr_error_new), and the checked-read TypeError @@ -133,7 +152,11 @@ function opaqueOperand(e: IrExpr): boolean { export class AmbientReceiverReaders { private readers: Set | null = null; - constructor(private readonly functions: ReadonlyMap) {} + constructor( + private readonly functions: ReadonlyMap, + /** The module's node index, read instead of walking the bodies it holds. */ + private readonly nodes?: IrNodeIndex, + ) {} ignores(fn: IrFunction): boolean { if (fn.async || fn.generator) return false; @@ -153,6 +176,7 @@ export class AmbientReceiverReaders { } const readers = new Set(); const callers = new Map(); + const scan = this.nodes?.scanExcept(TRIVIALLY_RECEIVER_FREE); for (const fn of this.functions.values()) { const callees = new Set(); let reads = fn.async === true || fn.generator !== undefined; @@ -161,32 +185,36 @@ export class AmbientReceiverReaders { callees.add(name); return true; }; + const expr = (expr: IrExpr): boolean => { + if (!RECEIVER_FREE_EXPRS.has(expr.kind)) return false; + switch (expr.kind) { + case "call": + return edge(expr.callee); + case "new": + return edge(`%${expr.className}.constructor`); + case "virtualCall": { + // Any class's override may run: admit only when every + // same-named method is itself receiver-free. + const targets = byMethod.get(expr.method) ?? []; + return targets.length > 0 && targets.every(edge); + } + case "libCall": + return isBorrowSafeMath(expr.fn) || RECEIVER_FREE_LIB_FNS.has(expr.fn); + case "toString": + return !opaqueOperand(expr.operand); + default: { + const operands = dataIntrinsicOperands(expr); + return operands === null || !operands.some(opaqueOperand); + } + } + }; if (!reads) - reads = !everyStmtList(fn.body, { - stmt: (stmt) => RECEIVER_FREE_STMTS.has(stmt.kind), - expr: (expr) => { - if (!RECEIVER_FREE_EXPRS.has(expr.kind)) return false; - switch (expr.kind) { - case "call": - return edge(expr.callee); - case "new": - return edge(`%${expr.className}.constructor`); - case "virtualCall": { - // Any class's override may run: admit only when every - // same-named method is itself receiver-free. - const targets = byMethod.get(expr.method) ?? []; - return targets.length > 0 && targets.every(edge); - } - case "libCall": - return isBorrowSafeMath(expr.fn) || RECEIVER_FREE_LIB_FNS.has(expr.fn); - case "toString": - return !opaqueOperand(expr.operand); - default: { - const operands = dataIntrinsicOperands(expr); - return operands === null || !operands.some(opaqueOperand); - } - } - }, + eachNodeOfKinds(fn, scan, (node, statement) => { + const free = statement + ? RECEIVER_FREE_STMTS.has((node as IrStmt).kind) + : expr(node as IrExpr); + if (!free) reads = true; + return free; }); if (reads) readers.add(fn.name); for (const callee of callees) { diff --git a/packages/compiler/src/ir/runtime-features.ts b/packages/compiler/src/ir/runtime-features.ts index 118f0ef38..85b21af7b 100644 --- a/packages/compiler/src/ir/runtime-features.ts +++ b/packages/compiler/src/ir/runtime-features.ts @@ -715,6 +715,10 @@ interface RuntimeFeatureRule { /** Gives every rule in the table one type. */ const rule = (r: RuntimeFeatureRule): RuntimeFeatureRule => r; +/* The node:url and WHATWG URL entry points: every URL value comes from + * one, and parsing one can reach the UTS #46 host processing. */ +const URL_PREFIXES: readonly string[] = ["url."]; + /* The HTTP, HTTPS, HTTP/2 and TLS surfaces, which link the net unit. */ const NET_PREFIXES: readonly string[] = ["net.", "http.", "tls.", "https.", "http2."]; @@ -1063,9 +1067,10 @@ const RUNTIME_FEATURE_RULES = { /** scr_url_legacy.c parses and stringifies queries. */ qs: rule({ fns: [...QS_LIB_FNS, ...LEGACY_URL_LIB_FNS] }), /** scr_url_legacy.c lowercases and normalizes hostnames with the regex - * unit's libunicode, NODE_DEBUG matching in scr_util_callable.c and - * readline's terminal unit use the regex unit, and locale case mapping - * runs its default conversion. */ + * unit's libunicode, the WHATWG URL host parser runs UTS #46 through + * scr_idna.c (which links with the regex unit), NODE_DEBUG matching in + * scr_util_callable.c and readline's terminal unit use the regex unit, + * and locale case mapping runs its default conversion. */ regex: rule({ fns: [ ...REGEX_LIB_FNS, @@ -1074,13 +1079,17 @@ const RUNTIME_FEATURE_RULES = { "rl.terminal", "str.localeCase", ], + prefixes: URL_PREFIXES, expr: (node, _tables) => node.kind === "regexLit" || node.kind === "regexIntrinsic" || (node.kind === "strIntrinsic" && CASE_MAPPING_STR_METHODS.has(node.method)), - types: ["regex"], + types: ["regex", "url"], }), searchParams: rule({ fns: ["url.searchParams"], prefixes: ["sp."], types: ["searchParams"] }), + /** URL programs install the IDNA unit's ToASCII for the WHATWG host + * parser (scr_url_idna_install). */ + urlIdna: rule({ prefixes: URL_PREFIXES, types: ["url"] }), /** stdin setEncoding / string data listeners / stdio 'resize' events. A * typed string 'data' listener takes setEncoding-decoded chunks; the * process stream object's setEncoding and 'resize' members are reached @@ -1177,6 +1186,7 @@ function noRuntimeFeatures(): RuntimeFeatures { qs: false, regex: false, searchParams: false, + urlIdna: false, stdioExtended: false, stream: false, streamDyn: false, diff --git a/packages/compiler/src/ir/self-captures.ts b/packages/compiler/src/ir/self-captures.ts index e009e9a11..74f634d4a 100644 --- a/packages/compiler/src/ir/self-captures.ts +++ b/packages/compiler/src/ir/self-captures.ts @@ -118,7 +118,12 @@ export function resolveSelfCaptures(mod: IrModule): IrModule { } if (resolved.size === 0) return mod; + // Only the resolved bodies and the functions that create their closures + // change. + const touched = new Set(resolved.keys()); + for (const name of resolved.keys()) for (const site of sites.get(name)!) touched.add(site.owner); const rewritten = mod.functions.map((fn) => { + if (!touched.has(fn.name)) return fn; const own = resolved.get(fn.name); const self = own === undefined ? null : fn.captures![own]!; let changed = false; diff --git a/packages/compiler/src/native/cache.ts b/packages/compiler/src/native/cache.ts index 937b4b109..712c1c232 100644 --- a/packages/compiler/src/native/cache.ts +++ b/packages/compiler/src/native/cache.ts @@ -118,22 +118,11 @@ export class NativeCache { } } - /** Whether a marker or entry exists, without reading or checking it. */ + /** Whether an entry exists, without reading or checking it. */ has(family: string, key: string): boolean { return existsSync(join(this.root, family, key)); } - /** Records an empty marker that has() finds. */ - mark(family: string, key: string): void { - try { - const directory = join(this.root, family); - mkdirSync(directory, { recursive: true, mode: 0o700 }); - writeFileSync(join(directory, key), ""); - } catch { - /* Cache failures never change a build's result. */ - } - } - /** Writes `bytes` under their content digest, and returns the digest. */ writeContent(family: string, bytes: Uint8Array): string { const digest = contentDigest(bytes); @@ -178,7 +167,7 @@ export class NativeCache { "object", "sanitizer", "executable", - "linked", + "kept", "binary", "dsym", "coverage", diff --git a/packages/compiler/src/native/compiler.ts b/packages/compiler/src/native/compiler.ts index 0d02b135f..320cd48d3 100644 --- a/packages/compiler/src/native/compiler.ts +++ b/packages/compiler/src/native/compiler.ts @@ -408,7 +408,9 @@ export class NativeCompiler { ffi, executablePack, ); - const restored = executableCache?.restore(stagedOutput) ?? false; + // An unchanged rebuild keeps the executable the last build left. + const kept = executableCache?.keep(output) ?? false; + const restored = kept || (executableCache?.restore(stagedOutput) ?? false); timing(restored ? "executable-cache-hit" : "executable-cache-miss"); if (!restored) { let programObjects = [object]; @@ -525,7 +527,10 @@ export class NativeCompiler { irPath = join(options.outDir, `${stem}.ir.json`); writeTextPartsSync(irPath, prepared.ir!); } - renameSync(stagedOutput, output); + if (!kept) { + renameSync(stagedOutput, output); + executableCache?.recordOutput(output); + } if (target.platform === "darwin" && (optimization !== "dev" || options.strip)) rmSync(output + ".dSYM", { recursive: true, force: true }); return { diff --git a/packages/compiler/src/native/executable-cache.test.ts b/packages/compiler/src/native/executable-cache.test.ts index 83a493721..97f9a3571 100644 --- a/packages/compiler/src/native/executable-cache.test.ts +++ b/packages/compiler/src/native/executable-cache.test.ts @@ -224,7 +224,7 @@ test("a failed or empty link trace cannot publish a completed executable", () => }); test.each(["ld", "crt", "libc"] as const)( - "a Linux key's second link records the linker, startup objects and libc; replacing the %s misses", + "a Linux key's first link records the linker, startup objects and libc; replacing the %s misses", (input) => { const f = fixture(); const lib = join(f.root, "lib"); @@ -257,26 +257,46 @@ test.each(["ld", "crt", "libc"] as const)( }; }); const open = () => new NativeExecutableCache(f.cache, f.key, "clang", false, [f.runtime], true); - // A key's first link is not traced; the caller links the output itself. - expect(open().trace(["-o", f.output], f.stage)).toBe(false); - expect(spawnSync).not.toHaveBeenCalled(); - expect(open().restore(f.destination)).toBe(false); - // Its next link traces both the driver and the linker in one run. + // A key's first link traces the driver and the linker in one run, and + // records the output it leaves instead of storing the executable. const entry = open(); expect(entry.trace(["-o", f.output], f.stage)).toBe(true); expect(spawnSync).toHaveBeenCalledTimes(1); expect(vi.mocked(spawnSync).mock.calls[0]![1]).toEqual(["-o", f.output, "-v", "-Wl,-t"]); entry.publish(f.output); + expect(f.cache.read("executable", f.key)).toBeNull(); + entry.recordOutput(f.output); + expect(open().keep(f.output)).toBe(true); const recorded = ( - JSON.parse(f.cache.read("executable", f.key)!.toString()) as { inputs: { path: string }[] } + JSON.parse(f.cache.read("kept", f.key)!.toString()) as { inputs: { path: string }[] } ).inputs.map((saved) => saved.path); expect(recorded).toEqual(expect.arrayContaining([...Object.values(paths), nonshared, lib])); expect(recorded).not.toContain(program); - expect(open().restore(f.destination)).toBe(true); const before = statSync(paths[input]); writeFileSync(paths[input] + ".new", "x".repeat(before.size)); utimesSync(paths[input] + ".new", before.atime, before.mtime); renameSync(paths[input] + ".new", paths[input]); - expect(open().restore(f.destination)).toBe(false); + expect(open().keep(f.output)).toBe(false); }, ); + +test("a Linux build whose output changed links again, and then stores the executable", () => { + const f = fixture(); + const open = () => new NativeExecutableCache(f.cache, f.key, "clang", false, [f.runtime], true); + const first = open(); + expect(first.trace(["-o", f.output], f.stage)).toBe(true); + first.publish(f.output); + first.recordOutput(f.output); + // The output is overwritten: it is not kept, and the next link publishes. + writeFileSync(f.output, Buffer.from([1, 2, 3, 4])); + expect(open().keep(f.output)).toBe(false); + const second = open(); + expect(second.restore(f.destination)).toBe(false); + expect(second.trace(["-o", f.output], f.stage)).toBe(true); + second.publish(f.output); + second.recordOutput(f.output); + expect(open().keep(f.output)).toBe(true); + // The stored executable restores at another path. + expect(open().restore(f.destination)).toBe(true); + expect(readFileSync(f.destination)).toEqual(readFileSync(f.output)); +}); diff --git a/packages/compiler/src/native/executable-cache.ts b/packages/compiler/src/native/executable-cache.ts index 1dfbf0b00..b9058e947 100644 --- a/packages/compiler/src/native/executable-cache.ts +++ b/packages/compiler/src/native/executable-cache.ts @@ -124,6 +124,41 @@ function tracePaths(output: string, driver: boolean, roots: string[]): string[] return [...paths].sort(); } +/** Linux: the output a build left, which an unchanged rebuild keeps while + * the output and every input keep their identities. */ +interface KeptEntry { + schema: "scriptc.native-kept.v1"; + output: InputIdentity; + inputs: InputIdentity[]; +} + +function validIdentity(input: InputIdentity): boolean { + return ( + input !== null && + typeof input === "object" && + typeof input.path === "string" && + isAbsolute(input.path) && + typeof input.canonical === "string" && + (input.kind === "file" || input.kind === "directory") && + [input.dev, input.ino, input.size, input.mtime, input.ctime].every( + (value) => typeof value === "number" && Number.isFinite(value), + ) + ); +} + +function validKept(value: unknown): value is KeptEntry { + if (value === null || typeof value !== "object") return false; + const entry = value as Partial; + return ( + entry.schema === "scriptc.native-kept.v1" && + entry.output !== undefined && + validIdentity(entry.output) && + Array.isArray(entry.inputs) && + entry.inputs.length > 0 && + entry.inputs.every(validIdentity) + ); +} + function validEntry(value: unknown): value is ExecutableEntry { if (value === null || typeof value !== "object") return false; const entry = value as Partial; @@ -135,18 +170,7 @@ function validEntry(value: unknown): value is ExecutableEntry { (typeof entry.symbols === "string" && /^[0-9a-f]{64}$/.test(entry.symbols))) && Array.isArray(entry.inputs) && entry.inputs.length > 0 && - entry.inputs.every( - (input) => - input !== null && - typeof input === "object" && - typeof input.path === "string" && - isAbsolute(input.path) && - typeof input.canonical === "string" && - (input.kind === "file" || input.kind === "directory") && - [input.dev, input.ino, input.size, input.mtime, input.ctime].every( - (value) => typeof value === "number" && Number.isFinite(value), - ), - ) + entry.inputs.every(validIdentity) ); } @@ -178,8 +202,8 @@ function stageSymbols(bytes: Buffer, binary: string): void { export class NativeExecutableCache { private inputs: InputIdentity[]; private traced = false; - /** Linux: no entry exists and the key was never linked before. */ - private readonly firstLink: boolean; + /** The inputs of the entry restore() used. */ + private restoredInputs: InputIdentity[] | null = null; constructor( private readonly cache: NativeCache, @@ -187,30 +211,46 @@ export class NativeExecutableCache { private readonly linker: string, private readonly debug: boolean, paths: string[], - /** Linux: trace with one verbose link, from a key's second link on. */ + /** Linux: trace with one verbose link, keep an unchanged output, and + * store the executable from a key's second link on. */ private readonly linux = false, ) { - this.firstLink = linux && !cache.has("executable", key) && !cache.has("linked", key); - // A first link neither restores nor traces, so it needs no identities. - this.inputs = this.firstLink ? [] : [...new Set(paths)].sort().map(identity); + this.inputs = [...new Set(paths)].sort().map(identity); + } + + /** Whether `recorded`, the inputs of an entry, still match and cover this + * build's inputs. */ + private covers(recorded: InputIdentity[]): boolean { + if (!stillMatches(recorded) || !stillMatches(this.inputs)) return false; + const saved = new Map(recorded.map((input) => [input.path, input])); + return this.inputs.every((input) => { + const found = saved.get(input.path); + return found !== undefined && sameIdentity(found, input); + }); + } + + /** Linux: true when `output` is the executable the last build of this key + * left there, unchanged, and its inputs are unchanged; the caller keeps it. */ + keep(output: string): boolean { + if (!this.linux) return false; + try { + const metadata = this.cache.read("kept", this.key); + if (metadata === null) return false; + const entry: unknown = JSON.parse(metadata.toString("utf8")); + if (!validKept(entry)) return false; + const current = identity(output); + return sameIdentity(current, entry.output) && this.covers(entry.inputs); + } catch { + return false; + } } restore(output: string): boolean { - if (this.firstLink) return false; try { const metadata = this.cache.read("executable", this.key); if (metadata === null) return false; const entry: unknown = JSON.parse(metadata.toString("utf8")); - if (!validEntry(entry) || !stillMatches(entry.inputs) || !stillMatches(this.inputs)) - return false; - const saved = new Map(entry.inputs.map((input) => [input.path, input])); - if ( - !this.inputs.every((input) => { - const recorded = saved.get(input.path); - return recorded !== undefined && sameIdentity(recorded, input); - }) - ) - return false; + if (!validEntry(entry) || !this.covers(entry.inputs)) return false; if (this.debug !== (entry.symbols !== null)) return false; const binary = this.cache.read("binary", entry.binary); const symbols = entry.symbols === null ? null : this.cache.read("dsym", entry.symbols); @@ -219,6 +259,7 @@ export class NativeExecutableCache { if (symbols !== null) stageSymbols(symbols, output); writeFileSync(output, binary); chmodSync(output, 0o755); + this.restoredInputs = entry.inputs; return true; } catch { return false; @@ -234,12 +275,6 @@ export class NativeExecutableCache { trace(args: string[], stage: string): boolean { this.traced = false; try { - // On Linux a key's first link only records that it was linked; a later - // link of the same key traces, so a build that runs once traces nothing. - if (this.firstLink) { - this.cache.mark("linked", this.key); - return false; - } const roots = [resolve(stage), realpathSync(stage)]; let driverInputs = this.linux ? [] @@ -250,7 +285,14 @@ export class NativeExecutableCache { const paths = tracePaths(linked, false, roots); if (paths.length === 0 || !stillMatches(this.inputs) || !stillMatches(driverInputs)) return false; - this.inputs.push(...driverInputs, ...paths.map(identity)); + // The driver's job line and the linker's trace name many of the same + // files; each is checked once. + const listed = new Set(this.inputs.map((input) => input.path)); + for (const input of [...driverInputs, ...paths.map(identity)]) + if (!listed.has(input.path)) { + listed.add(input.path); + this.inputs.push(input); + } this.traced = true; return true; } catch { @@ -260,7 +302,10 @@ export class NativeExecutableCache { publish(output: string): void { try { - if (!this.traced || !stillMatches(this.inputs)) return; + // On Linux the first build of a key only records its output (see + // recordOutput); the executable is stored once a later build relinks. + if (!this.traced || (this.linux && !this.cache.has("kept", this.key))) return; + if (!stillMatches(this.inputs)) return; const binary = readFileSync(output); const symbols = this.debug ? readSymbols(output) : null; const binaryKey = this.cache.writeContent("binary", binary); @@ -277,6 +322,24 @@ export class NativeExecutableCache { /* Build artifacts remain valid when caching is unavailable. */ } } + + /** Linux: records `output`, the executable this build left in place, for + * keep(). */ + recordOutput(output: string): void { + if (!this.linux) return; + try { + const inputs = this.traced ? this.inputs : this.restoredInputs; + if (inputs === null || !stillMatches(inputs)) return; + const entry: KeptEntry = { + schema: "scriptc.native-kept.v1", + output: identity(output), + inputs, + }; + this.cache.write("kept", this.key, JSON.stringify(entry)); + } catch { + /* Build artifacts remain valid when caching is unavailable. */ + } + } } export function openNativeExecutableCache( diff --git a/packages/compiler/src/native/prepare.test.ts b/packages/compiler/src/native/prepare.test.ts index 7ae2d1d5f..5d0f9faae 100644 --- a/packages/compiler/src/native/prepare.test.ts +++ b/packages/compiler/src/native/prepare.test.ts @@ -70,7 +70,7 @@ function fixture() { loads++; return runFrontend(path, loadProgram, npmStatic, externalTypes); }; - const prepare = (extra: { emitIr?: boolean } = {}) => + const prepare = (extra: { emitIr?: boolean; optimization?: "release" | "dev" } = {}) => prepareNativeExecutable(entry, { ...options, ...extra }, null, toolchain, frontend, cache); return { root, entry, dependency, cache, options, prepare, loads: () => loads }; } @@ -123,6 +123,38 @@ test("comment and whitespace edits reuse the frontend output a fresh build repro expect(edited.ok && edited.llvm.join("")).not.toBe(first.llvm.join("")); }); +test("a dev build reuses its frontend output after edits that keep every token's position", () => { + const f = fixture(); + const dev = { optimization: "dev" as const }; + const first = f.prepare(dev); + expect(first.ok, JSON.stringify(first)).toBe(true); + if (!first.ok) return; + // Debug information records lines and columns: a comment after the last + // token and trailing whitespace keep them. + const appended = 'export const value = "first"; \n// The exported value.\n'; + writeFileSync(f.dependency, appended); + const reused = f.prepare(dev); + expect(f.loads()).toBe(1); + expect(reused.ok).toBe(true); + if (!reused.ok) return; + expect(reused.llvm.join("")).toBe(first.llvm.join("")); + expect(new Map(reused.sources).get(f.dependency)).toBe(appended); + const fresh = prepareNativeExecutable( + f.entry, + { ...f.options, ...dev }, + null, + fixtureToolchain(), + (path, npmStatic, externalTypes) => runFrontend(path, loadProgram, npmStatic, externalTypes), + new NativeCache(join(f.root, "fresh-cache")), + ); + expect(fresh.ok && fresh.llvm.join("")).toBe(first.llvm.join("")); + // A comment that moves the token after it is rebuilt. + writeFileSync(f.dependency, '// The exported value.\nexport const value = "first";\n'); + const moved = f.prepare(dev); + expect(f.loads()).toBe(2); + expect(moved.ok && moved.llvm.join("")).not.toBe(first.llvm.join("")); +}); + test("a comment edit inside source text the output carries is rebuilt", () => { const f = fixture(); // A failing assert(value) reports its call's source text. diff --git a/packages/compiler/src/native/prepare.ts b/packages/compiler/src/native/prepare.ts index 67a4b5a9e..7118e47b9 100644 --- a/packages/compiler/src/native/prepare.ts +++ b/packages/compiler/src/native/prepare.ts @@ -2,6 +2,7 @@ import { targetCpuField, targetCpuTargetRefusal } from "../backend/target-cpu.js import { nativeCodegenDiag, type ScrDiagnostic } from "../diagnostics/diagnostic.js"; import type { CompilationTiming } from "../timing.js"; import { readFileSync } from "node:fs"; +import { availableParallelism } from "node:os"; import { basename, dirname, join, resolve } from "node:path"; import type { CompileFailure, @@ -31,7 +32,10 @@ import { import { compilerOwnedFileRoots } from "../frontend/dts-paths.js"; import { provenanceSources } from "../frontend/provenance-registry.js"; import { processFork } from "../frontend/process-fork.js"; -import { tokenEquivalentSource } from "../frontend/token-equivalence.js"; +import { + tokenEquivalentSource, + tokenPositionEquivalentSource, +} from "../frontend/token-equivalence.js"; import { llvmBytes, wtf8Bytes } from "../backend/literals.js"; import { prepareLibrary } from "../library/prepare.js"; import type { LibraryProfile } from "../library/library-profile.js"; @@ -142,9 +146,11 @@ function readFrontendCache( /** The stale entry's input when every changed source is token-equivalent to * the text the entry was built from (edits to whitespace and to comments * that cannot affect compilation; see token-equivalence.ts). Equivalent - * sources lower to the same IR apart from source locations, which only dev - * builds' debug information and the serialized IR carry. Diagnostics carry - * locations too, so an entry that reported warnings is rebuilt. A changed + * sources lower to the same IR apart from source locations, which only the + * debug information of dev builds and the serialized IR carry. A build + * with debug information accepts only edits that keep every token's line + * and column, the positions the debug information records. Diagnostics + * carry locations too, so an entry that reported warnings is rebuilt. A changed * source whose raw text or line positions the frontend read (call-site text * in messages, comptime callbacks, test locations) is rebuilt as well: * frontendInputsSemanticallyMatch refuses its `text` probe. */ @@ -154,17 +160,17 @@ function equivalentFrontendInput( exclusions: FrontendInputExclusions, ): { input: NativeExecutableInput; probes: FrontendInputSnapshot } | null { if ( - options.optimization === "dev" || options.sanitize === true || options.emitIr === true || stale.ir !== null || (stale.input.warnings?.length ?? 0) > 0 ) return null; + const debugInfo = options.optimization === "dev" && !options.strip; const match = frontendInputsSemanticallyMatch( stale.probes, new Map(stale.input.sources), - tokenEquivalentSource, + debugInfo ? tokenPositionEquivalentSource : tokenEquivalentSource, exclusions, ); if (match === null || match.changed.length === 0) return null; @@ -546,10 +552,12 @@ export function prepareNativeLibrary( return input; } -/** Worker processes that emit function bodies at a time - * (SCRIPTC_EMIT_WORKERS; 0 or 1 emits in this process). */ +/** Worker processes that emit function bodies at a time: up to eight, one + * per available core (SCRIPTC_EMIT_WORKERS sets the count; 0 or 1 emits in + * this process). */ function emitWorkers(): number { - const value = Number(process.env["SCRIPTC_EMIT_WORKERS"] ?? "4"); + const configured = process.env["SCRIPTC_EMIT_WORKERS"]; + const value = configured === undefined ? Math.min(8, availableParallelism()) : Number(configured); return Number.isFinite(value) && value > 1 ? Math.floor(value) : 1; } diff --git a/packages/compiler/test/chunked-emission.test.ts b/packages/compiler/test/chunked-emission.test.ts index b4912c9f6..2aec1e49b 100644 --- a/packages/compiler/test/chunked-emission.test.ts +++ b/packages/compiler/test/chunked-emission.test.ts @@ -9,9 +9,10 @@ import { inlineRcOption } from "../src/backend/llvm/shapes.js"; // Programs whose function bodies request module-level helpers and tables // that the chunk journal (emit-journal.ts) carries: checked-dynamic -// conversions, inspection, publication walkers, live record views, stream -// adapters, regex instances, colliding literal digests, interned C strings, -// speculative loop sites, typed message listeners and readers. +// conversions (with their discriminant tables), inspection, publication +// walkers, live record views, stream adapters, regex instances, colliding +// literal digests, interned C strings, speculative loop sites, typed message +// listeners and readers. const PROGRAMS = [ "1003-json-parse-unions.ts", "114-live-record-delete.ts", @@ -21,6 +22,7 @@ const PROGRAMS = [ "4900-console-instances.ts", "5612-readable-from-iterables.cjs", "5702-publish-candidate-buffers.ts", + "6305-json-discriminant-dispatch.ts", "7321-speculative-loop-bail-sites.ts", "7360-worker-typed-records.ts", "node24-text-utilities-coercion.js", diff --git a/packages/compiler/test/js-ic-emission.test.ts b/packages/compiler/test/js-ic-emission.test.ts index 47d641bff..54ea13d26 100644 --- a/packages/compiler/test/js-ic-emission.test.ts +++ b/packages/compiler/test/js-ic-emission.test.ts @@ -64,6 +64,50 @@ console.log(read(o), at(dyn([5, 6]), 1), call(o), o.n); } }); +test("property sites keep their caches in programs that declare classes and subclasses", async () => { + const dir = await mkdtemp(join(tmpdir(), "scriptc-js-ic-class-")); + try { + const entry = join(dir, "main.js"); + const outPath = join(dir, "main.ir.json"); + // Declared classes give every property helper class branches and + // property-bag lookups, and a subclass adds a normalization step; plain + // objects still read through the caches. + await writeFile( + entry, + ` +class Point { + constructor(x) { this.x = x; } +} +class Tagged extends Point { + constructor(x, tag) { super(x); this.tag = tag; } +} +function dyn(v) { return v; } +function read(o) { return o.x; } +function write(o, v) { o.x = v; } +function at(a, i) { return a[i]; } +var o = dyn({ x: 1 }); +var p = dyn(new Point(4)); +var t = dyn(new Tagged(6, "t")); +write(o, 3); +write(p, 5); +write(t, 7); +console.log(read(o), read(p), read(t), at(dyn([5, 6]), 1), at(p, "x"), at(t, "tag")); +`, + ); + const result = await compile(entry, { outDir: dir, outPath, outputKind: "ir" }); + if (!result.ok) + throw new Error(result.diagnostics.map((d) => `${d.code}: ${d.message}`).join("\n")); + const mod = deserializeModule(await readFile(outPath, "utf8")); + expect(validateModule(mod)).toEqual([]); + const llvm = emitLlvmModule(mod); + expect(body(llvm, "read")).toContain("@scr_js_load_probe("); + expect(body(llvm, "write")).toContain("@scr_js_store_probe("); + expect(body(llvm, "at")).toContain("@scr_js_keyed_probe("); + } finally { + await rm(dir, { recursive: true, force: true }); + } +}); + test("typed programs have no cache sites and keep the ordinary runtime", async () => { const dir = await mkdtemp(join(tmpdir(), "scriptc-js-ic-typed-")); try { diff --git a/packages/compiler/test/keyed-method-writes.test.ts b/packages/compiler/test/keyed-method-writes.test.ts new file mode 100644 index 000000000..407530945 --- /dev/null +++ b/packages/compiler/test/keyed-method-writes.test.ts @@ -0,0 +1,79 @@ +import { mkdtemp, readFile, rm, writeFile } from "node:fs/promises"; +import { tmpdir } from "node:os"; +import { join } from "node:path"; +import { expect, test } from "vitest"; +import { compile, deserializeModule } from "../src/index.js"; +import type { IrModule } from "../src/ir/ir.js"; +import { everyStmtList } from "../src/ir/traverse.js"; + +async function lowerSource(source: string): Promise { + const dir = await mkdtemp(join(tmpdir(), "scriptc-keyed-writes-")); + try { + const entry = join(dir, "main.ts"); + await writeFile(entry, source); + const outPath = join(dir, "main.ir.json"); + const result = await compile(entry, { outPath, outDir: dir, outputKind: "ir" }); + if (!result.ok) throw new Error(result.diagnostics.map((d) => d.message).join("\n")); + return deserializeModule(await readFile(outPath, "utf8")); + } finally { + await rm(dir, { recursive: true, force: true }); + } +} + +/** Calls in `run` that go through a dynamic value. */ +function dynamicCalls(mod: IrModule): number { + const fn = mod.functions.find((f) => f.name === "run"); + if (!fn) throw new Error("no run function"); + let n = 0; + everyStmtList(fn.body, { + expr: (e) => { + if (e.kind === "dynCall" || e.kind === "dynInvoke") n++; + return true; + }, + stmt: () => true, + }); + return n; +} + +const shape = ` +type Step = "scale" | "clamp"; +class Transform { + scale(x: number): number { + return x * 2; + } + clamp(x: number): number { + return x < 0 ? 0 : x; + } +} +`; + +test("element writes of data under method names keep the methods native", async () => { + const mod = await lowerSource(`${shape} +const enabled: Record = { scale: false, clamp: false }; +function enable(name: Step): void { + enabled[name] = true; +} +const totals: Record = {}; +function run(t: Transform): number { + enable("scale"); + totals["clamp"] = 1; + return t.clamp(t.scale(3)); +} +console.log(run(new Transform())); +`); + expect(dynamicCalls(mod)).toBe(0); +}); + +test("element writes through untyped receivers can still replace a method", async () => { + const mod = await lowerSource(`${shape} +function replace(target: object, name: Step): void { + (target as any)[name] = (x: number) => x + 100; +} +function run(t: Transform): number { + replace(t, "scale"); + return t.clamp(t.scale(3)); +} +console.log(run(new Transform())); +`); + expect(dynamicCalls(mod)).toBeGreaterThan(0); +}); diff --git a/packages/compiler/test/native-partitions.test.ts b/packages/compiler/test/native-partitions.test.ts index ce1f14335..fcd0132e8 100644 --- a/packages/compiler/test/native-partitions.test.ts +++ b/packages/compiler/test/native-partitions.test.ts @@ -29,8 +29,7 @@ afterEach(async () => { /** A program whose internal functions, constants and state must stay * reachable across partitions. `edit` changes one function's addend. */ -function program(edit = 0, extra = ""): string { - const count = 24; +function program(edit = 0, extra = "", count = 24): string { const lines = [ "; partitioned program", "%Pair = type { i64, i64 }", @@ -74,11 +73,12 @@ function program(edit = 0, extra = ""): string { return lines.join("\n") + "\n"; } -function expectedExit(edit = 0): number { - const addends = Array.from({ length: 24 }, (_, index) => (index === 5 ? index + edit : index)); +function expectedExit(edit = 0, count = 24): number { + const addends = Array.from({ length: count }, (_, index) => (index === 5 ? index + edit : index)); let total = 5; // strlen("total") - for (let start = 0; start < 24; start++) - for (let index = start; index < 24; index++) total += (start === 0 ? 4 : 1) * addends[index]!; + for (let start = 0; start < count; start++) + for (let index = start; index < count; index++) + total += (start === 0 ? 4 : 1) * addends[index]!; return total % 251; } @@ -162,6 +162,56 @@ describe.runIf(helper !== null)("LLVM program partitions", () => { }); } + test("functions under one group line share a partition", async () => { + const dir = await mkdtemp(join(tmpdir(), "scriptc-partitions-groups-")); + dirs.push(dir); + // Groups of four steps, each smaller than the mean partition; main ends + // the last group. Placement by group applies from 32 partitions. + const steps = 128; + const grouped = (edit: number, count = steps) => + program(edit, "", count) + .split("\n") + .flatMap((line) => { + const step = /^define internal void @step(\d+)\(/.exec(line); + if (step !== null && Number(step[1]) % 4 === 0) + return [`; scriptc-group ${Number(step[1]) / 4}`, line]; + return line.startsWith("define i32 @main(") ? ["; scriptc-group", line] : [line]; + }) + .join("\n"); + const cache = join(dir, "cache"); + const options = { outputs: 32, optLevel: "0", cache } as const; + const objects = await emit(dir, grouped(0), { ...options, name: "grouped" }); + expect(await exitCode(dir, objects, "grouped")).toBe(expectedExit(0, steps)); + const owner = new Map(); + for (const [index, object] of objects.entries()) { + const { stdout } = await execFileAsync("nm", [object]); + for (const line of stdout.split("\n")) { + const defined = /^[0-9a-fA-F]+ [A-Za-z] _?step(\d+)$/.exec(line.trim()); + if (defined !== null) owner.set(Number(defined[1]), index); + } + } + expect(owner.size).toBe(steps); + // step0's only caller is main, so it goes with main. + for (let group = 0; group < steps / 4; group++) { + const members = [0, 1, 2, 3].map((offset) => group * 4 + offset).filter((step) => step > 0); + expect(new Set(members.map((step) => owner.get(step))).size).toBe(1); + } + + // An edit to one function still changes one partition's source. + const sources = await cacheEntries(cache, ".o"); + const edited = await emit(dir, grouped(7), { ...options, name: "edited" }); + const after = await cacheEntries(cache, ".o"); + expect([...after].filter((entry) => !sources.has(entry))).toHaveLength(1); + expect(await exitCode(dir, edited, "edited")).toBe(expectedExit(7, steps)); + + // With fewer partitions, functions keep their placement by name. + const contents = (paths: string[]) => Promise.all(paths.map((path) => readFile(path))); + const few = { outputs: 4, optLevel: "0" } as const; + expect( + await contents(await emit(dir, grouped(0, 24), { ...few, name: "few-grouped" })), + ).toEqual(await contents(await emit(dir, program(), { ...few, name: "few-plain" }))); + }); + test("a table written one element per line keeps the text partitioner and its reuse", async () => { const dir = await mkdtemp(join(tmpdir(), "scriptc-partitions-table-")); dirs.push(dir); diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/6305-json-discriminant-dispatch.ts.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/6305-json-discriminant-dispatch.ts.json new file mode 100644 index 000000000..b83ada276 --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/6305-json-discriminant-dispatch.ts.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/6305-json-discriminant-dispatch.ts" + ], + "diags": [] +} diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/6306-discriminant-getter-dispatch.ts.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/6306-discriminant-getter-dispatch.ts.json new file mode 100644 index 000000000..ba7cba509 --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/6306-discriminant-getter-dispatch.ts.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/6306-discriminant-getter-dispatch.ts" + ], + "diags": [] +} diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/7984-untyped-array-methods.mjs.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/7984-untyped-array-methods.mjs.json new file mode 100644 index 000000000..2c823a24f --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/7984-untyped-array-methods.mjs.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/7984-untyped-array-methods.mjs" + ], + "diags": [] +} diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8002-while-byte-length-companions.ts.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8002-while-byte-length-companions.ts.json new file mode 100644 index 000000000..e2a36c22a --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8002-while-byte-length-companions.ts.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/8002-while-byte-length-companions.ts" + ], + "diags": [] +} diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8105-url-idna-hosts.js.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8105-url-idna-hosts.js.json new file mode 100644 index 000000000..50ed1b401 --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8105-url-idna-hosts.js.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/8105-url-idna-hosts.js" + ], + "diags": [] +} diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8106-shared-typed-array-set.ts.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8106-shared-typed-array-set.ts.json new file mode 100644 index 000000000..8b7a07fac --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8106-shared-typed-array-set.ts.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/8106-shared-typed-array-set.ts" + ], + "diags": [] +} diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8107-shared-typed-array-set-workers.ts.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8107-shared-typed-array-set-workers.ts.json new file mode 100644 index 000000000..338b4bc03 --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8107-shared-typed-array-set-workers.ts.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/8107-shared-typed-array-set-workers.ts" + ], + "diags": [] +} diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8215-js-any-logical-module-bindings/main.ts.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8215-js-any-logical-module-bindings/main.ts.json new file mode 100644 index 000000000..0ae9f6f2b --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/8215-js-any-logical-module-bindings/main.ts.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/8215-js-any-logical-module-bindings/main.ts" + ], + "diags": [] +} diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/instanceof-primitive-wrappers-typed.ts.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/instanceof-primitive-wrappers-typed.ts.json new file mode 100644 index 000000000..fba491f13 --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/instanceof-primitive-wrappers-typed.ts.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/instanceof-primitive-wrappers-typed.ts" + ], + "diags": [] +} diff --git a/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/instanceof-primitive-wrappers.mjs.json b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/instanceof-primitive-wrappers.mjs.json new file mode 100644 index 000000000..e2dddae48 --- /dev/null +++ b/packages/compiler/test/ts7/baselines/order-parity/tests/corpus/instanceof-primitive-wrappers.mjs.json @@ -0,0 +1,6 @@ +{ + "order": [ + "/tests/corpus/instanceof-primitive-wrappers.mjs" + ], + "diags": [] +} diff --git a/packages/compiler/test/untyped-push-emission.test.ts b/packages/compiler/test/untyped-push-emission.test.ts new file mode 100644 index 000000000..2abdf7f41 --- /dev/null +++ b/packages/compiler/test/untyped-push-emission.test.ts @@ -0,0 +1,42 @@ +import { mkdtempSync, readFileSync, rmSync, writeFileSync } from "node:fs"; +import { tmpdir } from "node:os"; +import { join } from "node:path"; +import { expect, test } from "vitest"; +import { compile, deserializeModule, validateModule } from "../src/index.js"; +import { emitLlvmModule } from "../src/backend/llvm/emitter.js"; + +// An untyped push in statement position drops the new length, so it calls +// the entry that does not box it; a push whose result is read keeps the +// ordinary entry, as do the other methods. +test("a statement-position untyped push skips boxing its result", async () => { + const dir = mkdtempSync(join(tmpdir(), "scriptc-untyped-push-")); + try { + const entry = join(dir, "main.mjs"); + const output = join(dir, "main.ir.json"); + writeFileSync( + entry, + `function fill(a, n) { + for (let i = 0; i < n; i++) a.push(i); + const length = a.push(n); + a.pop(); + return length; +} +console.log(fill(JSON.parse("[]"), 3)); +`, + ); + const lowered = await compile(entry, { outputKind: "ir", outDir: dir, outPath: output }); + if (!lowered.ok) throw new Error(JSON.stringify(lowered.diagnostics)); + const module = deserializeModule(readFileSync(output, "utf8")); + expect(validateModule(module)).toEqual([]); + const llvm = emitLlvmModule(module, {}); + const body = llvm.split(/\n(?=define )/).find((fn) => /@sc_[a-z]*f_fill\b/.test(fn))!; + const discarded = [...body.matchAll(/call void @scr_dyn_invoke_discard_id\(/g)]; + const kept = [...body.matchAll(/= call ptr @scr_dyn_invoke_prepared_id\(/g)]; + expect(discarded).toHaveLength(1); + // `a.push(n)` for its length, and `a.pop()`, whose result is dropped + // but which has no unboxed entry. + expect(kept).toHaveLength(2); + } finally { + rmSync(dir, { recursive: true, force: true }); + } +}); diff --git a/packages/compiler/test/while-induction-emission.test.ts b/packages/compiler/test/while-induction-emission.test.ts new file mode 100644 index 000000000..08fbab849 --- /dev/null +++ b/packages/compiler/test/while-induction-emission.test.ts @@ -0,0 +1,94 @@ +import { mkdtemp, readFile, rm, writeFile } from "node:fs/promises"; +import { tmpdir } from "node:os"; +import { join } from "node:path"; +import { expect, test } from "vitest"; +import { compile } from "../src/index.js"; + +async function emitLlvm(source: string): Promise { + const dir = await mkdtemp(join(tmpdir(), "scriptc-while-induction-")); + try { + const entry = join(dir, "main.ts"); + const outPath = join(dir, "main.ll"); + await writeFile(entry, source); + const result = await compile(entry, { + outDir: dir, + outPath, + outputKind: "llvm", + }); + expect(result.ok, JSON.stringify(result.diagnostics)).toBe(true); + return await readFile(outPath, "utf8"); + } finally { + await rm(dir, { recursive: true, force: true }); + } +} + +/** The definition holding the function's statements: the borrowed-argument + * body when there is one, otherwise the function itself. */ +function body(ll: string, name: string): string { + const definition = (prefix: string): number => + ll.search(new RegExp(`define [^\\n]*@${prefix}_${name}\\(`)); + const start = definition("sc_bf") >= 0 ? definition("sc_bf") : definition("sc_f"); + expect(start).toBeGreaterThanOrEqual(0); + return ll.slice(start, ll.indexOf("\n}\n", start)); +} + +test("a while loop bounded by a byte length keeps its counter and companions in integers", async () => { + const ll = await emitLlvm(` +function reversed(line: Uint8Array, target: Uint8Array): number { + const len = line.length; + let read = 0; + let written = 0; + let lines = 0; + while (read < len) { + const c = line[len - 1 - read]!; + read++; + if (c === 10) { + if (read === len) break; + read++; + } + target[written] = c; + written++; + if (written % 60 === 0) { + target[written] = 10; + written++; + lines++; + } + } + return written + lines; +} +console.log(reversed(new Uint8Array([65, 10, 66]), new Uint8Array(8))); +`); + const fn = body(ll, "reversed"); + expect(fn).toContain("alloca i64 ; integer read"); + expect(fn).toContain("alloca i64 ; integer written"); + expect(fn).toContain("alloca i64 ; integer lines"); +}); + +test("a companion a nested loop advances, or a counter a continue can skip, stays a double", async () => { + const ll = await emitLlvm(` +function nested(line: Uint8Array): number { + let read = 0; + let total = 0; + while (read < line.length) { + read++; + for (let j = 0; j < 2; j++) total++; + } + return total; +} +function skipped(line: Uint8Array): number { + let read = 0; + let total = 0; + while (read < line.length) { + if (line[read] === 0) continue; + read++; + total++; + } + return total; +} +console.log(nested(new Uint8Array(3)), skipped(new Uint8Array([1, 2]))); +`); + expect(body(ll, "nested")).not.toContain("; integer total"); + // The counter compared with the length directly is an integer. + expect(body(ll, "skipped")).toContain("; integer read"); + expect(body(ll, "skipped")).not.toContain("; integer total"); +}); diff --git a/packages/runtime-pack-common/scripts/build.mjs b/packages/runtime-pack-common/scripts/build.mjs index d4794b7d1..479d8aad4 100644 --- a/packages/runtime-pack-common/scripts/build.mjs +++ b/packages/runtime-pack-common/scripts/build.mjs @@ -386,6 +386,10 @@ async function build() { ? [ await archive("libregexp", lreSources, quickjs, [ ...vendorTarget, + // Section elimination keeps only the libunicode tables a + // program reaches: the URL host parser's NFC links the + // normalization tables without the regex property ones. + ...matrix.executable_section_elimination.compile_flags, "-std=c11", "-Os", "-I", diff --git a/packages/runtime/src/scr_array.c b/packages/runtime/src/scr_array.c index 4c7b84d98..39998b8ff 100644 --- a/packages/runtime/src/scr_array.c +++ b/packages/runtime/src/scr_array.c @@ -1724,6 +1724,14 @@ static void scr_frozen_undeletable(double index) { scr_frozen_throw("Cannot delete property '", index, "' of [object Array]"); } +void scr_array_not_extensible_error(double index) { + scr_frozen_not_extensible(index); +} + +void scr_array_undeletable_error(double index) { + scr_frozen_undeletable(index); +} + /* Each case finds the operation's first write in the order the spec makes * its writes. */ bool scr_array_frozen_error(int op, size_t len, double x, double y, double z) { diff --git a/packages/runtime/src/scr_array.h b/packages/runtime/src/scr_array.h index 23390a183..48fa4f26f 100644 --- a/packages/runtime/src/scr_array.h +++ b/packages/runtime/src/scr_array.h @@ -231,6 +231,10 @@ bool scr_arr_is_frozen(const ScrArr *a); * end), 9 copyWithin (x = target, y = start, z = end), 10 splice (x = * start, y = delete count, z = item count). */ bool scr_array_frozen_error(int op, size_t len, double x, double y, double z); +/* Node's TypeErrors for a write that a non-extensible or sealed array + * refuses: adding the element at `index`, or deleting it. */ +void scr_array_not_extensible_error(double index); +void scr_array_undeletable_error(double index); /* The start of the dense slots' allocation: `data` less any head room. */ static inline uint64_t *scr_arr_block(const ScrArr *a) { diff --git a/packages/runtime/src/scr_bigint.c b/packages/runtime/src/scr_bigint.c index dd0a70ff7..5af2eafaa 100644 --- a/packages/runtime/src/scr_bigint.c +++ b/packages/runtime/src/scr_bigint.c @@ -215,30 +215,81 @@ static int bi_cmp_limbs(const bi_limb *a, size_t an, const bi_limb *b, size_t bn return 0; } -/* r = a + b over n limbs; returns the carry. r may alias a or b. */ +/* x + y + carry_in (0 or 1), with the carry out in *carry: the add-with- + * carry builtins compile to one adc chain. */ +#if defined(__has_builtin) +#if __has_builtin(__builtin_addcll) && __has_builtin(__builtin_subcll) && \ + __has_builtin(__builtin_addc) && __has_builtin(__builtin_subc) +#define BI_HAVE_ADDC 1 +#endif +#endif +static inline bi_limb bi_addc(bi_limb x, bi_limb y, bi_limb *carry) { +#if defined(BI_HAVE_ADDC) && BI_LIMB_BITS == 64 + unsigned long long c; + bi_limb s = __builtin_addcll(x, y, *carry, &c); + *carry = c; + return s; +#elif defined(BI_HAVE_ADDC) + unsigned c; + bi_limb s = __builtin_addc(x, y, *carry, &c); + *carry = c; + return s; +#else + bi_limb s = x + y; + bi_limb c1 = s < x; + bi_limb t = s + *carry; + *carry = c1 | (t < s); + return t; +#endif +} + +/* x - y - borrow_in (0 or 1), with the borrow out in *borrow. */ +static inline bi_limb bi_subc(bi_limb x, bi_limb y, bi_limb *borrow) { +#if defined(BI_HAVE_ADDC) && BI_LIMB_BITS == 64 + unsigned long long b; + bi_limb d = __builtin_subcll(x, y, *borrow, &b); + *borrow = b; + return d; +#elif defined(BI_HAVE_ADDC) + unsigned b; + bi_limb d = __builtin_subc(x, y, *borrow, &b); + *borrow = b; + return d; +#else + bi_limb d = x - y; + bi_limb b1 = x < y; + bi_limb t = d - *borrow; + *borrow = b1 | (d < *borrow); + return t; +#endif +} + +/* r = a + b over n limbs; returns the carry. r may alias a or b. Four + * limbs a pass keep the carry in one chain. */ static bi_limb bi_add_n(bi_limb *r, const bi_limb *a, const bi_limb *b, size_t n) { bi_limb carry = 0; - for (size_t i = 0; i < n; i++) { - bi_limb x = a[i], s = x + b[i]; - bi_limb c1 = s < x; - bi_limb t = s + carry; - carry = c1 | (t < s); - r[i] = t; + size_t i = 0; + for (; i + 4 <= n; i += 4) { + r[i] = bi_addc(a[i], b[i], &carry); + r[i + 1] = bi_addc(a[i + 1], b[i + 1], &carry); + r[i + 2] = bi_addc(a[i + 2], b[i + 2], &carry); + r[i + 3] = bi_addc(a[i + 3], b[i + 3], &carry); } + for (; i < n; i++) r[i] = bi_addc(a[i], b[i], &carry); return carry; } /* r = a - b over n limbs; returns the borrow. r may alias a or b. */ static bi_limb bi_sub_n(bi_limb *r, const bi_limb *a, const bi_limb *b, size_t n) { bi_limb borrow = 0; - for (size_t i = 0; i < n; i++) { - bi_limb x = a[i], y = b[i]; - bi_limb d = x - y; - bi_limb b1 = x < y; - bi_limb t = d - borrow; - borrow = b1 | (d < borrow); - r[i] = t; + size_t i = 0; + for (; i + 4 <= n; i += 4) { + r[i] = bi_subc(a[i], b[i], &borrow); + r[i + 1] = bi_subc(a[i + 1], b[i + 1], &borrow); + r[i + 2] = bi_subc(a[i + 2], b[i + 2], &borrow); + r[i + 3] = bi_subc(a[i + 3], b[i + 3], &borrow); } + for (; i < n; i++) r[i] = bi_subc(a[i], b[i], &borrow); return borrow; } @@ -277,9 +328,22 @@ static bi_limb bi_sub_into(bi_limb *r, size_t rn, const bi_limb *a, size_t an) { return borrow; } -/* r = a * m + carry over n limbs; returns the high limb. r may alias a. */ +/* r = a * m + carry over n limbs; returns the high limb. r may alias a. + * Four limbs a pass: the four products, then one carry chain adding each + * product's high limb into the next low limb. */ static bi_limb bi_mul1(bi_limb *r, const bi_limb *a, size_t n, bi_limb m, bi_limb carry) { - for (size_t i = 0; i < n; i++) { + size_t i = 0; + for (; i + 4 <= n; i += 4) { + bi_wide p0 = (bi_wide)a[i] * m, p1 = (bi_wide)a[i + 1] * m; + bi_wide p2 = (bi_wide)a[i + 2] * m, p3 = (bi_wide)a[i + 3] * m; + bi_limb c = 0; + r[i] = bi_addc((bi_limb)p0, carry, &c); + r[i + 1] = bi_addc((bi_limb)p1, (bi_limb)(p0 >> BI_LIMB_BITS), &c); + r[i + 2] = bi_addc((bi_limb)p2, (bi_limb)(p1 >> BI_LIMB_BITS), &c); + r[i + 3] = bi_addc((bi_limb)p3, (bi_limb)(p2 >> BI_LIMB_BITS), &c); + carry = (bi_limb)(p3 >> BI_LIMB_BITS) + c; + } + for (; i < n; i++) { bi_wide p = (bi_wide)a[i] * m + carry; r[i] = (bi_limb)p; carry = (bi_limb)(p >> BI_LIMB_BITS); @@ -287,10 +351,27 @@ static bi_limb bi_mul1(bi_limb *r, const bi_limb *a, size_t n, bi_limb m, bi_lim return carry; } -/* r += a * m over n limbs; returns the high limb. */ +/* r += a * m over n limbs; returns the high limb. Four limbs a pass: one + * carry chain forms the four product limbs, a second adds them to r. */ static bi_limb bi_addmul1(bi_limb *r, const bi_limb *a, size_t n, bi_limb m) { bi_limb carry = 0; - for (size_t i = 0; i < n; i++) { + size_t i = 0; + for (; i + 4 <= n; i += 4) { + bi_wide p0 = (bi_wide)a[i] * m, p1 = (bi_wide)a[i + 1] * m; + bi_wide p2 = (bi_wide)a[i + 2] * m, p3 = (bi_wide)a[i + 3] * m; + bi_limb c = 0, k = 0; + bi_limb t0 = bi_addc((bi_limb)p0, carry, &c); + bi_limb t1 = bi_addc((bi_limb)p1, (bi_limb)(p0 >> BI_LIMB_BITS), &c); + bi_limb t2 = bi_addc((bi_limb)p2, (bi_limb)(p1 >> BI_LIMB_BITS), &c); + bi_limb t3 = bi_addc((bi_limb)p3, (bi_limb)(p2 >> BI_LIMB_BITS), &c); + r[i] = bi_addc(r[i], t0, &k); + r[i + 1] = bi_addc(r[i + 1], t1, &k); + r[i + 2] = bi_addc(r[i + 2], t2, &k); + r[i + 3] = bi_addc(r[i + 3], t3, &k); + /* r + a * m over these limbs stays below B^5, so this cannot wrap */ + carry = (bi_limb)(p3 >> BI_LIMB_BITS) + c + k; + } + for (; i < n; i++) { bi_wide p = (bi_wide)a[i] * m + r[i] + carry; r[i] = (bi_limb)p; carry = (bi_limb)(p >> BI_LIMB_BITS); @@ -301,7 +382,23 @@ static bi_limb bi_addmul1(bi_limb *r, const bi_limb *a, size_t n, bi_limb m) { /* r -= a * m over n limbs; returns the borrow limb. */ static bi_limb bi_submul1(bi_limb *r, const bi_limb *a, size_t n, bi_limb m) { bi_limb carry = 0; - for (size_t i = 0; i < n; i++) { + size_t i = 0; + for (; i + 4 <= n; i += 4) { + bi_wide p0 = (bi_wide)a[i] * m, p1 = (bi_wide)a[i + 1] * m; + bi_wide p2 = (bi_wide)a[i + 2] * m, p3 = (bi_wide)a[i + 3] * m; + bi_limb c = 0, k = 0; + bi_limb t0 = bi_addc((bi_limb)p0, carry, &c); + bi_limb t1 = bi_addc((bi_limb)p1, (bi_limb)(p0 >> BI_LIMB_BITS), &c); + bi_limb t2 = bi_addc((bi_limb)p2, (bi_limb)(p1 >> BI_LIMB_BITS), &c); + bi_limb t3 = bi_addc((bi_limb)p3, (bi_limb)(p2 >> BI_LIMB_BITS), &c); + r[i] = bi_subc(r[i], t0, &k); + r[i + 1] = bi_subc(r[i + 1], t1, &k); + r[i + 2] = bi_subc(r[i + 2], t2, &k); + r[i + 3] = bi_subc(r[i + 3], t3, &k); + /* a * m + carry over these limbs stays below B^5: the borrow fits */ + carry = (bi_limb)(p3 >> BI_LIMB_BITS) + c + k; + } + for (; i < n; i++) { bi_wide p = (bi_wide)a[i] * m + carry; bi_limb lo = (bi_limb)p; carry = (bi_limb)(p >> BI_LIMB_BITS); @@ -543,6 +640,56 @@ static void bi_divmod_knuth(bi_limb *q, bi_limb *r, const bi_limb *a, size_t an, free(un); } +/* a / b for a and b of n >= 2 limbs each, b's top limb nonzero: the + * quotient is one limb. Knuth's estimate comes from the top limbs as + * they read normalized (three of a, two of b) and is the quotient or one + * more, which one pass of a - qhat * b over the operands as they are + * settles, with no normalized copies. r, unless NULL, receives the n + * limbs of a % b. */ +static bi_limb bi_div_qlimb(bi_limb *r, const bi_limb *a, const bi_limb *b, size_t n) { + unsigned s = bi_clz(b[n - 1]); + bi_limb d1 = b[n - 1], d0 = b[n - 2], u2 = 0, u1 = a[n - 1], u0 = a[n - 2]; + if (s) { + unsigned t = BI_LIMB_BITS - s; + bi_limb b3 = n >= 3 ? b[n - 3] : 0, a3 = n >= 3 ? a[n - 3] : 0; + d1 = (d1 << s) | (d0 >> t); + d0 = (d0 << s) | (b3 >> t); + u2 = u1 >> t; + u1 = (u1 << s) | (u0 >> t); + u0 = (u0 << s) | (a3 >> t); + } + bi_limb qhat, rhat; + bool rhat_big; /* rhat >= B: the estimate needs no further test */ + if (u2 >= d1) { + qhat = BI_LIMB_MAX; + rhat = u1 + d1; + rhat_big = rhat < d1; + } else { + bi_divisor top = bi_divisor_of(d1); + qhat = bi_div21(u2, u1, top.d, top.v, &rhat); + rhat_big = false; + } + while (!rhat_big && (bi_wide)qhat * d0 > (((bi_wide)rhat << BI_LIMB_BITS) | u0)) { + qhat--; + rhat += d1; + rhat_big = rhat < d1; + } + /* a - qhat * b: negative when the product's high limb or the borrow + * reaches past a's n limbs */ + bi_limb carry = 0, borrow = 0; + for (size_t i = 0; i < n; i++) { + bi_wide p = (bi_wide)b[i] * qhat + carry; + carry = (bi_limb)(p >> BI_LIMB_BITS); + bi_limb d = bi_subc(a[i], (bi_limb)p, &borrow); + if (r) r[i] = d; + } + if (carry || borrow) { + qhat--; + if (r) bi_add_n(r, r, b, n); + } + return qhat; +} + /* ---- comparisons and hashing ----------------------------------------- */ static int bi_cmp_abs(const ScrBigInt *a, const ScrBigInt *b) { @@ -709,6 +856,13 @@ static void bi_divmod_abs(const ScrBigInt *a, const ScrBigInt *b, ScrBigInt **qo if (rout) *rout = bi_u64_signed(rem, 1); return; } + if (an == bn) { + ScrBigInt *r = rout ? bi_alloc(bn) : NULL; + bi_limb q = bi_div_qlimb(r ? r->limb : NULL, a->limb, b->limb, bn); + if (qout) *qout = bi_u64_signed(q, 1); + if (rout) *rout = bi_finish(r, bn, 1); + return; + } size_t qn = an - bn + 1; ScrBigInt *q = bi_alloc(qn); ScrBigInt *r = rout ? bi_alloc(bn) : NULL; diff --git a/packages/runtime/src/scr_bytes.c b/packages/runtime/src/scr_bytes.c index 3926bf83e..e6df837c8 100644 --- a/packages/runtime/src/scr_bytes.c +++ b/packages/runtime/src/scr_bytes.c @@ -755,12 +755,29 @@ void scr_bytes_set_from(ScrBytes *dst, const ScrBytes *src, double offset) { } size_t width = scr_bytes_elem_size(dst->elem); if (dst->shared || src->shared) { - /* Convert a private copy of the source, then publish the target range. */ SCR_BYTES_SNAPSHOT(src); size_t bytes = src->len * width; + uint8_t *target = dst->data + (size_t)t * width; + if (dst->elem == src->elem) { + /* The same element kind copies the source's bytes, NaN payloads + * included. A source range that overlaps the target range (views of + * one buffer, where no snapshot was taken) is copied out first. */ + uintptr_t to = (uintptr_t)target, from = (uintptr_t)src->data; + bool overlap = to <= from ? from - to < bytes : to - from < bytes; + if (!overlap) { + scr_bytes_raw_out(dst, target, src->data, bytes); + return; + } + uint8_t *copy = scr_bytes_scratch(bytes); + scr_bytes_raw_in(src, copy, src->data, bytes); + scr_bytes_raw_out(dst, target, copy, bytes); + free(copy); + return; + } + /* Convert a private copy of the source, then publish the target range. */ uint8_t *scratch = scr_bytes_scratch(bytes); scr_bytes_convert_span(scratch, dst->elem, src->data, src->elem, src->len); - scr_bytes_raw_out(dst, dst->data + (size_t)t * width, scratch, bytes); + scr_bytes_raw_out(dst, target, scratch, bytes); free(scratch); return; } diff --git a/packages/runtime/src/scr_bytes.test.c b/packages/runtime/src/scr_bytes.test.c index 58f52fa63..8b2ff3cc1 100644 --- a/packages/runtime/src/scr_bytes.test.c +++ b/packages/runtime/src/scr_bytes.test.c @@ -37,6 +37,52 @@ static void external_set(void) { } } +/* Same-kind sets into and out of shared storage copy raw bytes, NaN + * payloads included, also between overlapping views of one shared buffer. */ +static void shared_set(void) { + enum { COUNT = 37 }; + for (ScrBytesElem elem = SCR_BYTES_U8; elem <= SCR_BYTES_F16; elem++) { + size_t width = scr_bytes_elem_size(elem), bytes = COUNT * width; + uint8_t pattern[COUNT * 8], expected[COUNT * 8]; + for (size_t i = 0; i < bytes; i++) pattern[i] = (uint8_t)(i * 37 + 0x7f); + /* Signaling NaNs, which a conversion through double would quiet. */ + const uint16_t nan16 = 0x7d01; + const uint32_t nan32 = 0x7fa00001; + const uint64_t nan64 = UINT64_C(0x7ff4000000000001); + for (size_t i = 1; i < COUNT; i += 6) { + if (elem == SCR_BYTES_F16) memcpy(pattern + i * width, &nan16, width); + if (elem == SCR_BYTES_F32) memcpy(pattern + i * width, &nan32, width); + if (elem == SCR_BYTES_F64) memcpy(pattern + i * width, &nan64, width); + } + for (int direction = 0; direction < 2; direction++) { + ScrBytes *shared = scr_bytes_new(elem, COUNT); + scr_bytes_make_shared(shared); + memcpy(shared->data, pattern, bytes); + ScrBytes *src = direction ? scr_bytes_subarray(shared, 3, COUNT) : scr_bytes_subarray(shared, 0, COUNT - 3); + double offset = direction ? 0 : 3; + memcpy(expected, pattern, bytes); + memmove(expected + (size_t)offset * width, pattern + (direction ? 3 : 0) * width, (COUNT - 3) * width); + scr_bytes_set_from(shared, src, offset); + assert(!scr_exc_pending()); + assert(memcmp(shared->data, expected, bytes) == 0); + scr_bytes_release(src); + + ScrBytes *local = scr_bytes_new(elem, COUNT); + memcpy(local->data, pattern, bytes); + ScrBytes *head = scr_bytes_subarray(local, 0, COUNT - 1); + scr_bytes_set_from(shared, head, 1); + memcpy(expected + width, pattern, bytes - width); + assert(memcmp(shared->data, expected, bytes) == 0); + memset(local->data, 0, bytes); + scr_bytes_set_from(local, shared, 0); + assert(memcmp(local->data, expected, bytes) == 0); + scr_bytes_release(head); + scr_bytes_release(local); + scr_bytes_release(shared); + } + } +} + static void copied_storage(void) { /* Same-kind copies preserve NaN payloads, signed zero and ownership. */ const uint64_t bits[] = {UINT64_C(0x7ff8000000000042), UINT64_C(0x8000000000000000), @@ -79,6 +125,7 @@ static void filled_storage(void) { void scr_bytes_test_bulk_storage(void) { long before = scr_bytes_live_count(); external_set(); + shared_set(); copied_storage(); filled_storage(); assert(scr_bytes_live_count() == before); diff --git a/packages/runtime/src/scr_dyn.h b/packages/runtime/src/scr_dyn.h index 452eb0363..82fe1a6d9 100644 --- a/packages/runtime/src/scr_dyn.h +++ b/packages/runtime/src/scr_dyn.h @@ -681,6 +681,14 @@ void scr_dyn_copy_entry(ScrDyn *target, const ScrDyn *source, const ScrDynEntry bool scr_dyn_field_eq_str(const ScrDyn *d, const ScrStr *key, const ScrStr *value); bool scr_dyn_field_eq_num(const ScrDyn *d, const ScrStr *key, double value); bool scr_dyn_field_eq_bool(const ScrDyn *d, const ScrStr *key, bool value); +/* The case a union's checked match takes from its string discriminant: the + * index in `values` of the string `d`'s own data property `key` holds, or + * -1 when every literal test above fails (`d` is not an object, or its own + * data property `key` holds no string among `values`), or -2 when only the + * tests themselves can tell (a typed capsule, an accessor, or a property + * `d` does not own). Borrowed args. */ +int32_t scr_dyn_discriminant_case(const ScrDyn *d, const ScrStr *key, ScrStr *const *values, + size_t count); /* Object.keys/values/entries over the checked-dynamic tree: JS own-key order (array-index * keys ascending first), dyn-array results (+1); values/entries RETAIN @@ -1189,6 +1197,10 @@ void scr_weak_dispose(void *object); double scr_dyn_arr_len(const ScrDyn *d); ScrDyn *scr_dyn_arr_at(const ScrDyn *d, double i); ScrDyn *scr_dyn_arr_named_get(const ScrDyn *d, const ScrStr *key); +/* The member `key` of an array found before Array.prototype: an own named + * property, or one on its custom prototype chain (Object.setPrototypeOf). + * NULL when the chain reaches Array.prototype without it. +1. */ +ScrDyn *scr_dyn_arr_custom_member(const ScrDyn *d, const char *key, size_t key_len); void scr_dyn_obj_set(ScrDyn *obj, const char *key, size_t key_len, ScrDyn *value); void scr_dyn_obj_set_field(ScrDyn *obj, const char *key, size_t key_len, ScrDyn *value); ScrDyn *scr_dyn_define_property(ScrDyn *target, ScrDyn *key, ScrDyn *descriptor); @@ -1391,6 +1403,9 @@ ScrDyn *scr_dyn_invoke_prepared(ScrDyn *recv, ScrDyn *callee, const char *method * *mid_slot, the site's -1-initialized slot, caches the name's method id * after the first call, so later calls skip the name lookup. */ ScrDyn *scr_dyn_invoke_prepared_id(ScrDyn *recv, ScrDyn *callee, const char *method, int32_t *mid_slot, ScrDyn *const *args, size_t argc, const char *what); +/* scr_dyn_invoke_prepared_id for a call whose result is discarded: a plain + * array's push appends without boxing its new length. */ +void scr_dyn_invoke_discard_id(ScrDyn *recv, ScrDyn *callee, const char *method, int32_t *mid_slot, ScrDyn *const *args, size_t argc, const char *what); /* scr_dyn_invoke_prepared_id for Buffer.prototype method names * (scr_bytes_dyn.c), one entry per method family: a Buffer receiver runs * the method, any other takes the ordinary dispatch. */ diff --git a/packages/runtime/src/scr_dyn_invoke.c b/packages/runtime/src/scr_dyn_invoke.c index ebbfa8411..4901fa694 100644 --- a/packages/runtime/src/scr_dyn_invoke.c +++ b/packages/runtime/src/scr_dyn_invoke.c @@ -775,6 +775,8 @@ ScrDyn *scr_dyn_prepare_method(ScrDyn *recv, const char *method) { return scr_dyn_invoke(recv, method, NULL, 0, method); if (recv->kind == SCR_DYN_OBJ) return scr_dyn_obj_read(recv, method, strlen(method)); + if (recv->kind == SCR_DYN_ARR && (recv->v.arr.properties || recv->prototype)) + return scr_dyn_arr_custom_member(recv, method, strlen(method)); if (recv->kind == SCR_DYN_PROXY) { ScrStr *key = scr_str_new(method, strlen(method)); ScrDyn *result = scr_dyn_proxy_get(recv, key); @@ -799,20 +801,117 @@ ScrDyn *scr_dyn_invoke(ScrDyn *recv, const char *method, return scr_dyn_invoke_impl(recv, method, dyn_method_of(method), args, argc, what, NULL); } -ScrDyn *scr_dyn_invoke_prepared_id(ScrDyn *recv, ScrDyn *callee, const char *method, - int32_t *mid_slot, ScrDyn *const *args, size_t argc, - const char *what) { - if (callee) return scr_dyn_invoke_prepared(recv, callee, method, args, argc, what); - /* The slot is the call site's, shared by every thread: each writes the - * same id. */ +/* A dense array without holes, named properties or a prototype of its + * own, whose length and elements are writable: its common methods below + * run without the dispatch ladder, which answers every other receiver and + * method as before. */ +static bool dyn_arr_plain(const ScrDyn *recv) { + return recv->kind == SCR_DYN_ARR && !recv->v.arr.presence && !recv->v.arr.properties && + !recv->prototype && !recv->null_proto && !recv->non_extensible && + !recv->v.arr.sealed && !recv->v.arr.frozen && !scr_dyn_is_array_prototype(recv); +} + +/* scr_dyn_strict_eq, with the number and string cases inline: an element + * search compares every element. */ +static inline bool dyn_elem_strict_eq(const ScrDyn *a, const ScrDyn *b) { + if (a->kind == SCR_DYN_STR && b->kind == SCR_DYN_STR) { + const ScrStr *x = a->v.str, *y = b->v.str; + return x == y || (x->len == y->len && memcmp(x->data, y->data, x->len) == 0); + } + if (a->kind == SCR_DYN_NUM && b->kind == SCR_DYN_NUM) return a->v.num == b->v.num; + return scr_dyn_strict_eq(a, b); +} + +/* push, pop, shift, at, includes and indexOf of a plain array, with the + * ladder's results. An argument that would run user code (a non-number + * index, a fromIndex) leaves the call to the ladder. True when answered. */ +static bool dyn_arr_fast(ScrDyn *recv, int mid, ScrDyn *const *args, size_t argc, + ScrDyn **result) { + size_t len = recv->v.arr.len; + switch (mid) { + case DYN_M_push: + for (size_t i = 0; i < argc; i++) scr_dyn_arr_push(recv, scr_dyn_retain(args[i])); + *result = scr_dyn_new_num((double)recv->v.arr.len); + return true; + case DYN_M_pop: + if (len == 0) { + *result = scr_dyn_retain(scr_dyn_undefined()); + return true; + } + recv->v.arr.len = len - 1; + *result = recv->v.arr.items[len - 1]; /* ownership moves out */ + return true; + case DYN_M_shift: + if (len == 0) { + *result = scr_dyn_retain(scr_dyn_undefined()); + return true; + } + *result = recv->v.arr.items[0]; /* ownership moves out */ + memmove(recv->v.arr.items, recv->v.arr.items + 1, (len - 1) * sizeof(ScrDyn *)); + recv->v.arr.len = len - 1; + return true; + case DYN_M_at: { + if (argc == 0 || args[0]->kind != SCR_DYN_NUM) return false; + double index = args[0]->v.num; + index = isnan(index) ? 0 : trunc(index); + if (index < 0) index += (double)len; + *result = index < 0 || index >= (double)len ? scr_dyn_retain(scr_dyn_undefined()) + : scr_dyn_retain(recv->v.arr.items[(size_t)index]); + return true; + } + case DYN_M_includes: + case DYN_M_indexOf: { + if (argc > 1) return false; + ScrDyn *needle = argc > 0 ? args[0] : scr_dyn_undefined(); + bool nan = mid == DYN_M_includes && needle->kind == SCR_DYN_NUM && needle->v.num != needle->v.num; + for (size_t i = 0; i < len; i++) { + const ScrDyn *item = recv->v.arr.items[i]; + if (nan ? item->kind == SCR_DYN_NUM && item->v.num != item->v.num + : dyn_elem_strict_eq(item, needle)) { + *result = mid == DYN_M_includes ? scr_dyn_new_bool(true) : scr_dyn_new_num((double)i); + return true; + } + } + *result = mid == DYN_M_includes ? scr_dyn_new_bool(false) : scr_dyn_new_num(-1); + return true; + } + default: + return false; + } +} + +/* The call site's method id. The slot is the site's, shared by every + * thread: each writes the same id. */ +static int32_t dyn_site_method(const char *method, int32_t *mid_slot) { int32_t mid = __atomic_load_n(mid_slot, __ATOMIC_RELAXED); if (mid < 0) { mid = dyn_method_of(method); __atomic_store_n(mid_slot, mid, __ATOMIC_RELAXED); } + return mid; +} + +ScrDyn *scr_dyn_invoke_prepared_id(ScrDyn *recv, ScrDyn *callee, const char *method, + int32_t *mid_slot, ScrDyn *const *args, size_t argc, + const char *what) { + if (callee) return scr_dyn_invoke_prepared(recv, callee, method, args, argc, what); + int32_t mid = dyn_site_method(method, mid_slot); + ScrDyn *result; + if (dyn_arr_plain(recv) && dyn_arr_fast(recv, mid, args, argc, &result)) return result; return scr_dyn_invoke_impl(recv, method, mid, args, argc, what, NULL); } +void scr_dyn_invoke_discard_id(ScrDyn *recv, ScrDyn *callee, const char *method, + int32_t *mid_slot, ScrDyn *const *args, size_t argc, + const char *what) { + /* A plain array's push whose new length nobody reads. */ + if (!callee && dyn_arr_plain(recv) && dyn_site_method(method, mid_slot) == DYN_M_push) { + for (size_t i = 0; i < argc; i++) scr_dyn_arr_push(recv, scr_dyn_retain(args[i])); + return; + } + scr_dyn_release(scr_dyn_invoke_prepared_id(recv, callee, method, mid_slot, args, argc, what)); +} + static ScrDyn *scr_dyn_invoke_impl( ScrDyn *recv, const char *method, int mid, ScrDyn *const *args, size_t argc, const char *what, ScrDyn *callback_recv) { @@ -1125,6 +1224,14 @@ static ScrDyn *scr_dyn_invoke_impl( if (frozen_op >= 0 && scr_array_frozen_error(frozen_op, len, (double)splice_start, frozen_op == 10 ? (double)splice_removed : items, items)) return NULL; + if ((mid == DYN_M_push) && grows && recv->non_extensible) { + scr_array_not_extensible_error((double)len); + return NULL; + } + if ((mid == DYN_M_pop) && shrinks && recv->v.arr.sealed) { + scr_array_undeletable_error((double)(len - 1)); + return NULL; + } if ((grows && recv->non_extensible) || (shrinks && recv->v.arr.sealed && !(mid == DYN_M_shift))) { static const char message[] = "Cannot modify a sealed or frozen array"; scr_throw_error_msg(SCR_ERR_TYPE, message, sizeof message - 1); @@ -1154,8 +1261,7 @@ static ScrDyn *scr_dyn_invoke_impl( scr_dyn_release(recv->v.arr.items[i - 1]); recv->v.arr.items[i - 1] = scr_dyn_retain(recv->v.arr.items[i]); } - static const char message[] = "Cannot delete non-configurable property"; - scr_throw_error_msg(SCR_ERR_TYPE, message, sizeof message - 1); + scr_array_undeletable_error((double)(len - 1)); return NULL; } ScrDyn *first = recv->v.arr.items[0]; diff --git a/packages/runtime/src/scr_idna.c b/packages/runtime/src/scr_idna.c index 34c9e3882..7507e5caa 100644 --- a/packages/runtime/src/scr_idna.c +++ b/packages/runtime/src/scr_idna.c @@ -1,8 +1,8 @@ /* UTS #46 domain processing for node:url — the conversions Node applies to - * url.parse() hostnames (ToASCII), url.domainToASCII()/domainToUnicode() - * and url.format()'s unicode option: nontransitional mapping, NFC, label - * validity (leading combining marks, ContextJ, and the RFC 5893 bidi rule), - * and Punycode. It matches Node 24's implementation, including its + * WHATWG URL hosts and url.parse() hostnames (ToASCII), + * url.domainToASCII()/domainToUnicode() and url.format()'s unicode option: + * nontransitional mapping, NFC, label validity (leading combining marks, + * ContextJ, and the RFC 5893 bidi rule), and Punycode. It matches Node 24's implementation, including its * validity-data snapshot and the shortcuts its label checks take; the * differential corpus and test/url-legacy-cases.txt (generated from Node) * are the oracle. @@ -511,3 +511,6 @@ ScrStr *scr_idna_to_unicode(const char *p, size_t n) { } return scr_jb_finish(&out); } + +/* The URL host parser's UTS #46 ToASCII (scr_url.c). */ +void scr_url_idna_install(void) { scr_url_host_to_ascii = &scr_idna_to_ascii; } diff --git a/packages/runtime/src/scr_js.h b/packages/runtime/src/scr_js.h index 6fa20fdb8..1d874b31e 100644 --- a/packages/runtime/src/scr_js.h +++ b/packages/runtime/src/scr_js.h @@ -195,10 +195,13 @@ bool scr_js_truthy(ScrJsVal value); * index and which keys an object lacks. * * Each property site of compiled dynamic-tier code owns one zeroed, - * thread-local ScrJsIc. The code at the site is fixed: it tests entry 0 - * inline and calls a probe for everything else. Probes never run user - * code: they answer from plain data properties or return a miss (NULL / - * false) with nothing done, and the site then runs its generic path. + * thread-local ScrJsIc with static storage: the reset that frees a + * thread's shapes (at thread exit, and when a library session ends) clears + * every record that holds entries. The code at the site is fixed: it tests + * entry 0 inline and calls a probe for everything else. Probes never run + * user code: they answer from plain data properties or return a miss + * (NULL / false) with nothing done, and the site then runs its generic + * path. * * load entry: shape = receiver shape; target = NULL for an own property * (slot indexes the receiver's entries), else the shape of diff --git a/packages/runtime/src/scr_js_object.c b/packages/runtime/src/scr_js_object.c index 63929ab57..0e0ec0908 100644 --- a/packages/runtime/src/scr_js_object.c +++ b/packages/runtime/src/scr_js_object.c @@ -18,6 +18,7 @@ #include "scr_key.h" #include +#include /* The compiler's inline cache tests read these in place (64-bit targets; * backend/llvm/js-ic.ts). */ @@ -37,10 +38,10 @@ _Static_assert(sizeof(ScrJsIcEntry) == 40 && sizeof(ScrJsIc) == 168, "cache layo * take the generic path and its hash index. */ #define SCR_JS_SHAPE_MAX 64 -/* Shapes live until the thread exits, so a thread creates at most this - * many, holding at most this many key bytes. Once either budget is spent, - * objects whose key sequence has no shape yet stay unshaped; shapes that - * exist keep serving. */ +/* Shapes live until the thread exits or a library session ends, so a thread + * creates at most this many, holding at most this many key bytes. Once + * either budget is spent, objects whose key sequence has no shape yet stay + * unshaped; shapes that exist keep serving. */ #define SCR_JS_SHAPE_BUDGET ((size_t)1 << 16) #define SCR_JS_SHAPE_KEY_BYTES_BUDGET ((size_t)4 << 20) @@ -63,12 +64,25 @@ static SCR_TL size_t scr_js_transition_count; static SCR_TL size_t scr_js_shape_key_bytes; static SCR_TL bool scr_js_shapes_registered; +/* The cache records that hold entries. Entries name shapes and prototype + * objects, so the reset that frees the shapes also clears these records: a + * library's next session runs the same sites with the same records. */ +static SCR_TL ScrJsIc **scr_js_filled_ics; +static SCR_TL size_t scr_js_filled_count; +static SCR_TL size_t scr_js_filled_capacity; + /* The shape word of an object whose key sequence has no shape within the * budgets. No cache entry names it, so its accesses take the probes, * which answer with a miss until the key set changes. */ static ScrJsShape scr_js_unshaped; static void scr_js_shapes_reset(void) { + for (size_t i = 0; i < scr_js_filled_count; i++) + memset(scr_js_filled_ics[i], 0, sizeof *scr_js_filled_ics[i]); + free(scr_js_filled_ics); + scr_js_filled_ics = NULL; + scr_js_filled_count = 0; + scr_js_filled_capacity = 0; while (scr_js_all_shapes) { ScrJsShape *shape = scr_js_all_shapes; scr_js_all_shapes = shape->all_next; @@ -192,6 +206,16 @@ static size_t scr_js_shape_find(const ScrJsShape *shape, const char *key, size_t /* Fills a cache entry: `stale` (a matching entry whose hit failed) when * there is one, else a free or replaceable one. */ static void scr_js_ic_add(ScrJsIc *ic, uint32_t stale, ScrJsIcEntry entry) { + if (ic->count == 0) { + if (scr_js_filled_count == scr_js_filled_capacity) { + size_t capacity = scr_js_filled_capacity ? scr_js_filled_capacity * 2 : 64; + ScrJsIc **grown = realloc(scr_js_filled_ics, capacity * sizeof *grown); + if (!grown) scr_trap("scriptc: out of memory\n"); + scr_js_filled_ics = grown; + scr_js_filled_capacity = capacity; + } + scr_js_filled_ics[scr_js_filled_count++] = ic; + } uint32_t index; if (stale < ic->count) { index = stale; diff --git a/packages/runtime/src/scr_js_object.test.c b/packages/runtime/src/scr_js_object.test.c index a14aa9ff6..6b6934dea 100644 --- a/packages/runtime/src/scr_js_object.test.c +++ b/packages/runtime/src/scr_js_object.test.c @@ -24,11 +24,17 @@ static ScrDyn *object_with(const ScrStr *key, double n) { return object; } +/* Cache records have static storage, as the compiled sites' do: the reset + * at exit clears every record that holds entries. Each check starts from + * an empty record. */ +static ScrJsIc load_ic; +static ScrJsIc store_ic; + /* Whether a load probe answers `object.key` and returns `n`. */ static bool probe_reads(ScrDyn *object, const ScrStr *key, double n) { - ScrJsIc ic; - memset(&ic, 0, sizeof ic); - ScrDyn *value = scr_js_load_probe(object, key, &ic); + ScrJsIc *ic = &load_ic; + memset(ic, 0, sizeof *ic); + ScrDyn *value = scr_js_load_probe(object, key, ic); if (!value) return false; assert(value->kind == SCR_DYN_NUM && value->v.num == n); scr_dyn_release(value); @@ -66,14 +72,13 @@ static void budget_bounds_shapes(size_t key_len, size_t budget) { /* A store probe that would add a new shape leaves the add to the generic * store; one that replaces through an existing shape still runs. */ - ScrJsIc ic; - memset(&ic, 0, sizeof ic); + memset(&store_ic, 0, sizeof store_ic); ScrStr *added = scr_str_new("added", 5); ScrDyn *value = scr_dyn_new_num(8); - assert(!scr_js_store_probe(object, added, value, &ic)); + assert(!scr_js_store_probe(object, added, value, &store_ic)); assert(object->v.obj.len == 1); - memset(&ic, 0, sizeof ic); - assert(scr_js_store_probe(object, first, value, &ic)); + memset(&store_ic, 0, sizeof store_ic); + assert(scr_js_store_probe(object, first, value, &store_ic)); assert(probe_reads(object, first, 8)); scr_dyn_release(value); scr_str_release(added); @@ -90,10 +95,9 @@ static void immortal_objects_stay_unshaped(void) { object->rc = SIZE_MAX; assert(!probe_reads(object, key, 3)); assert(object->v.obj.shape == NULL); - ScrJsIc ic; - memset(&ic, 0, sizeof ic); + memset(&store_ic, 0, sizeof store_ic); ScrDyn *value = scr_dyn_new_num(4); - assert(!scr_js_store_probe(object, key, value, &ic)); + assert(!scr_js_store_probe(object, key, value, &store_ic)); assert(object->v.obj.shape == NULL); scr_dyn_release(value); scr_str_release(key); diff --git a/packages/runtime/src/scr_json.c b/packages/runtime/src/scr_json.c index fd6db7ce8..49af072ed 100644 --- a/packages/runtime/src/scr_json.c +++ b/packages/runtime/src/scr_json.c @@ -1286,10 +1286,14 @@ void scr_dyn_release(ScrDyn *d) { return; } #ifndef SCR_RC_AUDIT - /* Number and string boxes own at most one string, are never cycle - * candidates and cannot be weak keys: return them to the free list. */ + /* Number and string boxes own at most one string and cannot be weak + * keys: return them to the free list. Their releases never buffer them, + * but a collector pass can (a box that traces its payload reaches them, + * and scr_xg_release buffers what it settles), so a buffered box leaves + * its candidate buffer first. */ if ((d->kind == SCR_DYN_NUM || d->kind == SCR_DYN_STR) && !d->prototype && !d->symbol_properties && !d->symbol_keys && scr_dyn_free_count < SCR_DYN_FREE_MAX) { + scr_cyc_dead(d); if (d->kind == SCR_DYN_STR) scr_str_release(d->v.str); d->v.str = (ScrStr *)scr_dyn_free_misc; scr_dyn_free_misc = d; @@ -1435,6 +1439,24 @@ bool scr_dyn_field_eq_num(const ScrDyn *d, const ScrStr *key, double value) { return matches; } +int32_t scr_dyn_discriminant_case(const ScrDyn *d, const ScrStr *key, ScrStr *const *values, + size_t count) { + if (!d) return -1; + if (d->kind == SCR_DYN_TYPED_REF) return -2; + if (d->kind != SCR_DYN_OBJ) return -1; + for (size_t i = 0; i < d->v.obj.len; i++) { + const ScrDynEntry *entry = &d->v.obj.entries[i]; + if (entry->key_len != key->len || memcmp(entry->key, key->data, key->len) != 0) continue; + if (entry->accessor) return -2; + const ScrDyn *member = entry->value; + if (!member || member->kind != SCR_DYN_STR) return -1; + for (size_t k = 0; k < count; k++) + if (scr_str_eq(member->v.str, values[k])) return (int32_t)k; + return -1; + } + return -2; +} + bool scr_dyn_field_eq_bool(const ScrDyn *d, const ScrStr *key, bool value) { ScrDyn *member = scr_dyn_discriminant(d, key); bool matches = member && member->kind == SCR_DYN_BOOL && member->v.b == value; @@ -1717,6 +1739,21 @@ ScrDyn *scr_dyn_arr_at(const ScrDyn *d, double i) { return result; } +ScrDyn *scr_dyn_arr_custom_member(const ScrDyn *d, const char *key, size_t key_len) { + if (d->v.arr.properties && scr_dyn_obj_get(d->v.arr.properties, key, key_len)) + return scr_dyn_obj_read_receiver(d->v.arr.properties, key, key_len, d); + for (const ScrDyn *link = d->prototype; link && link != scr_builtin_array_prototype; + link = link->prototype) { + const ScrDyn *table = link->kind == SCR_DYN_OBJ ? link + : link->kind == SCR_DYN_ARR ? link->v.arr.properties + : NULL; + if (table && scr_dyn_obj_get(table, key, key_len)) + return scr_dyn_obj_read_receiver(table, key, key_len, d); + if (link->kind != SCR_DYN_OBJ && link->kind != SCR_DYN_ARR) break; + } + return NULL; +} + ScrDyn *scr_dyn_arr_named_get(const ScrDyn *d, const ScrStr *key) { if (d->v.arr.properties && scr_dyn_obj_get(d->v.arr.properties, key->data, key->len)) return scr_dyn_obj_read_receiver(d->v.arr.properties, key->data, key->len, d); diff --git a/packages/runtime/src/scr_json.test.c b/packages/runtime/src/scr_json.test.c index a28759876..29ba4bfe3 100644 --- a/packages/runtime/src/scr_json.test.c +++ b/packages/runtime/src/scr_json.test.c @@ -391,6 +391,30 @@ static bool buffered_after_read(ScrDyn *d) { return buffered(d); } +/* A collector pass can buffer a number or string box that it reached + * through an owner that traces it: scr_xg_release, settling an edge a + * restricted pass skipped, calls scr_cyc_on_release on the target. The + * box's last release must take it out of its candidate buffer before its + * block returns to the free list, which hands it to the next box. */ +static void buffered_leaf_release(void) { + for (uint16_t gen = SCR_CYC_NURSERY; gen < SCR_CYC_NGENS; gen++) { + ScrStr *text = scr_str_new("buffered", 8); + ScrDyn *leaves[2] = {scr_dyn_new_str(text), scr_dyn_new_num(0.5)}; + scr_str_release(text); + for (int i = 0; i < 2; i++) { + scr_cyc_set_gen(scr_cyc_hdr(leaves[i]), gen); + scr_cyc_on_release(leaves[i]); + scr_dyn_release(leaves[i]); + } + ScrDyn *fresh[2] = {scr_dyn_new_num(1.5), scr_dyn_new_num(2.5)}; + assert(!buffered(fresh[0]) && !buffered(fresh[1])); + scr_collect_cycles(); + assert(fresh[0]->v.num == 1.5 && fresh[1]->v.num == 2.5); + scr_dyn_release(fresh[0]); + scr_dyn_release(fresh[1]); + } +} + static ScrDyn *parse_text(const char *text) { ScrStr *input = scr_str_new(text, strlen(text)); ScrDyn *parsed = scr_json_parse(input); @@ -823,6 +847,7 @@ int main(void) { shared_property_keys(); checked_leaf_cycles(); checked_number_storage(); + buffered_leaf_release(); json_intact_regions(); json_schema_parse(); json_schema_special_names(); diff --git a/packages/runtime/src/scr_url.c b/packages/runtime/src/scr_url.c index 09ccb87fe..a61286c63 100644 --- a/packages/runtime/src/scr_url.c +++ b/packages/runtime/src/scr_url.c @@ -21,10 +21,8 @@ * case preserved, C0-encoded), a single "/" roots a normalized path, * anything else is an OPAQUE path with the C0 control set encoded * (data:text/plain,hi there) - * - * DOCUMENTED DIVERGENCE: a special host that holds non-ASCII text after - * percent-decoding is rejected with "Invalid URL" (no UTS #46 processing), - * and "xn--" labels are kept without Punycode validation. + * - UTS #46 processing of non-ASCII hosts through the IDNA unit + * (scr_url_idna_install) * * Failures THROW catchable TypeErrors through the exception cell with * Node's messages; callers are compiler-emitted pending checks. */ @@ -430,16 +428,21 @@ static int hex_val(char c) { return -1; } +/* UTS #46 ToASCII (scr_idna.c), which programs that use URLs install at + * startup (scr_url_idna_install). NULL refuses hosts that hold non-ASCII + * text. */ +ScrStr *(*scr_url_host_to_ascii)(const char *p, size_t n) = NULL; + /* The URL Standard's host parser for the constructor and the setters, * with canonical IPv6 serialization: * - a non-special (opaque) host refuses forbidden host code points and * percent-encodes the C0 control set; - * - a special host is percent-decoded, lowercased, refused when it holds a - * forbidden domain code point, and serialized as dotted decimal when it - * ends in a number (or refused when that number is invalid). - * Divergence: a special host that decodes to non-ASCII needs UTS #46 - * processing, which this unit does not link, so it is refused; "xn--" - * labels are kept without Punycode validation. */ + * - a special host is percent-decoded and lowercased; when that leaves + * non-ASCII text it goes through UTS #46 ToASCII (as Node's host parser + * does, an ASCII host, "xn--" labels included, is only lowercased). The + * result is refused when it holds a forbidden domain code point, and + * serialized as dotted decimal when it ends in a number (or refused when + * that number is invalid). */ static bool parse_host(const char *hp, size_t host_len, bool special, bool is_file, ScrStr **host) { bool ipv6 = host_len > 0 && hp[0] == '['; @@ -467,6 +470,7 @@ static bool parse_host(const char *hp, size_t host_len, bool special, bool is_fi *host = ub_take(&hb); return true; } + bool non_ascii = false; for (size_t i = 0; i < host_len; i++) { unsigned char c = (unsigned char)hp[i]; int hi, lo; @@ -475,12 +479,27 @@ static bool parse_host(const char *hp, size_t host_len, bool special, bool is_fi c = (unsigned char)(hi * 16 + lo); i += 2; } - if (url_host_byte_class[c] & URL_FORBIDDEN_DOMAIN) { + if (c >= 0x80) non_ascii = true; + if (c >= 'A' && c <= 'Z') c = (unsigned char)(c - 'A' + 'a'); + ub_push(&hb, (char)c); + } + if (non_ascii) { + /* UTS #46 ToASCII, or refused without the IDNA unit. */ + ScrStr *ascii = scr_url_host_to_ascii ? scr_url_host_to_ascii(hb.data, hb.len) : NULL; + free(hb.data); + if (!ascii || ascii->len == 0) { + scr_str_release(ascii); + return false; + } + ub_init(&hb); + ub_append(&hb, ascii->data, ascii->len); + scr_str_release(ascii); + } + for (size_t i = 0; i < hb.len; i++) { + if (url_host_byte_class[(unsigned char)hb.data[i]] & URL_FORBIDDEN_DOMAIN) { free(hb.data); return false; } - if (c >= 'A' && c <= 'Z') c = (unsigned char)(c - 'A' + 'a'); - ub_push(&hb, (char)c); } if (url_ends_in_number(hb.data, hb.len)) { UrlBuf ip; diff --git a/packages/runtime/src/scr_url.h b/packages/runtime/src/scr_url.h index c7c9c63c8..13a7d093b 100644 --- a/packages/runtime/src/scr_url.h +++ b/packages/runtime/src/scr_url.h @@ -46,6 +46,10 @@ void scr_url_set_checked(ScrUrl *u, ScrStr *field, const ScrDyn *value); /* borr * null and has no lowering for an object that is not a URL. Borrowed; +1, * or NULL with the error pending. */ ScrDyn *scr_url_to_http_options(ScrUrl *u); +/* UTS #46 ToASCII for the host parser: NULL until scr_url_idna_install + * (scr_idna.c, emitted at startup by programs that use URLs) sets it. */ +extern ScrStr *(*scr_url_host_to_ascii)(const char *p, size_t n); +void scr_url_idna_install(void); ScrDyn *scr_url_to_http_options_dyn(const ScrDyn *value); /* node:url's legacy surface (scr_url_legacy.c — linked with the regex, * querystring and dyn-async units). url.parse() answers a checked-dynamic diff --git a/scripts/zero-cost-compare.mjs b/scripts/zero-cost-compare.mjs index 160e33438..a1ab3ca60 100644 --- a/scripts/zero-cost-compare.mjs +++ b/scripts/zero-cost-compare.mjs @@ -394,8 +394,9 @@ export function dynamicTierSymbols(names) { * not yet traced to its typed root. bench/thread-slots, * bench/channel-messages and bench/worker-messages, added later, link them * through worker messages: - * node:worker_threads values are dyn objects. bench/dynamic-methods is - * untyped JavaScript. Every other bench workload is fully typed and must link none. The gate and + * node:worker_threads values are dyn objects. bench/json-unions, added + * later too, decodes JSON.parse trees. bench/dynamic-methods and + * bench/dynamic-properties are untyped JavaScript. Every other bench workload is fully typed and must link none. The gate and * tests/harness/bench-dynamic-tier.test.ts enforce that, and the test also * fails when a listed workload stops linking them, so remove a workload here * when a leak fix makes its binary clean. */ @@ -410,10 +411,12 @@ export const DYNAMIC_WORKLOADS = new Set([ "bench/thread-slots", "bench/channel-messages", "bench/dynamic-methods", + "bench/dynamic-properties", "bench/worker-messages", "bench/typed-array-scan-workers", "bench/log-summary", "bench/inventory-report", + "bench/json-unions", ]); /** Names of the functions (text and weak symbols) in `nm` output. Mach-O diff --git a/tests/corpus/6305-json-discriminant-dispatch.ts b/tests/corpus/6305-json-discriminant-dispatch.ts new file mode 100644 index 000000000..ac4d992da --- /dev/null +++ b/tests/corpus/6305-json-discriminant-dispatch.ts @@ -0,0 +1,123 @@ +// JSON cast into a union with many string discriminant literals: the cast +// reads the discriminant once and tests only the arm whose literal it +// holds, plus the arms that are not records. An arm can list several +// literals, and values that are not objects reach the other arms in +// canonical order. Valid casts only (see 1002-json-parse-cast.ts); JSON keys +// and record fields alphabetical for stringify parity. + +type Expr = + | { kind: "num"; value: number } + | { kind: "str"; text: string } + | { kind: "neg"; operand: Expr } + | { kind: "add" | "mul"; left: Expr; right: Expr } + | { args: Expr[]; kind: "call"; name: string } + | { kind: "method"; name: string; target: Expr } + | { kind: "list"; items: Expr[] }; + +function show(e: Expr): string { + switch (e.kind) { + case "num": + return String(e.value); + case "str": + return JSON.stringify(e.text); + case "neg": + return `-(${show(e.operand)})`; + case "add": + return `(${show(e.left)} + ${show(e.right)})`; + case "mul": + return `(${show(e.left)} * ${show(e.right)})`; + case "call": + return `${e.name}(${e.args.map(show).join(", ")})`; + case "method": + return `${show(e.target)}.${e.name}()`; + case "list": + return `[${e.items.map(show).join(", ")}]`; + } +} + +function evaluate(e: Expr): number { + switch (e.kind) { + case "num": + return e.value; + case "str": + return e.text.length; + case "neg": + return -evaluate(e.operand); + case "add": + return evaluate(e.left) + evaluate(e.right); + case "mul": + return evaluate(e.left) * evaluate(e.right); + case "call": + return e.args.reduce((sum, arg) => sum + evaluate(arg), 0); + case "method": + return evaluate(e.target); + case "list": + return e.items.length; + } +} + +const source = + '{"kind":"add","left":{"kind":"mul","left":{"kind":"num","value":3},' + + '"right":{"kind":"neg","operand":{"kind":"num","value":4}}},' + + '"right":{"args":[{"kind":"str","text":"abc"},{"kind":"num","value":2}],"kind":"call","name":"sum"}}'; +const tree = JSON.parse(source) as Expr; +console.log(show(tree)); +console.log(evaluate(tree)); +console.log(JSON.stringify(tree) === source); + +const method = JSON.parse( + '{"kind":"method","name":"size","target":{"items":[{"kind":"num","value":1}],"kind":"list"}}', +) as Expr; +console.log(show(method), evaluate(method)); + +// A deep list of every arm, built from text. +const parts: string[] = []; +for (let i = 0; i < 200; i++) { + parts.push( + i % 3 === 0 + ? `{"kind":"num","value":${i}}` + : i % 3 === 1 + ? `{"kind":"neg","operand":{"kind":"num","value":${i}}}` + : `{"kind":"mul","left":{"kind":"num","value":${i}},"right":{"kind":"str","text":"xy"}}`, + ); +} +const list = JSON.parse(`{"items":[${parts.join(",")}],"kind":"list"}`) as Expr; +if (list.kind === "list") { + let total = 0; + for (const item of list.items) total += evaluate(item); + console.log(list.items.length, total); +} + +// Arms that are not records: a value that is not an object can only be one +// of them. +type Token = + | { kind: "ident"; name: string } + | { kind: "number"; value: number } + | { kind: "string"; value: string } + | { kind: "punct"; symbol: string } + | { kind: "comment"; line: number; text: string } + | string + | number; + +function describe(token: Token): string { + if (typeof token === "string") return `raw ${token}`; + if (typeof token === "number") return `count ${token}`; + switch (token.kind) { + case "ident": + return `ident ${token.name}`; + case "number": + return `number ${token.value}`; + case "string": + return `string ${JSON.stringify(token.value)}`; + case "punct": + return `punct ${token.symbol}`; + case "comment": + return `comment ${token.line}: ${token.text}`; + } +} + +const tokens = JSON.parse( + '[{"kind":"ident","name":"x"},{"kind":"punct","symbol":"="},{"kind":"number","value":42},' + + '{"kind":"string","value":"hi"},{"kind":"comment","line":3,"text":"note"},"spare",7]', +) as Token[]; +for (const token of tokens) console.log(describe(token)); diff --git a/tests/corpus/6306-discriminant-getter-dispatch.ts b/tests/corpus/6306-discriminant-getter-dispatch.ts new file mode 100644 index 000000000..4f0a1244f --- /dev/null +++ b/tests/corpus/6306-discriminant-getter-dispatch.ts @@ -0,0 +1,55 @@ +// A checked cast into a union with many string discriminant literals reads +// the discriminant once, but a getter that runs while an arm's fields are +// checked can change it. The cast then continues with the arms after the +// one that failed, testing each arm's literal again, and matches the arm +// the new discriminant names. Each getter changes the discriminant on its +// first call only; the program reads the getter once itself, so Node (where +// the cast runs no code) reaches the same state. Valid casts only (see +// 1002-json-parse-cast.ts). + +type Shape = + | { kind: "circle"; radius: number } + | { kind: "square"; side: number } + | { kind: "rect"; height: number; width: number } + | { kind: "tri"; a: number; b: number; c: number }; + +function area(shape: Shape): number { + switch (shape.kind) { + case "circle": + return Math.round(Math.PI * shape.radius * shape.radius); + case "square": + return shape.side * shape.side; + case "rect": + return shape.width * shape.height; + case "tri": { + const s = (shape.a + shape.b + shape.c) / 2; + return Math.sqrt(s * (s - shape.a) * (s - shape.b) * (s - shape.c)); + } + } +} + +// The circle arm reads `radius`, whose getter turns the value into a square. +let squareReads = 0; +const square: any = { + kind: "circle", + get radius() { + squareReads++; + if (squareReads === 1) square.kind = "square"; + return "none"; + }, + side: 3, +}; +const asSquare = square as Shape; +void square.radius; +console.log(square.kind, asSquare.kind, area(asSquare)); + +// Values without getters take the arm their discriminant names. +const plain: any[] = [ + { kind: "rect", width: 2, height: 5 }, + { kind: "tri", a: 3, b: 4, c: 5 }, + { kind: "circle", radius: 2 }, +]; +for (const value of plain) { + const shape = value as Shape; + console.log(shape.kind, area(shape)); +} diff --git a/tests/corpus/7984-untyped-array-methods.mjs b/tests/corpus/7984-untyped-array-methods.mjs new file mode 100644 index 000000000..a68ec305e --- /dev/null +++ b/tests/corpus/7984-untyped-array-methods.mjs @@ -0,0 +1,120 @@ +// Untyped calls of push, pop, shift, at, includes and indexOf: plain arrays +// take a direct path; arrays with holes, frozen, sealed or non-extensible +// arrays, Array.prototype, proxies, and arrays whose own property or +// prototype defines the method take the full dispatch. Each call is made +// both for its result and as a statement, whose result is dropped. +function attempt(label, f) { + try { + console.log(label, f()); + } catch (e) { + console.log(label, "threw", e.constructor.name, e.message); + } +} +function push(a, v) { + return a.push(v); +} +function pushOnly(a, v) { + a.push(v); + return a.length; +} +function push2(a, v, w) { + return a.push(v, w); +} +function pop(a) { + return a.pop(); +} +function shift(a) { + return a.shift(); +} +function at(a, i) { + return a.at(i); +} +function includes(a, v) { + return a.includes(v); +} +function indexOf(a, v) { + return a.indexOf(v); +} +function indexFrom(a, v, from) { + return a.indexOf(v, from); +} +function len(a) { + return a.length; +} + +const holes = JSON.parse("[1, 2, 3]"); +delete holes[1]; +attempt("holes push", () => push(holes, 4)); +attempt("holes push statement", () => pushOnly(holes, 5)); +attempt("holes includes", () => [includes(holes, undefined), indexOf(holes, undefined), at(holes, 1)]); +attempt("holes pop", () => [pop(holes), len(holes)]); +attempt("holes shift", () => [shift(holes), len(holes), 1 in holes]); + +const frozen = Object.freeze(JSON.parse("[1, 2]")); +attempt("frozen push", () => push(frozen, 3)); +attempt("frozen push statement", () => pushOnly(frozen, 3)); +attempt("frozen pop", () => pop(frozen)); +attempt("frozen shift", () => shift(frozen)); +attempt("frozen includes", () => [includes(frozen, 2), indexOf(frozen, 2), at(frozen, -1)]); + +const sealed = Object.seal(JSON.parse("[1, 2]")); +attempt("sealed push", () => push(sealed, 3)); +attempt("sealed push statement", () => pushOnly(sealed, 3)); +attempt("sealed pop", () => pop(sealed)); +attempt("sealed shift", () => [shift(sealed), JSON.stringify(sealed)]); +attempt("sealed includes", () => [includes(sealed, 2), at(sealed, 0)]); + +const fixed = Object.preventExtensions(JSON.parse("[1, 2]")); +attempt("fixed push", () => push(fixed, 3)); +attempt("fixed push statement", () => pushOnly(fixed, 3)); +attempt("fixed pop", () => [pop(fixed), len(fixed)]); +attempt("fixed shift", () => [shift(fixed), len(fixed)]); + +attempt("prototype includes", () => [includes(Array.prototype, 1), indexOf(Array.prototype, 1), at(Array.prototype, 0)]); + +const handler = new Proxy({ push(v) { return "proxied push " + v; }, at(i) { return "proxied at " + i; } }, {}); +attempt("proxy push", () => push(handler, 7)); +attempt("proxy push statement", () => pushOnly(handler, 7)); +attempt("proxy at", () => at(handler, 1)); + +const own = JSON.parse("[1, 2]"); +own.push = function (v) { + return "own push " + v; +}; +own.includes = function (v) { + return "own includes " + v; +}; +attempt("own push", () => [push(own, 9), len(own)]); +attempt("own push statement", () => pushOnly(own, 9)); +attempt("own includes", () => [includes(own, 1), indexOf(own, 1)]); +const notCallable = JSON.parse("[1, 2]"); +notCallable.pop = 5; +attempt("own non-callable pop", () => pop(notCallable)); + +const customProto = Object.create(Array.prototype); +customProto.push = function (v) { + return "proto push " + v + " onto " + this.length; +}; +const reproto = JSON.parse("[1, 2]"); +Object.setPrototypeOf(reproto, customProto); +attempt("reproto push", () => push(reproto, 9)); +attempt("reproto push statement", () => pushOnly(reproto, 9)); +attempt("reproto pop", () => [pop(reproto), len(reproto)]); + +const plain = JSON.parse("[]"); +for (let i = 0; i < 5; i++) push2(plain, i, i * 2); +for (let i = 0; i < 3; i++) pushOnly(plain, "s" + i); +attempt("plain", () => [len(plain), pop(plain), shift(plain), at(plain, -2), at(plain, 1.9), at(plain, -0), at(plain, NaN), at(plain, 100), at(plain, -100)]); +attempt("plain search", () => [includes(plain, 8), indexOf(plain, 8), includes(plain, "s1"), indexOf(plain, "s1"), indexOf(plain, "s9"), indexFrom(plain, 2, 3), includes(plain)]); +attempt("plain at string", () => [at(plain, "1"), at(plain, undefined), at(plain, { valueOf() { return 2; } })]); +const nans = JSON.parse("[1]"); +pushOnly(nans, NaN); +pushOnly(nans, -0); +attempt("nan", () => [includes(nans, NaN), indexOf(nans, NaN), includes(nans, 0), indexOf(nans, 0), indexOf(nans, -0)]); +const empty = JSON.parse("[]"); +attempt("empty", () => [pop(empty), shift(empty), at(empty, 0), includes(empty, undefined), indexOf(empty, undefined), len(empty)]); +const objects = JSON.parse("[]"); +const key = { k: 1 }; +pushOnly(objects, key); +pushOnly(objects, { k: 1 }); +attempt("objects", () => [includes(objects, key), indexOf(objects, key), indexOf(objects, { k: 1 }), at(objects, 0) === key]); diff --git a/tests/corpus/8002-while-byte-length-companions.ts b/tests/corpus/8002-while-byte-length-companions.ts new file mode 100644 index 000000000..e37ec9e57 --- /dev/null +++ b/tests/corpus/8002-while-byte-length-companions.ts @@ -0,0 +1,201 @@ +// While loops bounded by a typed array or Buffer length, with companion +// counters that only grow: reversed copies with line breaks, steps of two, +// early exits, skipped passes, nested loops and loops inside loops. + +function reverseLines( + line: Uint8Array, + target: Uint8Array, + start: number, +): number { + const len = line.length; + const right = len - 1; + let read = 0; + let written = 0; + let count = 0; + let lines = 0; + while (read < len) { + let c = line[right - read]!; + read++; + if (c === 10) { + if (read === len) break; + c = line[right - read]!; + read++; + } + target[written] = c + 1; + written++; + count++; + if ((count + start - lines) % 5 === 0) { + target[written] = 10; + written++; + count++; + lines++; + } + } + return written * 1000 + lines * 10 + count; +} + +const text = Buffer.from("ABCDEFGHIJ\nKLMNOPQRST\nUVWXYZ\n"); +const out = new Uint8Array(64); +for (const start of [0, 1, 3, 7]) { + const summary = reverseLines(text, out, start); + console.log( + start, + summary, + Buffer.from(out.subarray(0, Math.floor(summary / 1000))).toString("latin1"), + ); +} + +function pairs(values: Uint16Array): number[] { + let i = 0; + let sums = 0; + let odd = 0; + const result: number[] = []; + while (i < values.length) { + i += 2; + sums++; + if ((values[i - 2]! & 1) === 1) odd += 3; + result.push(values[i - 2]! + (values[i - 1] ?? -1)); + } + result.push(i, sums, odd); + return result; +} +console.log(pairs(new Uint16Array([1, 2, 3, 4, 5]))); +console.log(pairs(new Uint16Array(0))); +console.log(pairs(new Uint16Array([65535, 65535]))); + +function firstZero(bytes: Uint8Array): string { + let i = 0; + let seen = 0; + let total = -5; + while (i < bytes.length) { + i++; + if (bytes[i - 1] === 0) break; + seen++; + total += bytes[i - 1]!; + } + return `${i} ${seen} ${total}`; +} +console.log( + firstZero(new Uint8Array([3, 4, 0, 9])), + firstZero(new Uint8Array([1, 2])), + firstZero(new Uint8Array(0)), +); + +function skipping(bytes: Uint8Array): string { + let i = 0; + let kept = 0; + let skipped = 0; + while (i < bytes.length) { + const b = bytes[i]!; + i++; + if (b > 100) { + skipped++; + continue; + } + kept++; + } + return `${i} ${kept} ${skipped}`; +} +console.log(skipping(new Uint8Array([1, 200, 3, 250, 255, 7]))); + +function repeated(bytes: Uint8Array): string { + let i = 0; + let inner = 0; + let outer = 0; + while (i < bytes.length) { + i++; + outer++; + let j = 0; + while (j < bytes[i - 1]!) { + j++; + inner++; + } + } + return `${i} ${outer} ${inner}`; +} +console.log(repeated(new Uint8Array([2, 0, 5, 1]))); + +function grid(rows: Float64Array[]): string { + let total = 0; + let cells = 0; + for (const row of rows) { + let k = 0; + let doubled = 0; + while (k < row.length) { + k++; + cells++; + doubled += 2; + total += row[k - 1]! * doubled; + } + total += k * 0.5; + } + return `${cells} ${total}`; +} +console.log( + grid([ + new Float64Array([0.5, 1.5]), + new Float64Array(0), + new Float64Array([-2, 3, 4.25]), + ]), +); + +function views(buffer: ArrayBuffer): string { + const view = new DataView(buffer, 2); + let offset = 0; + let words = 0; + const parts: string[] = []; + while (offset < view.byteLength) { + offset += 2; + words++; + if (offset <= view.byteLength) + parts.push(view.getUint16(offset - 2, true).toString(16)); + } + return `${offset} ${words} ${parts.join(",")}`; +} +const backing = new Uint8Array([9, 9, 1, 2, 3, 4, 5]).buffer; +console.log(views(backing)); + +function captured(bytes: Uint8Array): number { + let i = 0; + let n = 0; + const read = (): number => n; + while (i < bytes.length) { + i++; + n++; + } + return read() + i; +} +console.log(captured(new Uint8Array(4))); + +function fromNegative(bytes: Int8Array): string { + let i = 0; + let level = -3; + let low = -0; + while (i < bytes.length) { + i++; + level++; + if (bytes[i - 1]! < 0) low++; + } + return `${level} ${low} ${Object.is(low, -0)}`; +} +console.log( + fromNegative(new Int8Array([-1, 2, -3])), + fromNegative(new Int8Array(0)), +); + +function throwing(bytes: Uint8Array): number { + let i = 0; + let n = 0; + while (i < bytes.length) { + i++; + n++; + if (bytes[i - 1] === 7) throw new Error(`stopped at ${i} after ${n}`); + } + return n; +} +try { + console.log(throwing(new Uint8Array([1, 2]))); + console.log(throwing(new Uint8Array([1, 7, 3]))); +} catch (error) { + console.log((error as Error).message); +} diff --git a/tests/corpus/8105-url-idna-hosts.js b/tests/corpus/8105-url-idna-hosts.js new file mode 100644 index 000000000..329566013 --- /dev/null +++ b/tests/corpus/8105-url-idna-hosts.js @@ -0,0 +1,58 @@ +// WHATWG URL hosts that hold non-ASCII text, directly or after +// percent-decoding, go through UTS #46 ToASCII as the host parser in Node +// does: mapped and normalized labels become Punycode, and disallowed code +// points, invalid joiners and bidi violations make the URL invalid. Cases +// from the WPT url test data, plus setters, pathToFileURL UNC hosts and +// URL.canParse. +import url from "node:url"; + +const parses = [ + ["http://é@é", null], + ["http://GOO\u200B\u2060\uFEFFgoo.com", "http://other.com/"], + ["http://www.foo。bar.com", "http://other.com/"], + ["http://Ｇｏ.com", "http://other.com/"], + ["http://你好你好", "http://other.com/"], + ["https://faß.ExAmPlE/", null], + ["http://０Ｘｃ０．０２５０．０１", "http://other.com/"], + ["ftp://%e2%98%83", null], + ["https://%e2%98%83", null], + ["file://a\u00ADb/p", null], + ["file://a%C2%ADb/p", null], + ["file://loC𝐀𝐋𝐇𝐨𝐬𝐭/usr/bin", null], + ["https://a%C2%ADb/", null], + ["http://münchen.de/straße", null], + ["https://日本語.jp/", null], + ["http://例え.テスト/", null], + ["http://xn--mnchen-3ya.de/", null], + ["http://a\u200Db.com/", null], + ["http://אa.com/", null], + ["http://ab�c/", null], + ["ws://ÉXAMPLE.org:8080/x", null], +]; +function describe(value) { + return [value.href, value.host, value.hostname, value.origin].join(" | "); +} +for (const [input, base] of parses) { + let line; + try { + line = describe(base === null ? new URL(input) : new URL(input, base)); + } catch (e) { + line = `throws ${e.code}`; + } + console.log(JSON.stringify(input), "=>", line, URL.canParse(input, base ?? undefined)); +} +const writes = [ + ["host", "https://x/", "ß"], + ["host", "https://example.com/", "a%C2%ADb"], + ["hostname", "https://example.com/", "a%C2%ADb"], + ["hostname", "https://example.com/", "BÜCHER.example"], +]; +for (const [name, href, value] of writes) { + const target = new URL(href); + if (name === "host") target.host = value; + else target.hostname = value; + console.log(name, JSON.stringify(value), "=>", describe(target)); +} +for (const path of ["\\\\münchen\\share\\f", "\\\\BÜCHER\\s\\f"]) { + console.log(JSON.stringify(path), url.pathToFileURL(path, { windows: true }).href); +} diff --git a/tests/corpus/8106-shared-typed-array-set.ts b/tests/corpus/8106-shared-typed-array-set.ts new file mode 100644 index 000000000..7e20aad72 --- /dev/null +++ b/tests/corpus/8106-shared-typed-array-set.ts @@ -0,0 +1,97 @@ +// TypedArray.prototype.set where the source, the target or both are views of +// a SharedArrayBuffer: same element kinds (bytes copied as is, NaN payloads +// included), different kinds (converted per element), offsets, overlapping +// ranges of one buffer, Buffer targets and out-of-bounds offsets. +function bits(label: string, buffer: ArrayBufferLike): void { + const words = new Uint32Array(buffer); + const out: string[] = []; + for (let i = 0; i < words.length; i++) out.push(words[i]!.toString(16)); + console.log(label, out.join(",")); +} + +const shared = new SharedArrayBuffer(32); +const bytes = new Uint8Array(shared); + +// Private source into a shared target, then shared into shared and shared into private. +bytes.set(new Uint8Array([1, 2, 3, 4, 5]), 2); +console.log("private into shared", bytes.subarray(0, 8).join(",")); +const other = new Uint8Array(new SharedArrayBuffer(8)); +other.set(bytes.subarray(2, 7)); +console.log("shared into shared", other.join(",")); +const local = new Uint8Array(12); +local.set(other, 3); +console.log("shared into private", local.join(",")); + +// Overlapping ranges of the same buffer, forward and backward. +const overlap = new Uint8Array(new SharedArrayBuffer(10)); +overlap.set([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]); +overlap.set(overlap.subarray(0, 6), 3); +console.log("overlap forward", overlap.join(",")); +overlap.set(overlap.subarray(4, 10), 1); +console.log("overlap backward", overlap.join(",")); + +// Same kind, wider elements: a signaling NaN keeps its bits. +const floatBits = new Uint32Array(new SharedArrayBuffer(8)); +floatBits[0] = 0x7fa00001; +floatBits[1] = 0xffc00002; +const floatsShared = new Float32Array(floatBits.buffer); +const floatsLocal = new Float32Array(2); +floatsLocal.set(floatsShared); +bits("f32 shared into private", floatsLocal.buffer); +const floatsBack = new Float32Array(new SharedArrayBuffer(8)); +floatsBack.set(floatsLocal); +bits("f32 private into shared", floatsBack.buffer); + +const doubleBits = new Uint32Array(new SharedArrayBuffer(16)); +doubleBits[0] = 5; +doubleBits[1] = 0x7ff40000; +doubleBits[2] = 0; +doubleBits[3] = 0xc0090000; +const doublesShared = new Float64Array(doubleBits.buffer); +const doublesCopy = new Float64Array(new SharedArrayBuffer(16)); +doublesCopy.set(doublesShared); +bits("f64 shared into shared", doublesCopy.buffer); +console.log("f64 values", doublesCopy[1]); + +const signed = new Int16Array(new SharedArrayBuffer(8)); +signed.set(new Int16Array([-32768, -1, 32767]), 1); +console.log("i16 private into shared", signed.join(",")); + +// Different kinds convert each element. +const clamped = new Uint8ClampedArray(new SharedArrayBuffer(4)); +clamped.set(new Int16Array([300, -5, 128, 255])); +console.log("i16 into shared clamped", clamped.join(",")); +const wrapped = new Uint8Array(new SharedArrayBuffer(4)); +wrapped.set(new Int16Array(signed.buffer, 2, 3)); +console.log("shared i16 into shared u8", wrapped.join(",")); +const narrowed = new Int8Array(4); +narrowed.set(new Float64Array(doublesShared.buffer, 8, 1)); +narrowed.set(new Uint8Array([200, 127]), 2); +console.log("f64 and u8 into i8", narrowed.join(",")); +const fromClamped = new Uint8Array(new SharedArrayBuffer(4)); +fromClamped.set(clamped); +console.log("shared clamped into shared u8", fromClamped.join(",")); + +// Buffer views over shared memory take the same paths. +const buf = Buffer.from(shared, 16, 8); +buf.set(Buffer.from("scriptc"), 1); +console.log("buffer", buf.toString("latin1", 1, 8), bytes[17], bytes[23]); +const into = Buffer.alloc(4); +into.set(buf.subarray(1, 4)); +console.log("buffer into private", into.toString("hex")); + +// Offsets past the end throw before writing. +for (const offset of [-1, 30, 31]) { + try { + bytes.set(new Uint8Array([9, 9]), offset); + console.log("set at", offset, bytes[30], bytes[31]); + } catch (error) { + console.log("set at", offset, (error as Error).name, (error as Error).message); + } +} +try { + floatsBack.set(new Float32Array(3)); +} catch (error) { + console.log("too long", (error as Error).name, (error as Error).message); +} +console.log("unchanged", floatsBack.join(",")); diff --git a/tests/corpus/8107-shared-typed-array-set-workers.ts b/tests/corpus/8107-shared-typed-array-set-workers.ts new file mode 100644 index 000000000..6be32668b --- /dev/null +++ b/tests/corpus/8107-shared-typed-array-set-workers.ts @@ -0,0 +1,63 @@ +// TypedArray.prototype.set on SharedArrayBuffer views in a program with +// workers: the same cases on the main thread and in a worker (same element +// kinds with NaN payloads, overlapping ranges of one buffer, converted +// kinds), and a worker's set into a buffer that the main thread reads after +// a message. +import { isMainThread, parentPort, Worker, workerData } from "node:worker_threads"; + +interface Job { + shared: SharedArrayBuffer; +} + +function hex(buffer: ArrayBufferLike): string { + const words = new Uint32Array(buffer); + const out: string[] = []; + for (let i = 0; i < words.length; i++) out.push(words[i]!.toString(16)); + return out.join(","); +} + +function cases(where: string): string[] { + const lines: string[] = []; + const bytes = new Uint8Array(new SharedArrayBuffer(12)); + bytes.set([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]); + bytes.set(bytes.subarray(0, 8), 4); + lines.push(`${where} overlap forward ${bytes.join(",")}`); + bytes.set(bytes.subarray(5, 12), 0); + lines.push(`${where} overlap backward ${bytes.join(",")}`); + + const floatBits = new Uint32Array(new SharedArrayBuffer(8)); + floatBits[0] = 0x7f800001; + floatBits[1] = 0x7fc12345; + const local = new Float32Array(2); + local.set(new Float32Array(floatBits.buffer)); + const back = new Float32Array(new SharedArrayBuffer(12)); + back.set(local, 1); + lines.push(`${where} f32 ${hex(local.buffer)} ${hex(back.buffer)}`); + + const halves = new Uint16Array(new SharedArrayBuffer(8)); + halves.set(new Uint16Array([65535, 1, 2]), 1); + const signed = new Int16Array(4); + signed.set(halves); + const clamped = new Uint8ClampedArray(new SharedArrayBuffer(4)); + clamped.set(signed); + lines.push(`${where} kinds ${halves.join(",")} ${signed.join(",")} ${clamped.join(",")}`); + return lines; +} + +if (isMainThread) { + for (const line of cases("main")) console.log(line); + const shared = new SharedArrayBuffer(16); + const job: Job = { shared }; + const worker = new Worker(new URL(import.meta.url), { workerData: job }); + worker.on("message", (message: string) => { + console.log(message); + console.log("worker wrote", new Uint8Array(shared).join(",")); + void worker.terminate(); + }); +} else { + const { shared } = workerData as Job; + const target = new Uint8Array(shared); + target.set(new Uint8Array([9, 8, 7]), 2); + target.set(target.subarray(2, 5), 10); + parentPort!.postMessage(cases("worker").join("\n")); +} diff --git a/tests/corpus/8215-js-any-logical-module-bindings/main.ts b/tests/corpus/8215-js-any-logical-module-bindings/main.ts new file mode 100644 index 000000000..3a1d5eb17 --- /dev/null +++ b/tests/corpus/8215-js-any-logical-module-bindings/main.ts @@ -0,0 +1,8 @@ +// Module bindings of a JavaScript package whose initializers combine an +// untyped value with a string, number or boolean through `||`, `&&`, `??` +// or `?:` (pi-coding-agent's config.js): other modules and functions read +// the values as in Node. +import { banner, usage } from "app-config"; + +console.log(banner); +console.log(usage("check")); diff --git a/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/config.js b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/config.js new file mode 100644 index 000000000..faecb8b2b --- /dev/null +++ b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/config.js @@ -0,0 +1,18 @@ +import { readFileSync } from "node:fs"; +import { join } from "node:path"; + +let pkg = {}; +try { + pkg = JSON.parse(readFileSync(join(import.meta.dirname, "..", "package.json"), "utf-8")); +} +catch (e) { + if (e.code !== "ENOENT") + throw e; +} +const configName = pkg.appConfig?.name; +export const APP_NAME = configName || "app"; +export const APP_TITLE = configName ? APP_NAME : "App"; +export const CONFIG_DIR = pkg.appConfig?.configDir || ".config"; +export const VERSION = pkg.version ?? "0.0.0"; +export const RETRIES = pkg.appConfig?.retries ?? 3; +export const isBundled = typeof APP_BUNDLED !== "undefined" && APP_BUNDLED; diff --git a/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/index.d.ts b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/index.d.ts new file mode 100644 index 000000000..4c4706926 --- /dev/null +++ b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/index.d.ts @@ -0,0 +1,2 @@ +export declare const banner: string; +export declare function usage(kind: "check" | "list"): string; diff --git a/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/index.js b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/index.js new file mode 100644 index 000000000..251303802 --- /dev/null +++ b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/index.js @@ -0,0 +1 @@ +export { banner, usage } from "./usage.js"; diff --git a/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/usage.js b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/usage.js new file mode 100644 index 000000000..736cef3c0 --- /dev/null +++ b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/dist/usage.js @@ -0,0 +1,12 @@ +import { APP_NAME, APP_TITLE, CONFIG_DIR, RETRIES, VERSION, isBundled } from "./config.js"; + +const USAGE = { + check: `${APP_NAME} check --dir ${CONFIG_DIR}`, + list: `${APP_NAME} list`, +}; + +export const banner = `${APP_TITLE} ${VERSION} (bundled: ${isBundled}, retries: ${RETRIES})`; + +export function usage(kind) { + return `${USAGE[kind]} [${APP_NAME.toUpperCase()}_HOME=${CONFIG_DIR}] ${typeof isBundled} ${RETRIES + 1}`; +} diff --git a/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/package.json b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/package.json new file mode 100644 index 000000000..918204988 --- /dev/null +++ b/tests/corpus/8215-js-any-logical-module-bindings/node_modules/app-config/package.json @@ -0,0 +1 @@ +{ "name": "app-config", "version": "2.1.0", "type": "module", "main": "dist/index.js", "types": "dist/index.d.ts", "appConfig": { "configDir": ".app" } } diff --git a/tests/corpus/instanceof-primitive-wrappers-typed.ts b/tests/corpus/instanceof-primitive-wrappers-typed.ts new file mode 100644 index 000000000..dc065e569 --- /dev/null +++ b/tests/corpus/instanceof-primitive-wrappers-typed.ts @@ -0,0 +1,14 @@ +// Typed operands tested against the primitive wrapper constructors: always +// false, with the operand's side effects kept. +let reads = 0; +function pick(n: number): unknown { + reads++; + return n === 0 ? "text" : n === 1 ? 7 : n === 2 ? { ok: true } : [n]; +} +const seen: string[] = []; +for (let i = 0; i < 4; i++) { + const value = pick(i); + seen.push(`${value instanceof String}/${value instanceof Number}/${value instanceof Boolean}`); +} +const record: object = { a: 1 }; +console.log(seen.join(" "), record instanceof Object, record instanceof String, reads); diff --git a/tests/corpus/instanceof-primitive-wrappers.mjs b/tests/corpus/instanceof-primitive-wrappers.mjs new file mode 100644 index 000000000..11bcaba07 --- /dev/null +++ b/tests/corpus/instanceof-primitive-wrappers.mjs @@ -0,0 +1,24 @@ +// `instanceof` against the primitive wrapper constructors (Boolean, Number, +// String, BigInt, Symbol), as typebox's value guards test: false for +// primitives and for every object, and the operand still evaluates. +function IsBoolean(value) { + return value instanceof Boolean; +} +function IsNumber(value) { + return value instanceof Number; +} +function IsString(value) { + return value instanceof String; +} +function IsUint8Array(value) { + return value instanceof globalThis.Uint8Array; +} +const values = [true, 1, "s", 2n, Symbol("k"), null, undefined, new Uint8Array(2), {}, [], () => 1, new Date(0)]; +for (const v of values) + console.log(typeof v, IsBoolean(v), IsNumber(v), IsString(v), IsUint8Array(v), v instanceof BigInt, v instanceof Symbol); +let calls = 0; +const next = () => { + calls++; + return { n: calls }; +}; +console.log(next() instanceof Number, next() instanceof globalThis.String, calls); diff --git a/tests/harness/ci-test-plan.test.ts b/tests/harness/ci-test-plan.test.ts index dd77039ca..29268c841 100644 --- a/tests/harness/ci-test-plan.test.ts +++ b/tests/harness/ci-test-plan.test.ts @@ -298,8 +298,10 @@ test("pull requests run the full suite on Linux and the darwin suite on macOS", // runner, and directly elsewhere. const linuxRun = linux.steps.at(-1)!.run as string; expect(linuxRun.startsWith("ulimit -c 0\n")).toBe(true); - expect(linuxRun).toContain("systemd-run --scope --quiet -p MemoryMax=14G -p OOMPolicy=continue"); - expect(linuxRun).toMatch(/-- env PATH="\$PATH" HOME="\$HOME" node scripts\/ci-test\.mjs\n/); + expect(linuxRun).toContain("systemd-run --scope --quiet -p MemoryMax=15G -p OOMPolicy=continue"); + expect(linuxRun).toMatch( + /-- env PATH="\$PATH" HOME="\$HOME" NODE_OPTIONS="--max-old-space-size=4096" node scripts\/ci-test\.mjs\n/, + ); expect(linuxRun.endsWith("fi\nnode scripts/ci-test.mjs\n")).toBe(true); expect(macos.steps.at(-1)!.run).toBe("node scripts/ci-test.mjs"); for (const job of [linux, macos]) expect(job.strategy.matrix.flavor).toEqual(["plain", "san"]); diff --git a/tests/harness/corpus-costs/6305-json-discriminant-dispatch.ts.json b/tests/harness/corpus-costs/6305-json-discriminant-dispatch.ts.json new file mode 100644 index 000000000..7e2a9d53c --- /dev/null +++ b/tests/harness/corpus-costs/6305-json-discriminant-dispatch.ts.json @@ -0,0 +1 @@ +{ "plain": 5844, "san": 7609 } diff --git a/tests/harness/corpus-costs/6306-discriminant-getter-dispatch.ts.json b/tests/harness/corpus-costs/6306-discriminant-getter-dispatch.ts.json new file mode 100644 index 000000000..2c49e2f02 --- /dev/null +++ b/tests/harness/corpus-costs/6306-discriminant-getter-dispatch.ts.json @@ -0,0 +1 @@ +{ "plain": 3800, "san": 4900 } diff --git a/tests/harness/corpus-costs/7984-untyped-array-methods.mjs.json b/tests/harness/corpus-costs/7984-untyped-array-methods.mjs.json new file mode 100644 index 000000000..5ae927c90 --- /dev/null +++ b/tests/harness/corpus-costs/7984-untyped-array-methods.mjs.json @@ -0,0 +1 @@ +{ "plain": 4100, "san": 4500 } diff --git a/tests/harness/corpus-costs/8002-while-byte-length-companions.ts.json b/tests/harness/corpus-costs/8002-while-byte-length-companions.ts.json new file mode 100644 index 000000000..d22ff8534 --- /dev/null +++ b/tests/harness/corpus-costs/8002-while-byte-length-companions.ts.json @@ -0,0 +1 @@ +{ "plain": 3948, "san": 3569 } diff --git a/tests/harness/corpus-costs/8215-js-any-logical-module-bindings/main.ts.json b/tests/harness/corpus-costs/8215-js-any-logical-module-bindings/main.ts.json new file mode 100644 index 000000000..d5d1c5f65 --- /dev/null +++ b/tests/harness/corpus-costs/8215-js-any-logical-module-bindings/main.ts.json @@ -0,0 +1 @@ +{ "plain": 2500, "san": 3500 } diff --git a/tests/harness/corpus-costs/instanceof-primitive-wrappers-typed.ts.json b/tests/harness/corpus-costs/instanceof-primitive-wrappers-typed.ts.json new file mode 100644 index 000000000..d06ba3ac5 --- /dev/null +++ b/tests/harness/corpus-costs/instanceof-primitive-wrappers-typed.ts.json @@ -0,0 +1 @@ +{ "plain": 2217 } diff --git a/tests/harness/corpus-costs/instanceof-primitive-wrappers.mjs.json b/tests/harness/corpus-costs/instanceof-primitive-wrappers.mjs.json new file mode 100644 index 000000000..68ee374b9 --- /dev/null +++ b/tests/harness/corpus-costs/instanceof-primitive-wrappers.mjs.json @@ -0,0 +1 @@ +{ "plain": 2265 } diff --git a/tests/harness/dyncheck.test.ts b/tests/harness/dyncheck.test.ts index 15068776a..69e97dc6d 100644 --- a/tests/harness/dyncheck.test.ts +++ b/tests/harness/dyncheck.test.ts @@ -396,6 +396,95 @@ console.log("unreachable", s.kind); expect(r.stderr).toContain("Uncaught TypeError: expected object | object at $, got object"); }); + test("a union read by its discriminant fails with every arm named, as arm-by-arm tests do", async () => { + // Five record arms with six literals, plus string and number arms: the + // check reads `kind` once and tests only the arm it names and the + // arms that are not records. + const r = await compileAndRun( + "union-discriminant-dispatch", + `type Expr = + | { kind: "num"; value: number } + | { kind: "str"; text: string } + | { kind: "neg"; operand: Expr } + | { kind: "add" | "mul"; left: Expr; right: Expr } + | { items: Expr[]; kind: "list" } + | string + | number; +const inputs = [ + '{"kind":"num"}', + '{"kind":"nope","value":1}', + '{"kind":7,"value":1}', + '{"value":1}', + '{"kind":"neg","operand":{"kind":"add","left":{"kind":"num","value":1},"right":{"kind":"str","text":5}}}', + 'true', + '{"items":["a",3,{"kind":"mul","left":"x","right":2}],"kind":"list"}', +]; +for (const input of inputs) { + try { + console.log(JSON.stringify(JSON.parse(input) as Expr)); + } catch (error) { + console.log((error as Error).message); + } +} +`, + ); + const any = "number | object | object | object | object | object | string"; + expect(r.exitCode).toBe(0); + expect(r.stdout).toBe( + [ + `expected ${any} at $, got object`, + `expected ${any} at $, got object`, + `expected ${any} at $, got object`, + `expected ${any} at $, got object`, + `expected ${any} at $, got object`, + `expected ${any} at $, got boolean`, + '{"items":["a",3,{"kind":"mul","left":"x","right":2}],"kind":"list"}', + "", + ].join("\n"), + ); + }); + + test("a getter that changes the discriminant during an arm's check continues with the later arms", async () => { + // Each getter sets `kind` and fails its arm's check: to a later arm's + // literal (that arm matches), to an earlier arm's literal (arm-by-arm + // tests do not go back), and to a literal no arm holds. + const r = await compileAndRun( + "union-discriminant-getter", + `type Shape = + | { kind: "circle"; radius: number } + | { kind: "square"; side: number } + | { kind: "rect"; height: number; width: number } + | { kind: "tri"; a: number; b: number; c: number }; +const flip = (value: any, to: string): string => { + value.kind = to; + return "none"; +}; +const later: any = { kind: "circle", get radius() { return flip(later, "square"); }, side: 3 }; +const earlier: any = { kind: "square", radius: 2, get side() { return flip(earlier, "circle"); } }; +const unknown: any = { kind: "square", get side() { return flip(unknown, "hexagon"); } }; +const inputs = [later, earlier, unknown]; +for (const input of inputs) { + try { + const shape = input as Shape; + console.log("matched", shape.kind); + } catch (error) { + console.log((error as Error).message); + } +} +`, + ); + const any = "object | object | object | object"; + expect(r.exitCode).toBe(0); + expect(r.stdout).toBe( + [ + "matched square", + `expected ${any} at $, got object`, + `expected ${any} at $, got object`, + "", + ].join("\n"), + ); + }); + test("nested paths point at the offending element", async () => { const r = await compileAndRun( "nested-path", diff --git a/tests/harness/event-loop-cases.ts b/tests/harness/event-loop-cases.ts index 50a7d2b85..e0ecfc899 100644 --- a/tests/harness/event-loop-cases.ts +++ b/tests/harness/event-loop-cases.ts @@ -37,9 +37,11 @@ export const eventLoopCases: EventLoopCase[] = [ { title: "stdio values: 'readable' drains with read(), then fires once more at EOF", fixture: "stdio-readable.mjs", + // The first chunk waits until both readers run: a reader that starts + // after the second chunk arrives reads the two in one 'readable'. script: { writes: [ - { delayMs: 0, data: "alpha " }, + { delayMs: 500, data: "alpha " }, { delayMs: 60, data: "beta\n" }, ], end: true, diff --git a/tests/harness/library-mode.test.ts b/tests/harness/library-mode.test.ts index 7eecd3c93..784dc747b 100644 --- a/tests/harness/library-mode.test.ts +++ b/tests/harness/library-mode.test.ts @@ -363,6 +363,17 @@ globals: undefined fenced expect(run.stdout).toBe(SESSION + SESSION + SESSION); }); + test("K4: untyped property caches start empty in every session", async () => { + // The shapes a session's cache entries name are freed when it ends; + // the next session's sites must not find those entries. + const { archive, outDir } = await buildLibrary("js-ic-sessions", emission); + const probe = buildProbe("js-ic-sessions", archive, outDir); + const run = runProbe(probe); + expect(run.signal).toBeNull(); + expect(run.status).toBe(0); + expect(run.stdout).toBe("run: 20000\n".repeat(3)); + }); + /* ── K5/K6/K7: the trap channel ──────────────────────────────────────── */ test("K5: a trap delivers to the sink exactly once, host survives", async () => { @@ -582,6 +593,15 @@ survived, sink_calls=1 expect(run.stdout).toBe(SESSION + SESSION + SESSION); }); + platformTest("K10: property caches across sessions under ASan", async () => { + const { archive, outDir } = await buildLibrary("js-ic-sessions", emission, { sanitize: true }); + const probe = buildProbe("js-ic-sessions", archive, outDir, { sanitize: true }); + const run = runProbe(probe); + expect(run.signal).toBeNull(); + expect(run.status).toBe(0); + expect(run.stdout).toBe("run: 20000\n".repeat(3)); + }); + platformTest("K10: K5/K7 under ASan", async () => { const { archive, outDir } = await buildLibrary("traps", emission, { sanitize: true }); const probe = buildProbe("traps", archive, outDir, { sanitize: true }); diff --git a/tests/library-mode/js-ic-sessions/lib.ts b/tests/library-mode/js-ic-sessions/lib.ts new file mode 100644 index 000000000..ebea59a54 --- /dev/null +++ b/tests/library-mode/js-ic-sessions/lib.ts @@ -0,0 +1,6 @@ +// The library's typed export over untyped code (work.mjs). +import { runAll } from "./work.mjs"; + +export function run(): number { + return runAll() as number; +} diff --git a/tests/library-mode/js-ic-sessions/probe.c b/tests/library-mode/js-ic-sessions/probe.c new file mode 100644 index 000000000..7dbcb56bf --- /dev/null +++ b/tests/library-mode/js-ic-sessions/probe.c @@ -0,0 +1,25 @@ +/* Three sessions of a library whose untyped code reads and writes through + * property caches: each prints the same result, and under the sanitize + * flavor no session reads a cache entry the previous session freed. */ +#include +#include + +extern void kj_init(void); +extern void kj_set_panic_sink(void (*fn)(void *, const uint8_t *, size_t, uint64_t), void *ctx); +extern void kj_collect(void); +extern double kj_run(void); + +static void sink(void *ctx, const uint8_t *msg, size_t len, uint64_t addr) { + (void)ctx; (void)addr; + printf("UNEXPECTED SINK: %.*s", (int)len, (const char *)msg); +} + +int main(void) { + kj_set_panic_sink(sink, NULL); + for (int session = 0; session < 3; session++) { + kj_init(); + printf("run: %.0f\n", kj_run()); + kj_collect(); + } + return 0; +} diff --git a/tests/library-mode/js-ic-sessions/profile.json b/tests/library-mode/js-ic-sessions/profile.json new file mode 100644 index 000000000..fd802b42c --- /dev/null +++ b/tests/library-mode/js-ic-sessions/profile.json @@ -0,0 +1,14 @@ +{ + "profile_format": 1, + "name": "conformance-js-ic-sessions", + "entry": "lib.ts", + "emission": "llvm", + "abi": { + "prefix": "kj_", + "init_symbol": "kj_init", + "sink_register_symbol": "kj_set_panic_sink", + "collect_symbol": "kj_collect", + "result_reset_symbol": null + }, + "exports": [{ "export": "run", "symbol": "kj_run", "params": [], "returns": "f64" }] +} diff --git a/tests/library-mode/js-ic-sessions/work.mjs b/tests/library-mode/js-ic-sessions/work.mjs new file mode 100644 index 000000000..5c1afcd18 --- /dev/null +++ b/tests/library-mode/js-ic-sessions/work.mjs @@ -0,0 +1,26 @@ +// Dynamic-tier property caches across library sessions: objects built empty +// and filled through untyped stores, read back through untyped loads, and a +// class instance at the same kind of sites. Each session must start with +// empty caches, since the shapes their entries named are freed when the +// previous session ends. +class Box { + constructor(v) { + this.v = v; + } +} +function fill(o, n) { + o.a = n; + o.b = n + 1; + return o; +} +function sum(o) { + return o.a + o.b; +} +function unbox(b) { + return b.v; +} +export function runAll() { + let total = 0; + for (let i = 0; i < 100; i++) total += sum(fill({}, i)); + return total + unbox(new Box(total)); +}