- Mar 21, 2019
-
-
Yihong Wang authored
For shared lib build, we leave the signal handling for embedding users. In these two test cases: - `parallel/test-process-external-stdio-close-spawn` - `parallel/test-process-external-stdio-close` The pipe is used for stdout and is destroied before child process uses it for logging. So the node executble that uses shared lib build receives SIGPIPE and the child process ends. This change ignores the SIGPIPE in node_main.cc for shared lib case. Refs: https://github.com/nodejs/node/issues/18535 Signed-off-by:
Yihong Wang <yh.wang@ibm.com> PR-URL: https://github.com/nodejs/node/pull/19211 Refs: https://github.com/nodejs/node/issues/18535 Reviewed-By:
Ben Noordhuis <info@bnoordhuis.nl> Reviewed-By:
Colin Ihrig <cjihrig@gmail.com> Reviewed-By:
James M Snell <jasnell@gmail.com> Reviewed-By:
Michael Dawson <michael_dawson@ca.ibm.com> Reviewed-By:
Gireesh Punathil <gpunathi@in.ibm.com> Signed-off-by:
Beth Griggs <Bethany.Griggs@uk.ibm.com>
-
John Barboza authored
On AIX the underlying fsync system call returns EBADF on a file descriptor for an open directory. So avoid running fsync on it. PR-URL: https://github.com/nodejs/node/pull/21298 Reviewed-By:
Richard Lau <riclau@uk.ibm.com> Reviewed-By:
-
Farazmand authored
Original commit message: PPC: fix Regex addi overflow using add insetad of addi when Operand is more than 16 bits long Change-Id: I7f9452381ed8b321ec71e68d0d90485508b69885 Reviewed-on: https://chromium-review.googlesource.com/c/1430619 Commit-Queue: Junliang Yan <jyan@ca.ibm.com> Reviewed-by:Junliang Yan <jyan@ca.ibm.com> Cr-Commit-Position: refs/heads/master@{#59049} Refs: https://github.com/v8/v8/commit/3cc69194b5ee967d0386fdd47de7a84141fef8cc PR-URL: https://github.com/nodejs/node/pull/25874 Reviewed-By:
George Adams <george.adams@uk.ibm.com> Reviewed-By:
-
Ben Noordhuis authored
There is a time window between the first and the last step of processing the clienthello event and the SecurePair may have been destroyed during that interval. Fixes: https://github.com/nodejs/node/issues/26428 PR-URL: https://github.com/nodejs/node/pull/26452 Reviewed-By:
Sam Roberts <vieuxtech@gmail.com> Reviewed-By:
-
Ben Noordhuis authored
This seems to have been broken ever since its introduction 5 years ago in commit 75ea11fc ("tls: introduce asynchronous `newSession`") and no one complained but that's not going to stop me from fixing it anyway because otherwise I can't write a regression test for issue #26428. Refs: https://github.com/nodejs/node/issues/26428 PR-URL: https://github.com/nodejs/node/pull/26452 Fixes: https://github.com/nodejs/node/issues/26428 Reviewed-By:
-
Yihong Wang authored
cctest depends on some internal APIs which don't declare `__declspec(dllexport)` and causes build failure when building node as shared lib on Windows. Since we already have good test coverage in static lib, we decide to skip the cctest in shared lib build on Windows. Signed-off-by:
Refael Ackermann <refack@gmail.com>
-
Sam Ruby authored
Building on Mac OS/X as follows: ``` ./configure --shared make -j4 test ``` Results in: ``` dyld: Library not loaded: @rpath/libnode.67.dylib Referenced from: /Users/rubys/git/node-shared/out/Release/cctest Reason: image not found make: *** [cctest] Abort trap: 6 ``` This change adds the loader path to the runtime path for the `cctest` executable. Backport-PR-URL: https://github.com/nodejs/node/pull/25681 PR-URL: https://github.com/nodejs/node/pull/23168 Reviewed-By:
-
Gabriel Schulhof authored
Calling into the engine from a weak callback is unsafe, however, the engine offers a way to attach a second-pass weak callback which gets called when it is safe to call into JavaScript. This moves the point at which the N-API finalize callback gets called to this latter point. Fixes: https://github.com/nodejs/node/issues/25927 PR-URL: https://github.com/nodejs/node/pull/25992 Backport-PR-URL: https://github.com/nodejs/node/pull/26060 Reviewed-By:
Anna Henningsen <anna@addaleax.net> Reviewed-By:
-
Brian White authored
`http_parser_execute(..., nullptr, 0)` returns either `0` or `1`. The expectation is that no error must be returned if it is `0`, and if it is `1` - a `Error` object must be returned back to user. The introduction of `llhttp` and the refactor that happened during it accidentally removed the error-returning code. This commit reverts it back to its original state. Backport-PR-URL: https://github.com/nodejs/node/pull/25938 Fix: #24585 PR-URL: #24738 Reviewed-By:
Luigi Pinca <luigipinca@gmail.com> Reviewed-By:
-
Julien Gilli authored
Reused parsers can be attached to the domain that corresponds to the active domain when the underlying socket was created, which is not necessarily correct. Instead, we attach parsers to the active domain if there is one when they're reused from the pool. Refs: https://github.com/nodejs/node/issues/25456 PR-URL: https://github.com/nodejs/node/pull/25459 Reviewed-By:
Ruben Bridgewater <ruben@bridgewater.de>
-
- Mar 20, 2019
-
-
Gabriel Schulhof authored
Fixes: https://github.com/nodejs/node/issues/24249 PR-URL: https://github.com/nodejs/node/pull/25556 Backport-PR-URL: https://github.com/nodejs/node/pull/25648 Reviewed-By:
-
Kyle Farnung authored
* Move `napi_get_uv_event_loop` into the `NAPI_VERSION >= 2` section * Move `napi_open_callback_scope`, `napi_close_callback_scope`, `napi_fatal_exception`, `napi_add_env_cleanup_hook`, and `napi_remove_env_cleanup_hook` into the `NAPI_VERSION >= 3` section * Added a missing `added` property to `napi_get_uv_event_loop` in the docs * Added a `napiVersion` property to the docs and updated the parser and generator to use it. * Added usage documentation PR-URL: https://github.com/nodejs/node/pull/19962 Backport-PR-URL: https://github.com/nodejs/node/pull/25648 Reviewed-By:
Gabriel Schulhof <gabriel.schulhof@intel.com> Reviewed-By:
-
Matteo Collina authored
Backport-PR-URL: https://github.com/nodejs/node/pull/25351 PR-URL: https://github.com/nodejs/node/pull/23051 Fixes: https://github.com/nodejs/node/issues/22814 Reviewed-By:
Matteo Collina <matteo.collina@gmail.com> Reviewed-By:
Trivikram Kamat <trivikr.dev@gmail.com> Reviewed-By:
-
Anna Henningsen authored
Make sure that `process.stdin.write()`, and in particular ending the stream, works. Backport-PR-URL: https://github.com/nodejs/node/pull/25351 PR-URL: https://github.com/nodejs/node/pull/23053 Reviewed-By:
-
Anna Henningsen authored
Allow reading from stdio streams that are conventionally associated with process output, since this is only convention. This involves disabling the oddness around closing stdio streams. Its purpose is to prevent the file descriptors 0 through 2 from being closed, since doing so can lead to information leaks when new file descriptors are being opened; instead, not doing anything seems like a more reasonable choice. Fixes: https://github.com/nodejs/node/issues/21203 Backport-PR-URL: https://github.com/nodejs/node/pull/25351 PR-URL: https://github.com/nodejs/node/pull/23053 Reviewed-By:
-
Anna Henningsen authored
Since faking TTY input is not otherwise fake-able, we need support in the test runner for it. Backport-PR-URL: https://github.com/nodejs/node/pull/25351 PR-URL: https://github.com/nodejs/node/pull/23053 Reviewed-By:
-
Richard Lau authored
https://github.com/nodejs/node/pull/17604 refactored the gyp files so that `-blibpath:` on AIX was only set if `node_shared=="true"`. Restore the setting for non-shared builds. Fixes: https://github.com/nodejs/node/issues/25444 Backport-PR-URL: https://github.com/nodejs/node/pull/25521 PR-URL: https://github.com/nodejs/node/pull/25447 Reviewed-By:
-
- Mar 16, 2019
-
-
Richard Lau authored
AIX 7.1 appears to return a different error message compared to AIX 6.1. PR-URL: https://github.com/nodejs/node/pull/25418 Reviewed-By:
-
- Feb 28, 2019
-
-
Rich Trott authored
Refs: https://github.com/nodejs/node/issues/25028 PR-URL: https://github.com/nodejs/node/pull/25032 Reviewed-By:
Daijiro Wachi <daijiro.wachi@gmail.com>
-
Gireesh Punathil authored
The test fails consistently on windows-fanned with vs2017. mark it as flaky while the issue is being progressed, and to keep CI green / amber. Ref: https://github.com/nodejs/node/issues/23621 Backport-PR-URL: https://github.com/nodejs/node/pull/25002 PR-URL: https://github.com/nodejs/node/pull/24714 Reviewed-By:
Rich Trott <rtrott@gmail.com>
-
Daniel Bevenius authored
Currently when building with --debug test/addons-napi/test_threadsafe_function will error: $ out/Debug/node test/addons-napi/test_threadsafe_function/test.js FATAL ERROR: v8::HandleScope::CreateHandle() Cannot create a handle without a HandleScope 1: 0x10004e287 node::DumpBacktrace(__sFILE*) [node/out/Debug/node] 2: 0x1000cd37b node::Abort() [/node/out/Debug/node] 3: 0x1000cd69f node::OnFatalError(char const*, char const*) [/node/out/Debug/node] 4: 0x1004df0b1 v8::Utils::ReportApiFailure(char const*, char const*) [/nodejs/node/out/Debug/node] 5: 0x100a8c0a9 v8::internal::HandleScope::Extend( v8::internal::Isolate*) [/node/out/Debug/node] 6: 0x1004e4229 v8::EmbedderDataFor(v8::Context*, int, bool, char const*) [/node/out/Debug/node] 7: 0x1004e43fa v8::Context::SlowGetAlignedPointerFromEmbedderData(int) [/node/out/Debug/node] 8: 0x10001c26b v8::Context::GetAlignedPointerFromEmbedderData(int) [/node/out/Debug/node] 9: 0x1000144ea node::Environment::GetCurrent(v8::Local<v8::Context>) [/node/out/Debug/node] 10: 0x1000f49e2 napi_env__::node_env() const [/node/out/Debug/node] 11: 0x1000f9885 (anonymous namespace)::v8impl::ThreadSafeFunction:: CloseHandlesAndMaybeDelete(bool) [/node/out/Debug/node] 12: 0x1000fb34f (anonymous namespace)::v8impl::ThreadSafeFunction:: DispatchOne() [/node/out/Debug/node] 13: 0x1000fb129 (anonymous namespace)::v8impl::ThreadSafeFunction:: IdleCb(uv_idle_s*) [/node/out/Debug/node] 14: 0x1011a1b69 uv__run_idle [/node/out/Debug/node] 15: 0x101198179 uv_run [/node/out/Debug/node] 16: 0x1000dfca1 node::Start(...) [/node/out/Debug/node] 17: 0x1000dae50 node::Start(...) [/node/out/Debug/node] 18: 0x1000da56f node::Start(int, char**) [/node/out/Debug/node] 19: 0x10141112e main [/node/out/Debug/node] 20: 0x100001034 start [/node/out/Debug/node] Abort trap: 6 This commit adds two HandleScope's, one to CloseHandlesAndMaybeDelete and one to the lambda. SlowGetAlignedPointerFromEmbedderData will only be called for debug builds: https://github.com/v8/v8/blob/2ef0aa662fe907a1b36ac1abe7d77ad2bcd27733 /include/v8.h#L10440-L10447 Backport-PR-URL: https://github.com/nodejs/node/pull/25002 PR-URL: https://github.com/nodejs/node/pull/24011 Reviewed-By: -
Lars-Magnus Skog authored
The thread_finalize_data and thread_finalize_cb parameters in napi_create_threadsafe_function are optional. Backport-PR-URL: https://github.com/nodejs/node/pull/25002 PR-URL: https://github.com/nodejs/node/pull/22998 Reviewed-By:
-
Gabriel Schulhof authored
* Move class `TsFn` to name space `v8impl` and rename it to `ThreadSafeFunction` * Remove `NAPI_EXTERN` from API declarations, because it's only needed in the header file. Backport-PR-URL: https://github.com/nodejs/node/pull/25002 PR-URL: https://github.com/nodejs/node/pull/22259 Reviewed-By:
Kyle Farnung <kfarnung@microsoft.com> Reviewed-By:
-
Lars-Magnus Skog authored
The idle_running member variable in TsFn is always false and can therefore be removed. Backport-PR-URL: https://github.com/nodejs/node/pull/25002 PR-URL: https://github.com/nodejs/node/pull/22520 Reviewed-By:
Gus Caplan <me@gus.host>
-
Michael Dawson authored
Backport-PR-URL: https://github.com/nodejs/node/pull/25002 PR-URL: https://github.com/nodejs/node/pull/21898 Reviewed-By:
Jon Moss <me@jonathanmoss.me> Reviewed-By:
-
Gabriel Schulhof authored
A condition variable is only created by the thread-safe function if the queue size is set to something larger than zero. This adds null-checks around the condition variable and tests for the case where the queue size is zero. Fixes: https://github.com/nodejs/help/issues/1387 PR-URL: https://github.com/nodejs/node/pull/21871 Backport-PR-URL: https://github.com/nodejs/node/pull/25002 Reviewed-By:
Benjamin Gruenbaum <benjamingr@gmail.com> Reviewed-By:
-
cjihrig authored
private field 'async_context' is not used [-Wunused-private-field] PR-URL: https://github.com/nodejs/node/pull/21597 Refs: https://github.com/nodejs/node/pull/17887 Backport-PR-URL: https://github.com/nodejs/node/pull/25002 Reviewed-By:
Daniel Bevenius <daniel.bevenius@gmail.com>
-
Gabriel Schulhof authored
Bundle a `uv_async_t`, a `uv_idle_t`, a `uv_mutex_t`, a `uv_cond_t`, and a `v8::Persistent<v8::Function>` to make it possible to call into JS from another thread. The API accepts a void data pointer and a callback which will be invoked on the loop thread and which will receive the `napi_value` representing the JavaScript function to call so as to perform the call into JS. The callback is run inside a `node::CallbackScope`. A `std::queue<void*>` is used to store calls from the secondary threads, and an idle loop is started by the `uv_async_t` callback on the loop thread to drain the queue, calling into JS with each item. Items can be added to the queue blockingly or non-blockingly. The thread-safe function can be referenced or unreferenced, with the same semantics as libuv handles. Re: https://github.com/nodejs/help/issues/1035 Re: https://github.com/nodejs/node/issues/20964 Fixes: https://github.com/nodejs/node/issues/13512 Backport-PR-URL: https://github.com/nodejs/node/pull/25002 PR-URL: https://github.com/nodejs/node/pull/17887 Reviewed-By:
-
Anna Henningsen authored
In 180f8650, the test was changed so that the `env` argument of `createInternalRepl()` also contained external environment variables, because keeping them can be necessary for spawning processes on some systems. However, this test does not spawn new processes, and relies on the fact that the environment variables it tests are not already set (and fails otherwise); therefore, reverting to the original state should fix this. Fixes: https://github.com/nodejs/node/issues/21451 Fixes: https://github.com/nodejs/build/issues/1377 Refs: https://github.com/nodejs/node/pull/25219 PR-URL: https://github.com/nodejs/node/pull/25226 Reviewed-By:
Tobias Nießen <tniessen@tnie.de> Reviewed-By:
Denys Otrishko <shishugi@gmail.com> Reviewed-By:
-
James Bunton authored
Refs: https://github.com/nodejs/node/pull/8895 PR-URL: https://github.com/nodejs/node/pull/25123 Reviewed-By:
-
Peter Marshall authored
Original commit message: [cpu-profiler] Fix a leak caused by re-logging existing functions. Don't re-log all existing functions during StartProcessorIfNotStarted(). They will already be in the CodeMap attached to the ProfileGenerator and re-logging them causes leaks. See the linked bug for more details. Bug: v8:8253 Change-Id: Ibb1a1ab2431c588e8c3a3a9ff714767cdf61a88e Reviewed-on: https://chromium-review.googlesource.com/1256763 Commit-Queue: Peter Marshall <petermarshall@chromium.org> Reviewed-by:Yang Guo <yangguo@chromium.org> Cr-Commit-Position: refs/heads/master@{#56336} Refs: https://github.com/v8/v8/commit/525b39619548b4e7fbf48934acb18d3f209321db PR-URL: https://github.com/nodejs/node/pull/25041 Reviewed-By:
Rod Vagg <rod@vagg.org>
-
Rod Vagg authored
PR-URL: https://github.com/nodejs-private/node-private/pull/165
-
Rod Vagg authored
This is a security release. All Node.js users should consult the security release summary at: https://nodejs.org/en/blog/vulnerability/february-2019-security-releases/ for details on patched vulnerabilities. Fixes for the following CVEs are included in this release: * Node.js: Slowloris HTTP Denial of Service with keep-alive (CVE-2019-5737) * OpenSSL: 0-byte record padding oracle (CVE-2019-1559) Notable Changes: * deps: OpenSSL has been upgraded to 1.0.2r which contains a fix for CVE-2019-1559 (https://www.openssl.org/news/secadv/20190226.txt). Under certain circumstances, a TLS server can be forced to respond differently to a client if a zero-byte record is received with an invalid padding compared to a zero-byte record with an invalid MAC. This can be used as the basis of a padding oracle attack to decrypt data. * http: Further prevention of "Slowloris" attacks on HTTP and HTTPS connections by consistently applying the receive timeout set by `server.headersTimeout` to connections in keep-alive mode. Reported by Marco Pracucci (https://voxnest.com). (CVE-2019-5737 / Matteo Collina) PR-URL: https://github.com/nodejs-private/node-private/pull/165
-
Shigeki Ohtsu authored
In openssl s_client on Windows, RAND_screen() is invoked to initialize random state but it takes several seconds in each connection. This added -no_rand_screen to openssl s_client on Windows to skip RAND_screen() and gets a better performance in the unit test of test-tls-server-verify. Do not enable this except to use in the unit test. Fixes: https://github.com/nodejs/io.js/issues/1461 PR-URL: https://github.com/nodejs/io.js/pull/1836 Reviewed-By:
-
Shigeki Ohtsu authored
Reapply b9106137 . Fixes: https://github.com/iojs/io.js/issues/589 PR-URL: https://github.com/iojs/io.js/pull/1389 Reviewed-By:
Fedor Indutny <fedor@indutny.com> Reviewed-By:
-
Shigeki Ohtsu authored
See https://mta.openssl.org/pipermail/openssl-dev/2015-February/000651.html iojs needs to stop using masm and move to nasm or yasm on Win32. Fixes: https://github.com/iojs/io.js/issues/589 PR-URL: https://github.com/iojs/io.js/pull/1389 Reviewed-By:
-
Fedor Indutny authored
`x86masm.pl` was mistakenly using .486 instruction set, why `cpuid` (and perhaps others) are requiring .686 . Fixes: https://github.com/iojs/io.js/issues/589 PR-URL: https://github.com/iojs/io.js/pull/1389 Reviewed-By:
Shigeki Ohtsu <ohtsu@iij.ad.jp>
-
Shigeki Ohtsu authored
All symlink files in `deps/openssl/openssl/include/openssl/` are removed and replaced with real header files to avoid issues on Windows. Two files of opensslconf.h in crypto and include dir are replaced to refer config/opensslconf.h.
-
Shigeki Ohtsu authored
This replaces all sources of openssl-1.0.2r.tar.gz into deps/openssl/openssl
-
- Feb 27, 2019
-
-
Matteo Collina authored
Fixes: https://github.com/nodejs-private/security/issues/214 PR-URL: https://github.com/nodejs-private/node-private/pull/162 Reviewed-By:
-
