HTTP/2 streams do not use the fact that the native
`StreamBase::Shutdown()` is asynchronous by default and
always finish synchronously.
Adding a status code for this scenario allows skipping an
expensive `MakeCallback()` C++/JS boundary crossing.
PR-URL: https://github.com/nodejs/node/pull/25609
Reviewed-By: Sam Roberts <vieuxtech@gmail.com>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
- Instead of creating the console extensions eagerly during bootstrap
and storing them on an object, wrap the code into a function to be
called during `installAdditionalCommandLineAPI` only when the
extensions are actually needed, which may not even happen if the
user do not use the console in an inspector session, and does not
need to happen during bootstrap unconditionally.
- Simplify the console methods wrapping and move the `consoleFromVM`
storage to `internal/util/inspector.js`
PR-URL: https://github.com/nodejs/node/pull/25450
Reviewed-By: John-David Dalton <john.david.dalton@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de>
This patch:
- Clarifies the dependency of the ESM loader initialization
(`process.cwd()` and the value of `--loader`) in `node.js`.
- Moves the initialization of the per-isolate `importModuleDynamically`
and `initializeImportMetaObject` callbacks into `node.js`
- Moves the initialization of the ESM loader into
`prepareUserCodeExecution()` since it potentially involves
execution of user code (similar to `--require` for CJS modules).
PR-URL: https://github.com/nodejs/node/pull/25530
Reviewed-By: Gus Caplan <me@gus.host>
For non-Windows platforms, the happy path behavior should be
identical. On Windows, uv_os_uname() attempts to use
RtlGetVersion() before falling back to the deprecated
GetVersionExW() that Node was previously using.
PR-URL: https://github.com/nodejs/node/pull/25600
Reviewed-By: Bartosz Sosnowski <bartosz@janeasystems.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
This patch replaces usage of `filter` in such a way that it will be
compatible with Python 3. Also, this patch replaces the usage of `map`
to do a side-effect work with normal `for` loop.
PR-URL: https://github.com/nodejs/node/pull/25583
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Denys Otrishko <shishugi@gmail.com>
Currently, there are a number of compiler warnings like the following:
In file included from ../src/node_api.h:11:
../src/node_api_types.h:13:21:x
warning: extra tokens at end of #ifdef directive [-Wextra-tokens]
#ifdef NAPI_VERSION >= 4
This commit changes the #ifdef macros to #if.
PR-URL: https://github.com/nodejs/node/pull/25635
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Pass along errors from `Revert()` when a security revert
is unknown (which currently applies to all possible values).
Previously, we would unconditionally call `exit()`, which is
not nice for embedding use cases, and could crash because we
were holding a lock for a mutex in `ProcessGlobalArgs()` that
would be destroyed by calling `exit()`.
Also, add a regression test that makes sure that the process
exits with the right exit code and not a crash.
PR-URL: https://github.com/nodejs/node/pull/25466
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
This gets better coverage of the codes, and is more explicit. It also
works around ordering differences in the errors produced by openssl.
The approach was tested with 1.1.0 and 1.1.1, as well as TLSv1.2 vs
TLSv1.3. OpenSSL 1.1.0 is relevant when node is built against a shared
openssl.
PR-URL: https://github.com/nodejs/node/pull/25381
Reviewed-By: Daniel Bevenius <daniel.bevenius@gmail.com>
Reviewed-By: Shigeki Ohtsu <ohtsu@ohtsu.org>
Because llvm on MacOS does not support AVX-512, asm files need to be limited to
AVX-2 support even when they are generated on Linux. fake_gcc.pl returns the
fake llvm banner version for MacOS as if the assembler supports upto AVX-2.
For Windows, makefiles for nmake were updated in OpenSSL-1.1.1 and they are
rewritten into GNU makefile format by hand.
PR-URL: https://github.com/nodejs/node/pull/25381
Reviewed-By: Daniel Bevenius <daniel.bevenius@gmail.com>
Reviewed-By: Shigeki Ohtsu <ohtsu@ohtsu.org>
- Generate and use a list of suppressions that reduce noisiness for
known (non-)issues.
- Use `--zero-fill-buffers` for tests, as they sometimes use
`Buffer.allocUnsafe()` and valgrind reports that as usage
of uninitialized memory.
PR-URL: https://github.com/nodejs/node/pull/25498
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
This commit updates the named curves P-192 (prime192v1), and secp192k1
to 256 bit versions.
The motivation for this is that in Red Hat Enterprise Linux (RHEL) all
ECC curves < 224 bits are removed from OpenSSL provided by the system.
I'm not sure if other distributions do this but these 256 bit curves are
availalbe in OpenSSL 1.1.0j (current version on master) and OpenSSL
1.1.1 so as far as I can tell it should be safe change to make.
PR-URL: https://github.com/nodejs/node/pull/25564
Reviewed-By: Sam Roberts <vieuxtech@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
This commit updates option ciphers from 'RC4' to 'no-such-cipher' in
test/parallel/test-tls-handshake-error.js.
The motivation for this change is that this test is verifying that a
'no ciphers match' error be thrown, but 'RC4' might be among the ciphers
supported by the OpenSSL version when dynamically linking. I ran into
this specific issue when dynamically linking against OpenSSL 1.1.1 on
RHEL8 using https://github.com/nodejs/node/pull/25381.
PR-URL: https://github.com/nodejs/node/pull/25534
Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de>
Reviewed-By: Sam Roberts <vieuxtech@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
