Use the abort-aware iterator wrapper in the no-transform pipeTo()
path so a pending source read does not block AbortSignal handling.

Fixes: #64014

Signed-off-by: Kamat, Trivikram <16024985+trivikr@users.noreply.github.com>
Assisted-by: openai:gpt-5.5
PR-URL: #64015
Fixes: #64014
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Filip Skokan <panva.ip@gmail.com>

PR #62782 updated the keepAliveTimeout default from 5s to 65s and
correctly updated server.keepAliveTimeout (line 2006), but the
http.createServer() options table at line 3775 was not updated.

Fix the remaining reference to reflect the current default of 65000ms.

Signed-off-by: Jahanzaib iqbal <jahanzaib@gsoftconsulting.com>
PR-URL: #63974
Reviewed-By: Tim Perry <pimterry@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>

Emit the standard JUnit timestamp (ISO 8601) on <testsuite> elements,
which the reporter was omitting. The suite start time is reconstructed
as end-minus-duration, because the runner reports a suite's test:start
lazily (when its first subtest reports), which would otherwise record a
time close to the suite's end.

Fixes: #64028
Signed-off-by: sangwook <rewq5991@gmail.com>
PR-URL: #64029
Reviewed-By: Moshe Atlow <moshe@atlow.co.il>
Reviewed-By: Jacob Smith <jacob@frende.me>

`EncodeUtf8String`, which backs `TextEncoder.prototype.encode()`, copied
the entire source string out of the V8 heap into a `MaybeStackBuffer`
(via `WriteOneByteV2`/`WriteV2`) before encoding, allocating on the heap
for strings larger than the stack buffer. `EncodeInto` already avoids
this by reading the flat content directly through
`v8::String::ValueView`.

Read the flat content via `ValueView` instead. Because `ValueView` holds
a `DisallowGarbageCollection` scope, the backing store cannot be
allocated while it is alive, so the view is used in two short scopes:
one to validate and compute the exact UTF-8 length, and one to encode
directly into the backing store after allocation. Flattening is cached
on the string, so re-acquiring the view is cheap. The rare unpaired
surrogate path still copies into a mutable buffer for in-place
`to_well_formed_utf16`.

benchmark/util/text-encoder.js (op=encode, n=1e6, 12 runs each):

                       len=256  len=1024  len=8192
  ascii                 +14.1%    +23.5%    +43.9%
  one-byte (latin1)     +14.4%    +22.2%    +12.3%
  two-byte (utf-16)     +16.7%    +20.4%    +15.5%

len=32 uses the unchanged small-string path (~noise). The untouched
encodeInto path stayed flat (-2.0%..+0.5%) across all configurations.

Signed-off-by: Yagiz Nizipli <yagiz@nizipli.com>
PR-URL: #63897
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Filip Skokan <panva.ip@gmail.com>

Signed-off-by: Matthew Aitken <maitken033380023@gmail.com>
PR-URL: #64036
Refs: w3c/FileAPI@cd1d1da
Refs: nodejs/undici@0d6ecc5
Reviewed-By: René <contact.9a5d6388@renegade334.me.uk>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Gürgün Dayıoğlu <hey@gurgun.day>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Mattias Buelens <mattias@buelens.com>

PR-URL: #64039
Reviewed-By: Antoine du Hamel <duhamelantoine1995@gmail.com>
Reviewed-By: Filip Skokan <panva.ip@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>

Merely loading the repl builtin used to eagerly require acorn and
acorn-walk (~250KB of JS) and create an entire V8 context via
vm.runInNewContext() just to enumerate global property names, even
though both are only needed once REPL input is actually parsed or
tab-completion is used.

Require acorn and acorn-walk at their function-level use sites
instead, and wrap the global builtins set in getLazy(). The cost
moves to the first preview/completion/recoverable-error check,
where the one-time ~1ms is imperceptible.

Benchmark results (Linux x64, misc/startup-core.js, 20 runs):

  require-builtins.js: +3.12% ops/s (t=3.57, p<0.01)
  import-builtins.mjs: +2.20% ops/s (t=3.00, p<0.01)

In isolation, require('repl') drops from 4.64ms to 2.84ms (-39%)
and interactive `node -i` startup improves by ~13%.

Signed-off-by: Daijiro Wachi <daijiro.wachi@gmail.com>
PR-URL: #63879
Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com>

Signed-off-by: Renegade334 <contact.9a5d6388@renegade334.me.uk>
PR-URL: #64043
Refs: #63078
Reviewed-By: Antoine du Hamel <duhamelantoine1995@gmail.com>
Reviewed-By: Michaël Zasso <targos@protonmail.com>
Reviewed-By: Filip Skokan <panva.ip@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>

V8's Minor Mark-Sweep collector, enabled with the --minor-ms flag
(available since Node.js 22), reports garbage collection performance
entries with kind kGCTypeMinorMarkSweep (value 2). perf_hooks exposed
constants for every other GC kind (major, minor, incremental, weakcb)
but not this one, so consumers inspecting performanceEntry.detail.kind
had no constant to compare against and saw an unmapped value.

Expose it as perf_hooks.constants.NODE_PERFORMANCE_GC_MINOR_MARK_SWEEP,
mirroring the existing v8::GCType mappings, and document it alongside
the other GC kinds.

Signed-off-by: Attila Szegedi <attila.szegedi@datadoghq.com>
PR-URL: #63877
Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de>
Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com>
Reviewed-By: Ilyas Shabi <ilyasshabi94@gmail.com>
Reviewed-By: Chengzhong Wu <legendecas@gmail.com>

Signed-off-by: Antoine du Hamel <duhamelantoine1995@gmail.com>
PR-URL: #64047
Reviewed-By: René <contact.9a5d6388@renegade334.me.uk>
Reviewed-By: Filip Skokan <panva.ip@gmail.com>
Reviewed-By: Richard Lau <richard.lau@ibm.com>

The method-proxying loop in Readable.prototype.wrap() started at
index 1, silently skipping streamKeys[0] and leaving a method at the
first own-enumerable key unproxied.

This was introduced in ee9e2a2 when ArrayPrototypeForEach(
ObjectKeys(stream), ...) — which iterates from index 0 — was
rewritten as a manual for loop for performance. The faithful
translation starts at 0; a sibling loop converted in the same commit
correctly does so. The bug stayed hidden because keys[0] is usually
_events (a non-function), which the loop's typeof guard skips anyway.

Signed-off-by: Daijiro Wachi <daijiro.wachi@gmail.com>
PR-URL: #64048
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Mattias Buelens <mattias@buelens.com>

Reset process.exitCode before running the coverage threshold check.
This lets the test verify the exit code set by that run() call without
depending on process-global state from earlier coverage runs.

Signed-off-by: Kamat, Trivikram <16024985+trivikr@users.noreply.github.com>
PR-URL: #64051
Refs: https://github.com/nodejs/reliability/issues?q=%22test-runner-run-coverage%22
Reviewed-By: Chemi Atlow <chemi@atlow.co.il>
Reviewed-By: Filip Skokan <panva.ip@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Jacob Smith <jacob@frende.me>

Signed-off-by: Guilherme Araújo <arauujogui@gmail.com>
PR-URL: #63702
Reviewed-By: Stephen Belanger <admin@stephenbelanger.com>
Reviewed-By: Gürgün Dayıoğlu <hey@gurgun.day>

Return false for Ed25519 and Ed448 one-shot verification when the public
key or signature R component is a known low-order point. This keeps key
import behavior unchanged while making WebCrypto verification match WPT
expectations across OpenSSL variants.

Remove the stale WPT expected-failure entry and add focused regression
coverage for both curves.

Closes: #54572

Signed-off-by: Filip Skokan <panva.ip@gmail.com>
PR-URL: #64026
Fixes: #54572
Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com>
Reviewed-By: Tobias Nießen <tniessen@tnie.de>

Signed-off-by: Ijtihed Kilani <ijtihedk@gmail.com>
PR-URL: #64022
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de>
Reviewed-By: Aviv Keller <me@aviv.sh>

Include the coverage filename when parsing V8 coverage data fails,
and make test-runner coverage assertions report spawned child process
output when the expected coverage report is missing.

Signed-off-by: Kamat, Trivikram <16024985+trivikr@users.noreply.github.com>
Assisted-by: openai:gpt-5.5
PR-URL: #64050
Refs: https://github.com/nodejs/reliability/issues?q=%22parallel%2Ftest-runner-coverage%22
Reviewed-By: Moshe Atlow <moshe@atlow.co.il>
Reviewed-By: Chemi Atlow <chemi@atlow.co.il>

Disable connection pooling for the client request so server.close()
does not wait on an idle keep-alive socket before retrying the test.

This keeps the test focused on request timeout behavior instead of
HTTP keep-alive teardown timing.

Signed-off-by: Kamat, Trivikram <16024985+trivikr@users.noreply.github.com>
PR-URL: #64052
Refs: https://github.com/nodejs/reliability/issues?q=%22test-http-server-consumed-timeout%22
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Filip Skokan <panva.ip@gmail.com>
Reviewed-By: Gürgün Dayıoğlu <hey@gurgun.day>

Signed-off-by: Rawal27 <obviouslykamal@gmail.com>
PR-URL: #63912
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Trivikram Kamat <trivikr.dev@gmail.com>

The basic.any.js test compares elapsed performance.now() with elapsed
Date.now(). On macos15-x64 CI, Date.now() can shift relative to the
monotonic clock and exceed the test's 30 ms tolerance.

Convert the status file to .cjs so the flaky expectation is scoped to
macOS 15 on x64.

Signed-off-by: Kamat, Trivikram <16024985+trivikr@users.noreply.github.com>
PR-URL: #64054
Refs: https://github.com/nodejs/reliability/issues?q=%22wpt%2Ftest-hr-time%22
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Filip Skokan <panva.ip@gmail.com>
Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com>
Reviewed-By: Jason Zhang <xzha4350@gmail.com>

Starting with Node.js 26.0.0, official binaries for AIX are built with
Clang 20.1.

Signed-off-by: Richard Lau <richard.lau@ibm.com>
PR-URL: #64068
Reviewed-By: Chengzhong Wu <legendecas@gmail.com>
Reviewed-By: Beth Griggs <bethanyngriggs@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Stewart X Addison <sxa@redhat.com>

Signed-off-by: parkhojeong <parkhj062@gmail.com>
PR-URL: #63403
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Deokjin Kim <deokjin81.kim@gmail.com>

Signed-off-by: parkhojeong <parkhj062@gmail.com>
PR-URL: #63403
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Deokjin Kim <deokjin81.kim@gmail.com>

Add a samplePerIteration option to monitorEventLoopDelay that records
event loop delay from libuv event loop iterations instead of the timer
interval sampler. The default remains interval-based; existing uses of
monitorEventLoopDelay() keep behaving the same unless the
samplePerIteration option is passed through.

Signed-off-by: Pablo Erhard <pablo.erhardhernandez@datadoghq.com>
PR-URL: #62935
Reviewed-By: Bryan English <bryan@bryanenglish.com>
Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de>
Reviewed-By: James M Snell <jasnell@gmail.com>

The close fixture expects the registered object to still be reachable
when the process emits exit. Keep a strong reference outside setup() so
the assertion does not depend on platform-specific GC timing.

Signed-off-by: Kamat, Trivikram <16024985+trivikr@users.noreply.github.com>
Assisted-by: openai:gpt-5.5
PR-URL: #64085
Refs: https://github.com/nodejs/reliability/issues?q=%22test-process-finalization%22
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Gürgün Dayıoğlu <hey@gurgun.day>
Reviewed-By: Stefan Stojanovic <stefan.stojanovic@janeasystems.com>

V8’s JIT infrastructure requires binaries such as mksnapshot to be run
during the build. However, these binaries must have the same bit-width
as the host platform (e.g. a x86_64 build platform targeting ARMv6 needs
to produce a 32-bit binary).

To work around this issue, allow building the binaries for the host
platform and running them on the build platform with an emulator.

Based on Buildroot’s nodejs-src 0001-add-qemu-wrapper-support.patch.
https://gitlab.com/buildroot.org/buildroot/-/blob/c1d5eada4d4db9eeaa1c44dd1dea95a67c8a70ca/package/nodejs/nodejs-src/0001-add-qemu-wrapper-support.patch

PR-URL: #53899
Reviewed-By: Antoine du Hamel <duhamelantoine1995@gmail.com>
Reviewed-By: Aviv Keller <me@aviv.sh>

Unfortunately, `std::optional<>` implements `operator<` in such a way
that this check will fail for very large generators. Since `bn_g` is
unsigned, if its value does not fit into a single word, we can be
certain that it is at least 2. By only checking the value if it does
indeed fit into a word, the check correctly ignores very large
generators.

Signed-off-by: Tobias Nießen <tniessen@tnie.de>
PR-URL: #64092
Reviewed-By: Filip Skokan <panva.ip@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>

If the exponent does not fit into a single word, previous versions of
Node.js would report an incorrect value or, recently, return `null`.
Change `GetExponentString()` to handle arbitrarily large RSA public
exponents properly.

Signed-off-by: Tobias Nießen <tniessen@tnie.de>
PR-URL: #64093
Refs: #63895
Reviewed-By: Filip Skokan <panva.ip@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>

Signed-off-by: Tobias Nießen <tniessen@tnie.de>
PR-URL: #64094
Reviewed-By: Juan José Arboleda <soyjuanarbol@gmail.com>
Reviewed-By: Filip Skokan <panva.ip@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Edy Silva <edigleyssonsilva@gmail.com>

Signed-off-by: Antoine du Hamel <duhamelantoine1995@gmail.com>
PR-URL: #64046
Reviewed-By: Paolo Insogna <paolo@cowtech.it>
Reviewed-By: René <contact.9a5d6388@renegade334.me.uk>

Rearm the stream's existing Timeout object instead of allocating a new
Timeout and a bound callback on every setTimeout() call. The HTTP
server disarms and rearms the keep-alive timeout twice per request, so
this saves two allocations and the associated async resource churn on
every request over a keep-alive connection.

The async resource is re-initialized on reuse, so async_hooks observes
the same init/destroy sequence as before.

A microbenchmark of the setTimeout(msecs)/setTimeout(0) cycle improves
by 2.2x (205.9ns to 92.8ns), and benchmark/http/simple.js
(type=buffer len=1024 chunks=1 chunkedEnc=0 c=50) improves by 3.70%
(t=2.15, 40 samples per binary with server and load generator pinned
to disjoint CPU sets).

Assisted-by: Claude Fable 5 (Claude Code)
Signed-off-by: Matteo Collina <hello@matteocollina.com>
PR-URL: #64254
Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Minwoo Jung <nodecorelab@gmail.com>

Signed-off-by: Renegade334 <contact.9a5d6388@renegade334.me.uk>
PR-URL: #64325
Refs: #63281
Reviewed-By: Michaël Zasso <targos@protonmail.com>
Reviewed-By: Richard Lau <richard.lau@ibm.com>
Reviewed-By: Marco Ippolito <marcoippolito54@gmail.com>
Reviewed-By: Antoine du Hamel <duhamelantoine1995@gmail.com>