socket.destroy() triggers a 'close' event from the socket which triggers
the onClose handler of HTTPAgent which calls self.removeSocket(). So by
calling self.removeSocket() prior to socket.destroy() we end up with two
calls to self.removeSocket().
If there are pending requests, removeSocket ends up creating a new socket.
So if there are pending requests, each time a request completes, we tear
down one socket and create two more. So the total number of sockets grows
exponentially and without regard for any maxSockets settings. This was
noticed in https://github.com/nodejs/node/issues/4050. Let's get rid of
the extra calls to removeSocket so we only call it once per completed
request.
PR-URL: https://github.com/nodejs/node/pull/4172
Reviewed-By: Brian White <mscdex@mscdex.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Fedor Indutny <fedor@indutny.com>
assert.deepEqual: when actual and expected are typed arrays,
wrap them in a new Buffer each to increase performance
significantly.
PR-URL: https://github.com/nodejs/node/pull/4330
Fixes: https://github.com/nodejs/node/issues/4294
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Roman Reiss <me@silverwind.io>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
On Windows there can exist some race condition where the
notification of the client's `socket.destroy()` isn't received
before the server writes to the socket. This race condition was
more evident/reproducible on a single core system.
This commit fixes the flakiness by waiting until the server's
connection event handler has been called to destroy the client
socket and perform the server socket write.
Fixes: https://github.com/nodejs/node/issues/4057
PR-URL: https://github.com/nodejs/node/pull/4342
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: João Reis <reis@janeasystems.com>
Load the certificate chain from the PFX file the same as we do it for a
regular certificate chain.
Fix: #4127
PR-URL: https://github.com/nodejs/node/pull/4165
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Improve performance of process._getActiveHandles by sending handles in
batches to JS to be set on the passed Array. Add test to check proper
active handles are returned.
Alter implementation of GetActiveRequests to match GetActiveHandles'
implementation.
PR-URL: https://github.com/nodejs/node/pull/3780
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Do not emit `upgrade` if the server is just advertising its protocols
support as per RFC 7230 Section 6.7.
A server MAY send an Upgrade header field in any other response
to advertise that it implements support for upgrading to the
listed protocols, in order of descending preference, when
appropriate for a future request.
Fix: https://github.com/nodejs/node/issues/4334
PR-URL: https://github.com/nodejs/node/pull/4337
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
This fixes breakage introduced in 94b9948d63 when writing the max
EventEmitter listeners warning to stderr.
PR-URL: https://github.com/nodejs/node/pull/4279
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Rewrite the test so that stderr reordering of the child processes won't
confuse the test's expectations.
PR-URL: https://github.com/nodejs/node/pull/4310
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
It does not currently have any explicit tests to verify the behavior.
PR-URL: https://github.com/nodejs/node/pull/4283
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
* Document that Symbol can used as event names.
* Add test for using Symbol as event names
PR-URL: https://github.com/nodejs/node/pull/4151
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
In dfee4e3712, the module wrapper
and line offset used when wrapping module code was changed to
better report errors on the first line of modules. However, that
commit did not update the runInThisContext() call used to
execute the core modules, so their error line numbers have been
off by one. This commit provides the correct lineOffset for core
modules.
Refs: https://github.com/nodejs/node/pull/2867
PR-URL: https://github.com/nodejs/node/pull/4254
Reviewed-By: Brian White <mscdex@mscdex.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
ARMv8 machines are typically quite fast and likely may not need
extended timeout times.
PR-URL: https://github.com/nodejs/node/pull/4248
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Rich Trott <rtrott@gmail.com>
Reviewed-By: Rod Vagg <rod@vagg.org>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Zero value of `maxCachedSessions` should disable TLS session caching in
`https.Agent`
PR-URL: https://github.com/nodejs/node/pull/4252
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Add `secureContext` option to `tls.connect`. It is useful for caching
client certificates, key, and CA certificates.
PR-URL: https://github.com/nodejs/node/pull/4246
Reviewed-By: James M Snell <jasnell@gmail.com>
Fix node exiting due to an exception being thrown rather than emitting
an `'uncaughtException'` event on the process object when:
1. no error handler is set on the domain within which an error is thrown
2. an `'uncaughtException'` event listener is set on the process
Also fix an issue where the process would not abort in the proper
function call if an error is thrown within a domain with no error
handler and `--abort-on-uncaught-exception` is used.
Finally, change the behavior of --abort-on-uncaught-exception so that,
if the domain within which the error is thrown has no error handler, but
a domain further up the domains stack has one, the process will not
abort.
Fixes#3607 and #3653.
PR: #3654
PR-URL: https://github.com/nodejs/node/pull/3654
Reviewed-By: Chris Dickinson <chris@neversaw.us>
When loading directory instead of file, no error message
is displayed. It's good to display error message for
this scenario.
PR-URL: https://github.com/nodejs/node/pull/4170
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
Add Buffer#includes() by wrapping an indexOf and performing a strict
equals check to -1.
The includes method takes the search value, byteOffset, and encoding as
arguments.
The test is a modified version of the indexOf test.
Fixes: https://github.com/nodejs/node/issues/3552
PR-URL: https://github.com/nodejs/node/pull/3567
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Reviewed-By: Evan Lucas <evanlucas@me.com>
Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
It can happen that the HTTP connection is closed before the server has received
all the requests, thus the server close condition is never reached. To solve
this, close the server when the socket is fully closed.
PR-URL: https://github.com/nodejs/node/pull/4041
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
Reviewed-By: Evan Lucas <evanlucas@me.com>
When sending a very large buffer (400000 bytes) the test fails due to
the client socket from the `a` server erroring with `ECONNRESET`.
There's a race condition between the closing of this socket and the `ssl`
socket closing on the other side of the connection. To improve things,
destroy the socket as soon as possible: in the `end` event of the `dest`
socket.
PR-URL: https://github.com/nodejs/node/pull/4195
Reviewed-By: Brian White <mscdex@mscdex.net>
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Make sure all the data is read before checking its validity.
Remove `gotHello` variable and just check that the ssl `end` event
is received.
Remove unused variables.
PR-URL: https://github.com/nodejs/node/pull/4195
Reviewed-By: Brian White <mscdex@mscdex.net>
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Fix module loading of third-party modules in the REPL by inheriting
module.paths from the REPL's parent module.
Commit ee72ee7 ("module,repl: remove repl require() hack") introduced
a regression where require() of modules in node_modules directories
no longer worked in the REPL (and fortunately only in the REPL.)
It turns out we didn't have test coverage for that but we do now.
Fixes: https://github.com/nodejs/node/issues/4208
PR-URL: https://github.com/nodejs/node/pull/4215
Reviewed-By: Roman Reiss <me@silverwind.io>
Before this commit you had to pass multiple CA certificates as an array
of strings. For convenience you can now pass them as a single string.
Fixes: https://github.com/nodejs/node/issues/4096
PR-URL: https://github.com/nodejs/node/pull/4099
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
This change cleans up outstanding comments on #3032. It improves error
handling when no isolate file is provided and adds the --prof-process
flag to the node binary which executes the tick processor on the
provided isolate file.
PR-URL: https://github.com/nodejs/node/pull/4021
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Evan Lucas <evanlucas@me.com>
Rename doNTCallback functions for clarity when profiling, these make
sense internally but the "NT" in particular is a bit obtuse to be
immediately understandable by non-core developers.
PR-URL: https://github.com/nodejs/node/pull/4167
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
These changes simplify using ECDH with private keys that are not
dynamically generated with ECDH.generateKeys.
Support for computing the public key corresponding to the given private
key was added. Validity checks to reduce the possibility of computing
a weak or invalid shared secret were also added.
Finally, ECDH.setPublicKey was softly deprecated.
PR-URL: https://github.com/nodejs/node/pull/3511
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Shigeki Ohtsu <ohtsu@iij.ad.jp>
If not running on Windows it skips the long path tests in:
* test-fs-long-path.js
* test-require-long-path.js
Fixes: https://github.com/nodejs/node/issues/2255
PR-URL: https://github.com/nodejs/node/pull/4116
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
As it can happen that the HTTP response is received in more than
one TCP chunk.
PR-URL: https://github.com/nodejs/node/pull/3961
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
Check if the worker 'isDead' instead of 'isConnected' as the
'disconnect' event is not guaranteed to be received before the
'exit' event.
Remove the 'net' dependency as it is not used.
PR-URL: https://github.com/nodejs/node/pull/3954
Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Michaël Zasso <mic.besace@gmail.com>
Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
If JS throws an object whose toString() method throws, then Node
attempts to print an empty message, but actually prints garbage.
This commit checks for this case, and prints a message instead.
Fixes: https://github.com/nodejs/node/issues/4079
PR-URL: https://github.com/nodejs/node/pull/4112
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
The modification time of a file is assumed to happen at the
exact time when it was requested. As the utime API specification
delcares that the resolution of the result is 1 second,
relax the constrain to 1 second helps the test case to be
robust and consistent under different load conditions in the system
PR-URL: https://github.com/nodejs/node/pull/3981
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Minwoo Jung <jmwsoft@gmail.com>
test-http-exit-delay was introduced to confirm the removal of a 1 second
delay that can occur when exiting node after an http request.
This change refactors the test for simplicity and also in the hopes of
either eliminating flakiness on CI or, if not that, at least making
the source of the flakiness easier to track down.
Ref: https://github.com/nodejs/node/commit/16b59cbc
Fixes: https://github.com/nodejs/node/issues/4045
PR-URL: https://github.com/nodejs/node/pull/4055
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: James M Snell <jasnell@gmail.com>
test-net-socket-local-address had a race condition that resulted in
unreliability on FreeBSD and Windows. This changes fixes the issue.
Fixes: https://github.com/nodejs/node/issues/2475
PR-URL: https://github.com/nodejs/node/pull/4109
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Evan Lucas <evanlucas@me.com>
Reviewed-By: Johan Bergström <bugs@bergstroem.nu>
The current implementation of tests for strings with length at or
exceeding kStringMaxLength allocate a temporary buffer inside a try
block to skip the test if there is insufficient memory. This commit adds
an invocation of the garbage collector after the temporary buffer is
allocated so that memory is freed for later allocations.
Change the corresponding catch block to rethrow the original exception
instead of asserting the exception message to provide more information
about the exception.
Add an additional check before trying to allocate memory to immediately
skip the test on machines with insufficient total memory.
PR-URL: https://github.com/nodejs/node/pull/3697
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Michaël Zasso <mic.besace@gmail.com>
Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com>
These properties were initially used to determine stream status
back in node v0.8 and earlier. Since streams2 however, these
properties were *always* true, which can be misleading for
example if you are trying to immediately determine whether
a Writable stream is still writable or not (to avoid a "write after
end" exception).
PR-URL: https://github.com/nodejs/node/pull/4083
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Dave
Claudio Rodriguez
Brian White
Fedor Indutny
Trevor Norris
Ben Noordhuis
Xotic750
cjihrig
Evan Lucas
Bryan English
Jeremiah Senkpiel
Julien Gilli
Prince J Wesley
Alexander Martin
Santiago Gimeno
Rich Trott
Matt Loring
Sakthipriyan Vairamani
Rod Vagg
Michael Ruddy
Rafał Pocztarski
Gireesh Punathil
Ali Ijaz Sheikh
Minwoo Jung
Michael Cornacchia