A record flowing into a union where several record arms could hold it by a
width conversion was refused with SC2003 as ambiguous, even when one arm
keeps every field of the record and the others hold it only by dropping
fields (a model record whose option union is narrower than the catalog's,
beside a classifier arm with a subset of its fields). That arm is the
value's own layout, widened: it now wins when it is the only one. Plans
that already resolved (a single candidate, the absent-field preference,
discriminant and presence routes) are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A promise or function value flowing into a union whose only arm of that
kind has a wider payload or result (`Promise<T>` into `Promise<unknown> |
undefined`, `(s) => Promise<string | null>` into an optional
`(s) => Promise<string | null | undefined>` member) was refused with SC2003:
an arm value wrapped into a union only when the union held an identical
arm. It now takes the arm-level conversion (the checked payload widening,
the function adapter) and wraps the result.
`p === u` against such a union compares both sides as untyped values, so a
promise keeps its identity although converting it into the union's arm
builds a new promise.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A class's constructor and methods are analyzed in receiver contexts
rooted at each member. Inside a per-site clone of the function that
declares the class, they resolved that function's bindings in the
default context: their writes through captured variables were lost,
and reads of the written objects folded to stale literals. A function
with a class declaration or expression anywhere inside is no longer
cloned per site.
The static-tables case adds a factory whose class method writes the
factory's object at initialization. It prints Node's `1,1,1` (it
printed `1,1,0`), and the count of constant reads stays at 8. zod3's
report is unchanged: 180 constant objects and 360 foldable reads.
An intersection that restates an optional member as required with a
separately written element type (`Opts & { files: Array<{ path: string }> }`
over `files?: Array<{ path: string }>`) gives the member the type
`A[] & B[]`. The checker keeps both element literals, and no intersection
rule mapped an intersection of arrays, so the record was refused (SC2009).
When every constituent maps to one IR type that is not the dynamic value,
the intersection now maps to it. The rule runs only after every other
intersection rule declined, so intersections that already mapped keep their
mapping.
pi's NormalizedBuildSystemPromptOptions is such an intersection; it sits on
the chain from main()'s options to the extension API and refused main().
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A cast to a generic type alias spelled over the enclosing function's own
type parameters (`return x as Catalog<G, P>`) stays symbolic in the
checker inside the body, so a mapped type with key remapping, a template
key filter or a conditional member refused with SC2001 in every instance.
The checker interns alias instantiations, so the body's `Catalog<G, P>` is
the same type the signature declares. Binding unification now records, per
call-keyed instance, each declared generic alias reference together with
the alias the call's resolved signature instantiates. When the symbolic
form does not map, mapType maps that concrete instantiation instead. A body
forwarding its own alias to another generic resolves through the enclosing
instance. Types that already mapped are unaffected.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The overlay validation suites (npm, npm-static, three, Effect 3 and 4) now
pass with the overlay on. Each cause had an npm repro or suite case that
failed with the overlay and passed without it:
Analysis facts:
- Symbol members: Symbol.for/keyFor are functions, well-known symbols are
symbols, other members unknown.
- Reads that may answer undefined: index reads past the end (unless the
enclosing for loop proves them in bounds of an array that never shrinks or
has holes), computed reads of objects, keys an object literal lacks at
creation, optional-chain short circuits, `let x;` before assignment.
- Object.create(proto) inherits proto members; arguments past the declared
parameters are unknown; `new` of a primitive answers nothing.
- Arithmetic on bigints answers bigints (number|bigint is never printed).
Overlay:
- Package JSDoc is a declaration: it stays as written and the overlay adds
only types the JSDoc does not declare.
- Objects held by module-level bindings (and what they hold) are never typed
as records: a typed return would copy them.
- Only checks of overlay facts carry the inferred-fact message.
- static-objects ignores nullish receivers outside optional chains; folded
literal keys are unparenthesized.
Fallback (the overlay never adds a failure the default build lacks):
- the overlay state is kept per load (a later load reset it, so the
fallback never ran);
- deferred refusals in unreached code count;
- a preflight, deferred refusals or IR the program without the overlay does
not have make the build compile without the overlay.
CI: the shape_overlay job builds its own native packages; a manual run with
scope shape_overlay runs it alone.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Three JavaScript lowering bugs threw a checked conversion where Node answers:
- concat with a checked-dynamic argument lowered it as an element, so an
array argument threw. The argument now spreads when it is an array at run
time and is appended otherwise.
- `let c; while ((c = str[i++]) !== "]" && c)`: the checker types c from the
index read, without the undefined a read past the end answers. Evolving-let
inference ignores index-read writes, and reads of a JS `let x;` statement
binding keep the actual value.
- Strict equality on a flow-narrowed dynamic JS name compares unchecked.
npm cases concat-dynamic-args and evolving-let-index match Node.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- A member that String.prototype, Number.prototype, Boolean.prototype and
Object.prototype lack reads undefined only while no write may add it.
Assignments to one of those prototypes, Object.defineProperty,
Object.assign, setPrototypeOf and delete on one, writes to an unknown
value, and passing a prototype to unknown code add the name (or every
name), and the reads already made take an unknown value as well. String
indices past the end follow the same rule.
- RegExp.prototype.exec results hold undefined as well as strings: a
capture group that did not take part in the match is undefined, as in
String.prototype.match results.
Two cases where a per-site clone lost writes, so later reads of the
written objects folded to their initial literals:
- A function walked in an enclosing clone's or receiver's context is
no longer cloned per site. The clone was its own context root, so the
variables it captures resolved to the default context's bindings, and
its writes through them reached no object.
- A clone's receiver now receives what the function's own receiver
does, such as the object `bind` supplies, and an arrow's clone keeps
the arrow's lexical `this`. Before, a bound function called at module
initialization wrote through a fresh, empty receiver.
The static-tables case writes at initialization through a helper's
captured binding and through a bound function. It prints Node's `1,1`
(it printed `0,0`), and the count of constant reads stays at 8. zod3's
report keeps its 180 constant objects and 360 foldable reads.
renderShapeOverlay rebuilt the whole source text once per JSDoc range, so
a file with many annotations cost time quadratic in its size. The ranges are
leading trivia of declarations, equal or disjoint, so the text is now built
once from the ranges in order.
`length` and `byteLength` loaded the typed array's header with
`!invariant.load` (#2016). LLVM may move such a load down to its use,
past the release that frees the view, so `const slots = this.slots;
this.slots = next; return slots.length` returned a garbage number in
release builds. These reads are now ordinary loads that keep the
header's alias tag. Header loads that only bound or address an element
access stay invariant: their values reach only that access, which runs
while the view is alive, so typed-array loops still hoist the length
and data pointer.
Corpus typed-array-length-after-release covers a replaced field, a
deleted map entry and a rebound global.
`f!()`, `t[k]!()` and `o.cb!()` on a value that holds undefined threw
scriptc's checked-extraction message ("undefined is not representable in
the target union …"). The non-null assertion is erased in JavaScript, so
Node throws the call's own "<callee> is not a function". The checked
extraction behind an asserted callee now carries that message, with
V8's spelling of the callee: parentheses and non-null assertions dropped
and call results printed as `f(...)`.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Static-object specialization's stage 3 prerequisite
(DESIGN-constant-objects.md): a function called from module
initialization (top-level code outside loops, or another such clone) is
analyzed once per call site, up to 64 sites per function and 4 nested
clones deep. The objects and arrays it creates are one allocation per
call site, and each clone runs at most once, so they are singletons: the
schemas zod's and valibot's builders create at init become constant
objects instead of "init-factory" near misses.
Stage 1 report (SCRIPTC_DEBUG_SHAPES=const):
- zod 3.25 (z.object bench): constant objects 7 -> 48 (fields 57 -> 155),
init-factory near misses 12 -> 1, foldable reads 22 -> 72.
- valibot 1.1: constant objects 2 -> 5, init-factory 6 -> 0.
The overlay's annotations and the zod and namespace benchmarks are
unchanged. The overlay counts union arms rather than atoms, so the
per-site allocations of one record shape stay one typedef.
SCRIPTC_SHAPE_SITE_CLONES sets the per-function limit (0: off).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Compile cost (Hono fixture: 53 s -> 99 s with the overlay before):
- The overlay is planned once per program text: a reload that only drops
refused annotations renders the cached plan (overlay.ts plan/render
split) and reuses the cached structural rewrites, instead of analyzing,
reporting and planning again.
- The overlay's analyses run without per-call-site clones, which only the
constant-graph report uses so far (SCRIPTC_SHAPE_SITE_CLONES opts in).
- A failing round also drops what the deferred refusals blame, saving a
reload; deferred refusals no longer blame every annotation of their
enclosing function (most of them exist without the overlay).
- A parameter whose `typeof` tests the proven values never pass stays
untyped: the typed parameter would narrow to never in that branch,
which the lowering refuses (zod 4's Doc.write).
- Deferred refusals in methods blame the arguments of the method's call
sites (implicit instantiations), like those in functions.
Package JSDoc: the overlay neutralized the package's JSDoc in every
overlaid file, so declarations it does not type lost their documented
types and the lowering instantiated them per call site instead. In
commander, this changed which refusal `--help` reaches: the run passed
the refusal the default build stops at and failed later, in
Command.helpInformation, at a checked conversion of the package's own
documented contract (a help width that is undefined off a TTY). Nothing
the analysis proved was contradicted (SCRIPTC_SHAPE_CHECK_LOG logs no
inferred-fact failure). Only the leading JSDoc of the declarations the
overlay annotates is neutralized now; commander's overlay build fails
exactly where its default build does, and the OVERLAY_KNOWN entries go.
The overlay npm suite (134 cases), the shape tests, lint and format pass.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The analysis exceeded its 20M-step budget on the Hono fixture (32.5 s)
and on moment 2.30 (60 s, two files), so it reported no facts for them.
- Widening: a value node holding more than 32 atoms (merged code that
every value flows through, such as moment's utilities) becomes ⊤. Its
values escape, so the facts stay sound, and later atoms escape instead
of multiplying constraints. SCRIPTC_SHAPE_WIDEN sets the limit.
- A member a string, number or boolean prototype lacks reads undefined
instead of creating a builtin atom per name (Hono's request and route
objects meet strings in shared helpers).
- Members of builtin instances (`RegExp.instance`) no longer chain into
static builtin paths.
- Symbol resolution is memoized per identifier.
Hono: 1.16M steps in 2.4 s with complete facts (55 constant objects).
moment: 115k steps in 0.4 s. zod 3 and the namespace benchmark are
unchanged (same steps and overlay annotations).
SCRIPTC_DEBUG_SHAPES_PROFILE prints where the solver's steps go.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Static-object specialization for namespace-style JavaScript, with the
shape overlay on (shape-inference/static-objects.ts):
- Functions stored in object literals that a top-level statement creates
(`ns.registry = { run: function () {} }`) are hoisted to module
functions, like those stored by assignment.
- `table[key](...)` on exactly one constant object whose members are all
hoisted functions, with a string key, compares the key against each
member and calls its function directly; any other key reads the table
as written. The members do not use `this`, the table has no accessors,
and comparing a string runs no code.
- A read of a primitive member written in the object's own literal
(`Kind = { CIRCLE: 1 }`) is replaced by the literal. Members stored
after the object was created stay reads (they could be read first).
- A read in another JavaScript file of a hoisted namespace member refers
to it through the defining module's exports: the member joins the
module's `module.exports = { ... }` literal, and the reading file binds
the module right after the top-level statement that requires it. This
applies when the exports object is constant and nothing enumerates,
reflects on, or lets it escape, and the module does not require the
reading file.
The overlay's cross-file class types and member function types are now
on by default (SCRIPTC_SHAPE_CROSS_FILE=0 and
SCRIPTC_SHAPE_MEMBER_FUNCTIONS=0 turn them off).
The analysis records computed callee reads as computed reads.
New npm case static-tables (with a JavaScript consumer) and
tests/harness/static-tables.test.ts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With the overlay on, a lowering refusal no longer fails the build. The
annotations the refusal depends on are dropped and the program is loaded
and lowered again, so those declarations compile untyped, as they do
without the overlay:
- Each overlay annotation carries an id (`@scriptcShape #<pos>`).
- shape-inference/fallback.ts blames the annotations of the declarations
enclosing a refusal and of the declarations its expression names,
following unannotated functions' returns.
- Rounds that stop converging drop whole files' overlays, then the
overlay for the build. SC6103 reports what was dropped.
Also:
- The tsc gate's project world is built over the files as written; it
used to see the overlay's files, which failed the typecheck of three
npm cases.
- Arrays a builtin creates (match results, `new Array(n)`) are no longer
typed by the overlay: they may have holes, and the library types them.
- compile() accepts `shapeOverlay: true`.
tests/harness/npm-shape-overlay.test.ts runs the whole npm suite with the
overlay on (npm.test.ts's body moves to npm-suite.ts): every case builds
and runs as in the default build, and a refused build that compiles with
the overlay matches Node.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The comments on net.sockWrite and net.sockWritableLength still described
the unsent bytes; the runtime counts each write not yet completely sent in
full.
The close step drops a socket's listener lists so that a listener which
captures its socket does not keep it alive; the 'drain' list was not
among them. A once('drain') listener still pending when the peer
disconnected, or an on('drain') listener, held its socket and the
socket's buffers after 'close'. The close step now drops the 'drain'
listeners with the others.
A loop to a typed array's length reads the length again on every
iteration, so a body that assigns the array's binding changes the limit
for the next iteration. The strip-mined form took the length once per
block of 64 iterations, and after the body assigned a shorter array it
ran on to the block's end, reading past the new length.
The loop to a typed array's length is now strip-mined only when the
array is a local of the function that the body never assigns
(stableReceiver on the match); others keep the budget and test the
length on every iteration.
New corpus program: native-worker-length-loop-reassign.ts.
RSA_PSS_SALTLEN_AUTO recovers the salt length from the signature and
does not enforce the key's minimum salt length, as in Node (OpenSSL).
The runtime's exemption now says so, and corpus 8178's header, which
said the minimum was still enforced, describes what it checks.
Untyped functions built each member of an object literal with
scr_dyn_key_set: a lookup of the key in the new object, a walk of its
empty prototype chain, a second lookup before the append, and a hash of
the key to share its bytes. The first cached read of the object then
computed its shape from its keys, hashing each key again.
A member whose key is a constant string other than __proto__ now owns a
store cache site and calls scr_js_store_probe, which appends the member
through the cached shape transition, takes the key bytes from the shape,
and leaves the object shaped. Members the probe does not take (Error
field names, members past the 64th, a site's first object of a shape)
keep scr_dyn_key_set, as do computed keys and literals in module
initializers.
push() of an empty Buffer outside object mode put a zero-length entry
into the read buffer, and the chunks pushed after it stalled behind it:
a 'data' listener saw nothing more and a readline interface never
closed. Node's readableAddChunk adds no empty chunk outside object mode;
it clears the reading flag and may read more. The runtime now does the
same, as it already did for an encoded push that decodes to nothing.
