Fixes: #44286
PR-URL: #44293
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>

Avoid magic numbers in the code and use an OpenSSL constant instead.

PR-URL: #44305
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Filip Skokan <[email protected]>

This function was introduced in 2684c90
as an internal helper function. The C++ implementation became a no-op in
a57e2f2 when building against OpenSSL
1.1.0 (instead of OpenSSL 1.0.2), and eventually became a no-op in all
supported OpenSSL versions in 970ce14.
Finally, eb20447 removed the only call
site of setFreeListLength (which was already a no-op at that point).

Refs: #1529
Refs: #10859
Refs: #19794
Refs: #38116
PR-URL: #44300
Reviewed-By: Feng Yu <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>

Remove an unnecessary static_cast<char*>().

Use OPENSSL_secure_zalloc() instead of OPENSSL_secure_malloc() +
memset().

Update the comment describing the function which predates support for
OpenSSL's secure heap.

PR-URL: #44302
Reviewed-By: Filip Skokan <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>

md and mgf1_md are internal variable names and should not appear in
JS error messages. Also include the invalid digest name in the error
message.

PR-URL: #44307
Reviewed-By: Filip Skokan <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

The THROW_ERR_* functions interpret the first argument as a printf-like
format string, which is problematic when it contains unsanitized user
input. This typically happens when a printf-like function is used to
produce the error message, which is then passed to a THROW_ERR_*
function, which again interprets the error message as a format string.

Fix such occurrences by properly formatting error messages using static
format strings only, and in a single step.

PR-URL: #44314
Reviewed-By: Matteo Collina <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Zeyu "Alex" Yang <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

PR-URL: #44322
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Moshe Atlow <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Ian Sutherland <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

PR-URL: #44351
Reviewed-By: Jacob Smith <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

Signed-off-by: Daeyeon Jeong [email protected]
PR-URL: #44355
Fixes: #44142
Reviewed-By: Darshan Sen <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Kohei Ueno <[email protected]>

Fixes: #44298
PR-URL: #44326
Reviewed-By: Paolo Insogna <[email protected]>
Reviewed-By: Matteo Collina <[email protected]>
Reviewed-By: Yongsheng Zhang <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>

Instead of creating an object template for every ContextifyContext,
we now create one object template that can be reused by all
contexts. The native pointer can be obtained through an embdder
pointer field in the creation context of the receiver in the
interceptors, because the interceptors are only meant to be invoked
on the global object of the contextified contexts. This makes
the ContextifyContext template context-independent and therefore
snapshotable.

PR-URL: #44252
Refs: #44014
Refs: #37476
Reviewed-By: Chengzhong Wu <[email protected]>

Include a minimally initialized contextify context in the embedded
snapshot. This paves the way for user-land vm context snapshots.

PR-URL: #44252
Refs: #44014
Refs: #37476
Reviewed-By: Chengzhong Wu <[email protected]>

Access to the global object from within a vm context is intercepted
so it's slow, therefore we should try to avoid unnecessary access
to it during the initialization of vm contexts.

- Remove the Atomics.wake deletion as V8 now does not install it
  anymore.
- Move the Intl.v8BreakIterator deletion into the snapshot.
- Do not query the Object prototype if --disable-proto is not set.

This should speed up the creation of vm contexts by about ~12%.

PR-URL: #44252
Refs: #44014
Refs: #37476
Reviewed-By: Chengzhong Wu <[email protected]>

PR-URL: #44319
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Matteo Collina <[email protected]>

PR-URL: #44241
Fixes: #44023
Fixes: #43675
Reviewed-By: Benjamin Gruenbaum <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>

By default, the debugger would query the specified inspector
sever port to see if it's available before starting the server,
and it would keep retrying until a timeout (previously 9999 ms)
is reached. This timeout seems to be longer than necessary. This
patch decreases the timeout to 3 seconds.

PR-URL: #44359
Reviewed-By: Moshe Atlow <[email protected]>
Reviewed-By: Richard Lau <[email protected]>

Add basic tests for providers when using OpenSSL 3.x.

PR-URL: #44148
Reviewed-By: Michael Dawson <[email protected]>
Reviewed-By: Rafael Gonzaga <[email protected]>

PR-URL: #44345
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Chengzhong Wu <[email protected]>

Allows APM vendors to generate a diagnostic report without calling into
JavaScript. Like, from their own message channels interrupting the
isolate and generating a report on demand.

PR-URL: #44255
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Gireesh Punathil <[email protected]>

Fixes: #41330
PR-URL: #43803
Reviewed-By: Joyee Cheung <[email protected]>

symmetric_key_len_ is always equal to symmetric_key_.size(). Storing it
separately is redundant and has no significant benefit.

PR-URL: #44346
Reviewed-By: Zeyu "Alex" Yang <[email protected]>
Reviewed-By: Filip Skokan <[email protected]>

The test in `test/parallel/test-stream-writable-callback-twice.js` is
duplicated in `test/parallel/test-stream-writable-write-cb-twice.js`.

PR-URL: #44313
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Yongsheng Zhang <[email protected]>

Functions declared in anonymous namespaces are not necessarily to be
marked as static.

PR-URL: #44301
Refs: #44141
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Michael Dawson <[email protected]>

Change the underlying type of DebugCategory to unsigned int so that it
can be safely cast to an unsigned type without having to worry about
negative values, removing the need for the DCHECK_GE statements.

To fully benefit from type safety, remove DebugCategory::CATEGORY_COUNT
and instead add a constexpr kDebugCategoryCount.

Remove the second argument from EnabledDebugList::set_enabled() and
EnabledDebugList::Parse() because it was always set to true.

PR-URL: #44350
Reviewed-By: Chengzhong Wu <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>

PR-URL: #44377
Reviewed-By: Chengzhong Wu <[email protected]>
Reviewed-By: Daeyeon Jeong <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Adrian Estrada <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>

PR-URL: #44344
Fixes: #43115
Reviewed-By: Rafael Gonzaga <[email protected]>
Reviewed-By: Matteo Collina <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Trivikram Kamat <[email protected]>

sequential/test-child-process-execsync and
parallel/test-child-process-spawnsync-timeout are both flaky
on azure Windows machines, where it may take longer for Node.js
to launch and receive output from child processes. These tests
work by spawning a child processes that is supposed to sleep
for a long time, but the option is configured so that Node.js
would terminate them early when a shorter timeout is reached.
Then the tests assert that the time taken for the whole thing
is shorter than the specified sleep time (meaning the process
don't actually get to sleep for that long). To make the tests
less brittle on azure Windows, this patch raises the sleep
times in those tests on Windows platform, so that the overhead
can be taken into account there.

PR-URL: #44375
Refs: nodejs/build#3014
Reviewed-By: Moshe Atlow <[email protected]>
Reviewed-By: Nitzan Uziely <[email protected]>

Unify the implementation and perform the same OpenSSL calls regardless
of whether the key and/or salt are empty. This simplifies the code and
improves coverage.

Refs: #44201
PR-URL: #44272
Reviewed-By: Filip Skokan <[email protected]>

PR-URL: #44367
Reviewed-By: Ruben Bridgewater <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Kohei Ueno <[email protected]>

This doesn't include the other tests for the result when
running a specific test in WPT.

Signed-off-by: Daeyeon Jeong [email protected]
PR-URL: #44238
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Filip Skokan <[email protected]>
Reviewed-By: Anto Aravinth <[email protected]>

Following the documentation at:
https://github.com/nodejs/node/blob/main/doc/contributing/maintaining-icu.md#time-zone-data

Refs: nodejs/Release#773
Refs: unicode-org/icu-data#29

PR-URL: #44283
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Michael Dawson <[email protected]>

There is no need to explicitly branch based on num_curves or on the
return value of the second call to EC_get_builtin_curves. Remove
unnecessary branches and replace the loop with a functional transform.

PR-URL: #44309
Reviewed-By: Filip Skokan <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #43476
Reviewed-By: Moshe Atlow <[email protected]>

The lint rule was not catching all occurences of unsafe primordials use,
and was too strict on some methods.

PR-URL: #43476
Reviewed-By: Moshe Atlow <[email protected]>

Doesn't change much but communicates to readers that these
are compile-time conditionals.

PR-URL: #44291
Reviewed-By: Matteo Collina <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Feng Yu <[email protected]>
Reviewed-By: Rich Trott <[email protected]>

This changeset fixes two consistency issues in this single example that
guides releasers on how to push the tag to the remote repo.

The first issue is that the remote git reference is consistently
referred as `upstream` elsewhere in the guide.

The second issue being that the version example used in the rest of the
guide should be a fictional `v1.2.3` as stated in the "How to create a
release" guide notes:

>  * Examples will use the fictional release version `1.2.3`.

PR-URL: #44385
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Rafael Gonzaga <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

Signed-off-by: Daeyeon Jeong [email protected]
PR-URL: #44292
Refs: https://streams.spec.whatwg.org/#default-reader-release-lock
Refs: https://streams.spec.whatwg.org/#byob-reader-release-lock
Refs: https://developer.mozilla.org/en-US/docs/Web/API/ReadableStreamDefaultReader/releaseLock
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Matteo Collina <[email protected]>

PR-URL: #44419
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>

Added an extra command to make sure to remind me (and other releasers)
to update the `main` branch _before_ cherry-picking the release commit
to the branch.

Adding this command will often avoid unnecessary extra-work in case
there are mismatches between the local state of the main branch and
remote.

PR-URL: #44384
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Rafael Gonzaga <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>

Return `undefined` when no source map is found for the given filename on
`findSourceMap`.

PR-URL: #44397
Fixes: #44391
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>

PR-URL: #44422
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Moshe Atlow <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

PR-URL: #44370
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Moshe Atlow <[email protected]>

Refs: v8/v8@10.2.154.13...10.2.154.15
PR-URL: #44294
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Jiawen Geng <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>

This will allow distribution packages to select an alternative
location for the unofficial libnode.so. For example, on Fedora it
will install into /usr/lib64 on 64-bit systems.

Signed-off-by: Stephen Gallagher <[email protected]>
PR-URL: #44361
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>

The previous implementation was typically compiled to a fair amount of
code even though all inputs are available at compile time.

The fact that GetOpenSSLVersion() returns a std::string and used an
uninitialized buffer with snprintf made it impossible to make
GetOpenSSLVersion() a constexpr, and compilers would typically emit code
to dynamically construct the resulting string.

The simplified implementation usually boils down to a few mov
instructions.

(Ideally, this function could be a constexpr returning a
std::string_view, but that does not have any advantage in the current
design of node::Metadata::Versions which stores versions as
std::string instances.)

Also make the function static since it is not in an anonymous namespace
and change the argument types and the return type of search() to
types that are more appropriate, semantically. (The use of snprintf
previously made this difficult.) Lastly, make the n argument of search()
optional because the simplified implementation always sets it to 0
except during recursive calls within search() itself.

PR-URL: #44395
Reviewed-By: Darshan Sen <[email protected]>
Reviewed-By: Shelley Vohr <[email protected]>
Reviewed-By: Minwoo Jung <[email protected]>

PR #43954 introduced the use of f-strings to `test.py` making the script
only work with python versions >= 3.6. Remove that use to make it
compatible again with older python versions.

PR-URL: #44407
Reviewed-By: Darshan Sen <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: James M Snell <[email protected]>

PR-URL: #44202
Reviewed-By: James M Snell <[email protected]>

Context's default IsCodeGenerationFromStringsAllowed value can be
changed by v8 flag `--disallow-code-generation-from-strings`. Restore
the value at runtime when delegating the code generation validation to
`node::ModifyCodeGenerationFromStrings`.

The context's settings are serialized in the snapshot. Reset the setting
values to its default values before the serialization so that it can be
correctly re-initialized after deserialization at runtime.

PR-URL: #44324
Fixes: #44287
Reviewed-By: Joyee Cheung <[email protected]>
Reviewed-By: James M Snell <[email protected]>

PR-URL: #44057
Reviewed-By: James M Snell <[email protected]>

On some machines the report OOM tests can take too long to complete,
resulting in a timeout. This splits the test into several different
smaller tests to reduce the flakiness of it.

PR-URL: #44389
Refs: nodejs/reliability#356
Reviewed-By: Chengzhong Wu <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

If `--max-old-space-size` is passed on the command line, it takes
precedence over `resourceLimits.maxOldSpaceSizeMb` passed to the worker
thread. IMO this is a bug, but seems unlikely to change(?), so let's
start by documenting it. See the attached issue for more details.

Refs: #43991
PR-URL: #43992
Reviewed-By: James M Snell <[email protected]>

PR-URL: #44416
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: James M Snell <[email protected]>

The "node_http2.h" include reordering enforced by clang-format broke
Electron's Node.js upgrade on Windows. ssize_t is a part of the POSIX
standard and it's not available on Windows, so the fix for this is to
include "node.h" which typedefs it on Windows in
https://github.com/nodejs/node/blob/bb4dff783ddb3b20c67041f7ccef796c335c2407/src/node.h#L212-L220.

Refs: electron/electron#35350 (comment)
Signed-off-by: Darshan Sen <[email protected]>
PR-URL: #44393
Reviewed-By: Matteo Collina <[email protected]>
Reviewed-By: Shelley Vohr <[email protected]>
Reviewed-By: Rafael Gonzaga <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: James M Snell <[email protected]>

PR-URL: #44408
Reviewed-By: Ruy Adorno <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Michaël Zasso <[email protected]>
Reviewed-By: James M Snell <[email protected]>

PR-URL: #44435
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Kohei Ueno <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: LiviaMedeiros <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>

PR-URL: #44446
Reviewed-By: Moshe Atlow <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>

PR-URL: #44411
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>

PR-URL: #44414
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

PR-URL: #44421
Refs: #44401
Reviewed-By: Chengzhong Wu <[email protected]>
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Juan José Arboleda <[email protected]>

This does not fix occurrences in test/parallel/test-dns-* because those
tests contain unrelated pre-existing bugs that would cause the tests to
fail with this fix. This unrelated bug in those tests should be fixed
separately before the use of mustNotCall() can be fixed in those files.

PR-URL: #44022
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Juan José Arboleda <[email protected]>
Reviewed-By: James M Snell <[email protected]>

oaep_label is the name of the C++ variable. Use oaepLabel instead, which
is the name of the documented option.

PR-URL: #44455
Reviewed-By: Filip Skokan <[email protected]>
Reviewed-By: theanarkh <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

node::base64_encode() uses ::base64_encode() when the mode is
Base64Mode::NORMAL, so base64_select_table is only ever called for
Base64Mode::URL and thus only ever returns base64_table_url, but never
base64_table.

Also move base64_table_url into base64-inl.h.

PR-URL: #44425
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Minwoo Jung <[email protected]>

Derived classes' fields are already destroyed if the virtual methods are
invoked in the base class's destructor. It is not safe to call virtual
methods in base's dtor.

PR-URL: #44424
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Minwoo Jung <[email protected]>

The original heap prof tests can take too long to complete on
azure Windows machines, resulting in timeouts. Split them into
smaller tests and move them into the parallel directory to
speed up the execution.

PR-URL: #44388
Refs: nodejs/reliability#356
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: James M Snell <[email protected]>

Previously the tests required that Node.js finish the initialization
of the watchdog thread and fires the timeout within 100ms, which
can be difficult on certain systems under load. This patch
relaxes the requirement to 2000ms. If there is a bug and the
timeout can actually be escaped, raising the timeout to 2000ms
would not make a difference anyway.

PR-URL: #44433
Refs: nodejs/reliability#333
Refs: nodejs/reliability#361
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Chengzhong Wu <[email protected]>
Reviewed-By: Stephen Belanger <[email protected]>

On Windows it might take too long for the parent to start the
communication with a child process, so by the time the parent
starts its own timer, the child process might have already
completed running, and the parent in those tests won't have a
chance to terminate these child processes after the timeout.
To address this issue, raise the time for which the child is
supposed to run to make sure that the parent starts
its own timer before the child terminates in the tests.
Also, split the test into smaller ones to reduce the overhead.

PR-URL: #44390
Refs: nodejs/reliability#356
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Stephen Belanger <[email protected]>

PR-URL: #44386
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: James M Snell <[email protected]>

PR-URL: #44437
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

There were missing imports in the example for using
`EventEmitterAsyncResource`, so I added them in both ESM and CJS.

PR-URL: #44337
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Kohei Ueno <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Franziska Hinkelmann <[email protected]>

PR-URL: #44429
Reviewed-By: Moshe Atlow <[email protected]>
Reviewed-By: Benjamin Gruenbaum <[email protected]>
Reviewed-By: Franziska Hinkelmann <[email protected]>

This commit updates the assert.snapshot() fixtures to better
match the code style of the rest of the codebase.

PR-URL: #44429
Reviewed-By: Moshe Atlow <[email protected]>
Reviewed-By: Benjamin Gruenbaum <[email protected]>
Reviewed-By: Franziska Hinkelmann <[email protected]>

PR-URL: #44429
Reviewed-By: Moshe Atlow <[email protected]>
Reviewed-By: Benjamin Gruenbaum <[email protected]>
Reviewed-By: Franziska Hinkelmann <[email protected]>

PR-URL: #44306
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>

The test previously created two fs.promises.open calls on the
same file with w+ back-to-back, and one of them could fail
when checking the contents of that file if the other happened
to be opening the file for write. Split them into different
tests (with different tmpdir) to avoid the race.

PR-URL: #44380
Refs: nodejs/reliability#354
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Evan Lucas <[email protected]>

Original commit message:

    Reland "Fix Context PromiseHook behaviour with debugger enabled"

    This is a reland of commit 872b7faa32d837f9b166d750328357f856168e72

    Original change's description:
    > Fix Context PromiseHook behaviour with debugger enabled
    >
    > This is a solution for #43148.
    >
    > Due to differences in behaviour between code with and without the debugger enabled, some promise lifecycle events were being missed and some extra ones were being added. This change resolves this and verifies the event sequence is consistent between code with and without the debugger.
    >
    > Change-Id: I3dabf1dceb14233226b1752083d659f1c2f97966
    > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3779922
    > Reviewed-by: Victor Gomes <[email protected]>
    > Commit-Queue: Camillo Bruni <[email protected]>
    > Reviewed-by: Camillo Bruni <[email protected]>
    > Cr-Commit-Position: refs/heads/main@{#82132}

    Change-Id: Ifdd407261c793887fbd012d5a04ba36b3744c349
    Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3805979
    Reviewed-by: Camillo Bruni <[email protected]>
    Reviewed-by: Clemens Backes <[email protected]>
    Commit-Queue: Camillo Bruni <[email protected]>
    Reviewed-by: Victor Gomes <[email protected]>
    Cr-Commit-Position: refs/heads/main@{#82575}

Refs: v8/v8@ff8d67c
Fixes: #43148
Fixes: #44415
PR-URL: #44423
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Stephen Belanger <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Jiawen Geng <[email protected]>
Reviewed-By: Gerhard Stöbich <[email protected]>

Allow add-ons to interrupt JavaScript execution, and wake up loop if it
is currently idle.

PR-URL: #44362
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Gerhard Stöbich <[email protected]>

Given that the divisor is not zero, the result of dividing unsigned
integers is an unsigned integer that is always rounded down, i.e.,
there is no need to call std::trunc(). Doing so unnecessarily yields
a floating-point number, requiring the result to be cast to an unsigned
integer again.

PR-URL: #44467
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Filip Skokan <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #44470
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Filip Skokan <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>

If sk_X509_new() returns NULL or if sk_X509_push() fails, return instead
of silently ignoring the error.

PR-URL: #44410
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Filip Skokan <[email protected]>

Refs: #11321
Refs: #17384
PR-URL: #44482
Reviewed-By: Filip Skokan <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>

`Environment::GetCurrent` may not available in the context of OOM.
Removes the cyclic `Environment::GetCurrent` and `env->isolate()`
calls to ensure both `isolate` and `env` is present if available.

However, this behavior is not guaranteed. As
`Environment::GetCurrent` didn't allocate new handles in the heap,
when a Context is entered it can still get the valid env pointer.
Removes the unstable assertion of the absence of env in the test.

PR-URL: #44398
Reviewed-By: Rafael Gonzaga <[email protected]>

When a new context with a different security token is entered, or
when no context is entered, `StackTrace::CurrentStackTrace` need to
be explicitly set with flag `kExposeFramesAcrossSecurityOrigins` to
avoid crashing.

PR-URL: #44398
Reviewed-By: Rafael Gonzaga <[email protected]>

PR-URL: #44486
Reviewed-By: Myles Borins <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>

This is not a function and should not use the term "return" to describe
its type or value.

PR-URL: #44481
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Harshitha K P <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>

We already do this in some places. This adds the digest name to
remaining uses of ERR_CRYPTO_INVALID_DIGEST except for one occurrence in
crypto_sig.cc that would require significant refactoring due to the
unusual error handling there.

PR-URL: #44468
Reviewed-By: Filip Skokan <[email protected]>
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>

Fixes #44471

PR-URL: #44475
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>

In 14.2 in the specification, `error` should be signal’s abort reason.
The current behavior seems to assume that only an `AbortError` instance
is given as signal’s abort reason.

Refs: https://streams.spec.whatwg.org/#readable-stream-pipe-to

Signed-off-by: Daeyeon Jeong [email protected]
PR-URL: #44418
Reviewed-By: Matteo Collina <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Benjamin Gruenbaum <[email protected]>

PR-URL: #44450
Reviewed-By: Benjamin Gruenbaum <[email protected]>
Reviewed-By: Antoine du Hamel <[email protected]>
Reviewed-By: Franziska Hinkelmann <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Minwoo Jung <[email protected]>

PR-URL: #44511
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Daeyeon Jeong <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>
Reviewed-By: Feng Yu <[email protected]>

PR-URL: #44510
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Daeyeon Jeong <[email protected]>
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>
Reviewed-By: Tobias Nießen <[email protected]>
Reviewed-By: Feng Yu <[email protected]>

Despite being named onend_arg, the pointer is passed both to the
onend_cb and to the onhello_cb. Rename it to cb_arg, which matches the
name of the class field cb_arg_.

PR-URL: #44500
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>

PR-URL: #44512
Reviewed-By: Mohammed Keyvanzadeh <[email protected]>
Reviewed-By: Darshan Sen <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>