A recorded baseline entry now holds a hash of its program's inputs: the
entry file and its relative imports (paths and contents) and the
optimization level. --baseline-cache uses an entry only when the program
still hashes the same, so a program or an import that changed since the
record builds its baseline, instead of a recorded build error or artifact
hash of the old sources settling the comparison as both-fail or identical.
Records take schema 2; older records are ignored.
The full gate takes 30 to 75 minutes on one Sandbox, which made it the
longest step of a validation round. It can now be split and merged:
- --shard=i/n splits the bench and corpus programs (the self-hosted
compiler runs in shard 1, or alone with --set=self), and
scripts/zero-cost-merge.mjs merges the shards' --json reports into one
verdict, recomputing the compile-time intervals over every shard.
- --record-baseline-cache records the baseline's artifact hashes once;
--baseline-cache skips the baseline build of every program whose
candidate artifacts hash the same. The record is tied to the built
baseline checkout, toolchain and paths.
- The candidate-source self build overlaps the comparison builds on
hosts with room for it.
pnpm test:sandbox can reuse a prepared snapshot's LLVM helper and runtime
pack (SCRIPTC_SANDBOX_NATIVE_REFERENCE) when their inputs are identical,
and retries rate-limited Sandbox creation with backoff.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A typed-array loop version runs a checked copy and one or two fast copies
of the loop, and each copy repeats every nested loop together with the
nested loop's own version. A loop whose guard proves only accesses outside
its nested loops, which run once per pass, now runs as written when a
nested loop accesses typed arrays; the nested loops still version their own
accesses. The loop versions as before when the guard proves the bounds of
an access inside a nested loop, or, in a worker program, the ownership of
an access inside a nested while loop, which never versions for ownership
alone.
Spectral-norm's paired row sums no longer copy their column loop for the
row loop's two stores. The worker program's au and atu functions are
102 KB of LLVM text instead of 236 KB each, and the program 519 KB instead
of 786 KB.
Corpus typed-array-nested-loop-versions covers row loops that store past
the end or at negative, fractional and NaN rows, views the nested loop
writes, and a nested while loop.
String.prototype.indexOf took its byte-search fast path on a proven-ASCII
receiver only when the start index was not positive. A positive start went
through the general path: ToIntegerOrInfinity, the surrogate-needle check,
and UTF-16 to byte conversions in both directions, although byte offsets
are UTF-16 indexes in an ASCII string. The fast path now covers every start
index: a positive start truncates and clamps to the length, where only the
empty needle is found, and the search runs from that byte offset.
The implicit toolchain probe includes every system header that the
runtime and vendored sources spell, under __has_include. clang ships
<riscv_vector.h> for every target, and through clang 18 including it
without the RISC-V vector extension is an #error. Since the vendored
zlib's adler32_simd.c spells it, the probe failed with clang 18, which
disabled the native cache and program sharding: a sanitized dev build
then compiled the whole program as one translation unit.
The probe now includes <riscv_vector.h> only where __riscv_vector is
defined, and the probe's header test covers it and a riscv64 target.
Every unbounded recursion repeats some cycle of the call graph, so a
stack check in one function of each cycle bounds it. stackCheckedFunctions
now picks a feedback vertex set of the stack-check graph greedily (the
function with the most caller-callee pairs first) instead of checking
every function on a cycle, and adds a check where a run of unchecked cycle
functions would exceed 32 calls, so the stack used between two checks
stays of the same order as before.
Worker entry polls follow: a synchronous function on a cycle that may
throw needs no entry poll, since a descent that never returns ends at a
checked function of its cycle. NoThrow contracts keep the model in which
every function on a cycle checks, a superset that does not depend on the
choice.
A small helper called once per element, such as
function pick(list: Gene[], rand: number): number {
for (const gene of list) if (rand < gene.c) return gene.sc;
return list[list.length - 1]!.sc;
}
called from a loop over the elements, is often past LLVM's inline cost
threshold once its loop versions and checks are emitted (fasta's
randomCode costs 430 against a threshold of 225), so every element pays a
call, the spills around it, and checks the caller's loop cannot hoist.
findLoopCallees (loop-callees.ts) finds functions that are synchronous,
non-generator, capture nothing, call nothing (no call, callValue,
virtualCall, dynCall, dynInvoke, new, newValue, newPromise, ffiCall or
await), have at most 12 statements and one loop, and have one or two
direct call sites in the module, each inside a loop of its caller. The
emitter marks those calls `alwaysinline` (a call-site attribute, so the
function's definition is unchanged). Library modules and debug builds
keep LLVM's cost decision.
stableFileName used node:path's relative(), which returns paths with `\`
separators on Windows. The module tags hash that name, so Windows builds
named the functions of every non-entry module differently from Linux
builds, and a profile or layout collected on Linux no longer matched them.
Structural positions used the same name.
portableRelativePath computes the relative path with the platform's rules
and writes it with `/` separators, as tsgoPath does for checker paths.
Linux and macOS names are unchanged.
`skip || (flag ? warnA() : warnB())`, with void calls in both arms, as
React's development warnings write it, failed with SC9001 "ternary must
not be void": the statement lowered the logical as a value, and a
conditional of void calls has none. A logical whose result is discarded
now lowers as `if (!skip) ...` (or `if (skip) ...` for `&&`) when its
right operand needs no value: its type maps to void, or it is a logical
or comma expression whose last operand needs none. This holds in
statements and in a for loop's initializer and update, and for a
generic callback's conditional whose type parameter maps to void.
Closes#2322
Co-authored-by: MathurAditya724 <57684218+MathurAditya724@users.noreply.github.com>
The runtime vendored zlib 1.3.1 and archived it at -Os, and host C-lane
builds linked the system libz, so deflate, raw deflate and gzip output
differed from Node's at every level. The vendored copy is now Node
v24.22.0's deps/zlib (Chromium's zlib, 1.3.2.1-motley-285e94b), built at
-O3 with the units and defines Node's zlib.gyp selects per target: the
SSSE3 Adler-32 and SSE2 inflate and slide-hash code on x86-64, the NEON
and ARMv8 CRC-32 code on arm64, and the portable code on wasm32. The SIMD
units check the CPU at run time and give the same output, so compressed
bytes match Node's on every target.
zlibInputs (runtime-pack-common/scripts/zlib-inputs.mjs for the runtime
packs, vendor-inputs.ts for Zig cross builds and the sanitized runtime)
holds the per-target table, and a unit test keeps the two identical.
Every C-lane driver now links the vendored objects instead of -lz.
Chromium's BSD license ships beside zlib's in the runtime packs.
Corpus 8405 compares compressed bytes with Node across levels,
strategies, memLevel, windowBits, dictionaries and flushing streams;
8406 covers chunked inflation with small input pieces and output chunks.
An integer clone reads a number array's element as an 8-byte double. A
number array proven to hold only int32s keeps 4-byte int32 slots
(SCR_ELEM_I32), so the clone read such an array at the wrong width: it
returned wrong values, and near the array's end it read past the slots.
The clone now tests the array's element kind: int32 slots load as 4-byte
int32s, double slots load as doubles and keep the exact-int32 test, and
any other storage bails to the ordinary body.
New corpus program: integer-clone-int32-slots.ts.
TypeScript 7.1 parses the elements of an interface's `extends` and a
class's `implements` as TypeReference nodes, whose names are entity names,
where 7.0 gives ExpressionWithTypeArguments nodes. heritage-elements.ts
recognizes both forms, and the readers that took an element's expression
read either:
- contract-types.ts finds `implements ProgramLifetime` and
`implements Acyclic` written as type references;
- narrowing-stability.ts takes a qualified name as it takes a property
access;
- generic-functions.ts reads an element's name in either form, and keeps
counting a name in a heritage clause as a value reference.
Chris Tate