Return the previous raw mode setting from the internal `_setRawMode`
so that is easier to reset it to its original state later.
PR-URL: https://github.com/nodejs/node/pull/6635
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
- Adds the `breakEvalOnSigint` option to `vm.runIn(This)Context`.
This uses a watchdog thread to wait for SIGINT and generally works
just like the existing `timeout` option.
- Adds a method to the existing timer-based watchdog to check if it
stopped regularly or by running into the timeout. This is used to
tell a SIGINT abort from a timer-based one.
- Adds (internal) `process._{start,stop}SigintWatchdog` methods to
start/stop the watchdog thread used by the above option manually.
This will be used in the REPL to set up SIGINT handling before
entering terminal raw mode, so that there is no time window in
which Ctrl+C fully aborts the process.
PR-URL: https://github.com/nodejs/node/pull/6635
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
read() performance is improved most by switching from an array to
a linked list for storing buffered data. However, other changes that
also contribute include: making some hot functions inlinable, faster
read() argument checking, and misc code rearrangement to avoid
unnecessary code execution.
PR-URL: https://github.com/nodejs/node/pull/7077
Reviewed-By: Calvin Metcalf <calvin.metcalf@gmail.com>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
This change introduces experimental v8-inspector support. This brings
the DevTools debug protocol allowing Node.js to be debugged with
Chrome DevTools native, or through other debuggers supporting that
protocol.
Partial WebSocket support, to the extent required by DevTools, is
included. This is derived from the implementation in Blink.
v8-inspector support can be disabled by the --without-inspector
configure flag.
PR-URL: https://github.com/nodejs/node/pull/6792
Reviewed-By: jasnell - James M Snell <jasnell@gmail.com>
Reviewed-By: addaleax - Anna Henningsen <anna@addaleax.net>
Reviewed-By: bnoordhuis - Ben Noordhuis <info@bnoordhuis.nl>
The require('constants') module is currently undocumented and mashes
together unrelated constants. This refactors the require('constants')
in favor of distinct os.constants, fs.constants, and crypto.constants
that are specific to the modules for which they are relevant. The
next step is to document those within the specific modules.
PR-URL: https://github.com/nodejs/node/pull/6534
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Robert Lindstaedt <robert.lindstaedt@gmail.com>
Conflicts:
doc/api/fs.md
Improves performance of allocating unsafe buffers, creating buffers from
an existing ArrayBuffer and creating .slice(...) from existing Buffer by
avoiding deoptimizing change of prototype after Uint8Array allocation
in favor of ES6 native subclassing.
This is done through an internal ES6 class that extends Uint8Array and
is used for allocations, but the regular Buffer function is exposed, so
calling Buffer(...) with or without `new` continues to work as usual
and prototype chains are also preserved.
Performance wins for .slice are +120% (2.2x), and, consequently, for
unsafe allocations up to +95% (1.9x) for small buffers, and for safe
allocations (zero-filled) up to +30% (1.3x).
PR-URL: https://github.com/nodejs/node/pull/7349
Ref: https://github.com/nodejs/node/pull/6893
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Сковорода Никита Андреевич <chalkerx@gmail.com>
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Guard against the call to write() inside pipe's ondata pushing more data
back onto the Readable, thus causing ondata to be called again.
This is fine but results in awaitDrain being increased more than once.
The problem with that is when the destination does drain, only a single
'drain' event is emitted, so awaitDrain in this case will never reach
zero and we end up with a permanently paused stream.
Fixes: https://github.com/nodejs/node/issues/7278
PR-URL: https://github.com/nodejs/node/pull/7292
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
When `maybeReadMore` kicks in on a first bytes of incoming data, the
`req.read(0)` will be invoked and the `req._consuming` will be set to
`true`. This seemingly harmless property leads to a dire consequences:
the server won't call `req._dump()` and the whole HTTP/1.1 pipeline will
hang (single connection).
PR-URL: https://github.com/nodejs/node/pull/7211
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
This commit optimizes checkInvalidHeaderChar() by unrolling the
character checking loop a bit.
Additionally, some changes to the benchmark runner are needed in
order for the included benchmark to be run correctly. Specifically,
the regexp used to parse `key=value` parameters contained a greedy
quantifier that was causing the `key` to match part of the `value`
if `value` contained an equals sign.
PR-URL: https://github.com/nodejs/node/pull/6570
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Сковорода Никита Андреевич <chalkerx@gmail.com>
This commit both makes checkIsHttpToken() inlinable and extracts
the character checking logic to a separate inlinable function so that
the main loop can be unrolled a bit.
PR-URL: https://github.com/nodejs/node/pull/6570
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Сковорода Никита Андреевич <chalkerx@gmail.com>
Reset the `readableState.awaitDrain` counter after manual calls to
`.resume()`.
What might happen otherwise is that a slow consumer at the end of the
pipe could end up stalling the piping in the following scenario:
1. The writable stream indicates that its buffer is full.
2. This leads the readable stream to `pause()` and increase its
`awaitDrain` counter, which will be decreased by the writable’s next
`drain` event.
3. Something calls `.resume()` manually.
4. The readable continues to pipe to the writable, but once again
the writable stream indicates that the buffer is full.
5. The `awaitDrain` counter is thus increased again, but since it has
now been increased twice for a single piping destination, the next
`drain` event will not be able to reset `awaitDrain` to zero.
6. The pipe is stalled and no data is passed along anymore.
The solution in this commit is to reset the `awaitDrain` counter to
zero when `resume()` is called.
Fixes: https://github.com/nodejs/node/issues/7159
PR-URL: https://github.com/nodejs/node/pull/7160
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Avoid sending messages if the IPC channel is already disconnected. It
avoids undesired errors when calling `process.disconnect` when there are
still pending IPC messages.
PR-URL: https://github.com/nodejs/node/pull/7132
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Remove obsolete `setTimeout()` introduced in 3148f1400. The fix for the
problem is in b266074347. (For the record, I mostly don't know what I'm
talking about here but am summarizing from an IRC #node-dev conversation
with @indutny on 04-Jun-2016.)
PR-URL: https://github.com/nodejs/node/pull/7154
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
When `free`ing the socket to be reused in keep-alive Agent wait for
both `prefinish` and `end` events. Otherwise the next request may be
written before the previous one has finished sending the body, leading
to a parser errors.
PR-URL: https://github.com/nodejs/node/pull/7149
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Illegal tokens are only recoverable in string literals, RegExp literals,
and block comments. If not in one of these constructs, immediately
return an error rather than giving the user false hope by giving them a
chance to try to recover.
PR-URL: https://github.com/nodejs/node/pull/7104
Fixes: https://github.com/nodejs/node/issues/3611
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
All the callback functions in `fs` module are supposed to be executed
with no context (`this` value should not be a valid object). But
`mkdtemp`'s callback will have the `FSReqWrap` object as the context.
Sample code to reproduce the problem
'use strict';
const fs = require('fs');
fs.mkdtemp('/tmp/abcd', null, function() {
console.log(this);
});
This would print
FSReqWrap { oncomplete: [Function] }
But that should have printed `null` and this patch fixes that.
PR-URL: https://github.com/nodejs/node/pull/7068
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
The command line flag `--debug-brk` was ignored when the `-e` flag was
also present. This change allows the flags to both be honored when they
are used in a single command line.
PR-URL: https://github.com/nodejs/node/pull/7089
Fixes: https://github.com/nodejs/node/issues/3589
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
OS X has a tiny 1kb hard-coded buffer size for stdout / stderr to
TTYs (terminals). Output larger than that causes chunking, which ends
up having some (very small but existent) delay past the first chunk.
That causes two problems:
1. When output is written to stdout and stderr at similar times, the
two can become mixed together (interleaved). This is especially
problematic when using control characters, such as \r. With
interleaving, chunked output will often have lines or characters erased
unintentionally, or in the wrong spots, leading to broken output.
CLI apps often extensively use such characters for things such as
progress bars.
2. Output can be lost if the process is exited before chunked writes
are finished flushing. This usually happens in applications that use
`process.exit()`, which isn't infrequent.
See https://github.com/nodejs/node/issues/6980 for more info.
This became an issue as result of the Libuv 1.9.0 upgrade. A fix to
an unrelated issue broke a hack previously required for the OS X
implementation. This resulted in an unexpected behavior change in node.
The 1.9.0 upgrade was done in c3cec1eefc,
which was included in v6.0.0.
Full details of the Libuv issue that induced this are at
https://github.com/nodejs/node/issues/6456#issuecomment-219974514
Refs: https://github.com/nodejs/node/pull/1771
Refs: https://github.com/nodejs/node/issues/6456
Refs: https://github.com/nodejs/node/pull/6773
Refs: https://github.com/nodejs/node/pull/6816
PR-URL: https://github.com/nodejs/node/pull/6895
Reviewed-By: Rod Vagg <rod@vagg.org>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Treat negative length arguments to `Buffer()`/`allocUnsafe()`
as if they were zero so the allocation does not affect the
pool’s offset.
Fixes: https://github.com/nodejs/node/issues/7047
PR-URL: https://github.com/nodejs/node/pull/7051
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Сковорода Никита Андреевич <chalkerx@gmail.com>
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Reviewed-By: Rod Vagg <rod@vagg.org>
There are several places in the cluster module where a version
of process.send() is called, but the result is swallowed. Most
of these cases are internal, but Worker.prototype.send(), which
is publicly documented, also suffers from this problem. This
commit exposes the return value to facilitate better error
handling, and bring Worker.prototype.send() into compliance
with the documentation.
PR-URL: https://github.com/nodejs/node/pull/6998
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ron Korving <ron@ronkorving.nl>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
When debug flags are passed to clustered applications, the debug
port is rewritten for each worker process to avoid collisions.
Prior to this commit, each debug flag would get a unique value.
This commit reworks the logic to assign the same port value to
all debug flags for a single worker.
PR-URL: https://github.com/nodejs/node/pull/7050
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
This commit provides a rewrite of StringDecoder that both improves
performance (for non-single-byte encodings) and understandability.
Additionally, StringDecoder instantiation performance has increased
considerably due to inlinability and more efficient encoding name
checking.
PR-URL: https://github.com/nodejs/node/pull/6777
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
It allows reopening a server after it has been closed.
Fixes: https://github.com/nodejs/node/issues/6693
PR-URL: https://github.com/nodejs/node/pull/6981
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Ron Korving <ron@ronkorving.nl>
Reviewed-By: James M Snell <jasnell@gmail.com>
This commit fix a possible crash situation in dgram send().
A crash is possible if an array is passed, and then altered after the
send call, as the call to libuv is wrapped in process.nextTick().
It also avoid sending an empty array to libuv by allocating an empty
buffer. It also does some cleanup inside send() to increase readability.
It removes test flakyness by use common.mustCall and
common.platformTimeout. Fixes situations were some events were not
asserted to be emitted.
Fixes: https://github.com/nodejs/node/issues/6616
PR-URL: https://github.com/nodejs/node/pull/6804
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Before this commit `node --debug-port=1234 debug t.js` ignored the
--debug-port= argument, binding to the default port 5858 instead,
making it impossible to debug more than one process on the same
machine that way.
This commit also reduces the number of places where the default port
is hard-coded by one.
Fixes: https://github.com/nodejs/node/issues/3345
PR-URL: https://github.com/nodejs/node/pull/3470
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Fedor Indutny <fedor.indutny@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
cluster's internal message handling includes a cache of callback
functions. Once the message for that callback is received, it is
removed from the cache. If, for any reason, the same message ID
is processed twice, the callback will be missing from the cache
and cluster will try to call undefined as a function. This commit
guards against this scenario.
Refs: https://github.com/nodejs/node/issues/6561
PR-URL: https://github.com/nodejs/node/pull/6902
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
When a worker is disconnecting, it shuts down all of the handles
it is waiting on. It is possible that a handle does not have an
owner, which causes a crash. This commit closes such handles
without accessing the missing owner.
Fixes: https://github.com/nodejs/node/issues/6561
PR-URL: https://github.com/nodejs/node/pull/6909
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Currently, if an IPC event handler throws an error, it can
cause the message to not be consumed, leading to messages piling
up. This commit causes IPC events to be emitted on the next tick,
allowing the channel's processing logic to move forward as
normal.
Fixes: https://github.com/nodejs/node/issues/6561
PR-URL: https://github.com/nodejs/node/pull/6909
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Don't cache the exported values of fully uninitialized builtins.
This works by adding an additional `loading` flag that is only
active during initial loading of an internal module and checking
that either the module is fully loaded or is in that state before
using its cached value.
This has the effect that builtins modules which could not be loaded
(e.g. because compilation failed due to missing stack space) can be
loaded at a later point.
Fixes: https://github.com/nodejs/node/issues/6899
PR-URL: https://github.com/nodejs/node/pull/6907
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
When the 'buffer' encoding is passed to spawnSync(), an exception
is thrown in Buffer's toString() method because 'buffer' is not
a valid encoding there. This commit special cases the 'buffer'
encoding.
Fixes: https://github.com/nodejs/node/issues/6930
PR-URL: https://github.com/nodejs/node/pull/6939
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
As it is, `fs.mkdtemp` crashes with a C++ assertion if the callback
function is not passed. This patch uses `maybeCallback` to create one,
if no callback function is passed.
PR-URL: https://github.com/nodejs/node/pull/6828
Reviewed-By: Brian White <mscdex@mscdex.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
This is purely cleanup and carries no visible behavioural changes.
Up to now, `this._closed` was used in zlib.js as a
synonym of `!this._handle`. This change makes this connection
explicit and removes the `_closed` property from zlib streams,
as the previous duplication has been the cause of subtle errors
like https://github.com/nodejs/node/issues/6034.
This also makes zlib errors lead to an explicit `_close()` call
rather than waiting for garbage collection to clean up the handle,
thus returning memory resources earlier in the case of an error.
Add a getter for `_closed` so that the property remains accessible
by legacy code.
PR-URL: https://github.com/nodejs/node/pull/6574
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Avoid using deprecated getter syntax plus other
miscellaneous updates.
PR-URL: https://github.com/nodejs/node/pull/6766
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Add the `--preserve-symlinks` flag. This makes the changes added
in #5950 conditional. By default the old behavior is used. With
the flag set, symlinks are preserved, switching to the new
behavior. This should be considered to be a temporary solution
until we figure out how to solve the symlinked peer dependency
problem in a more general way that does not break everything
else.
Additional test cases are included.
PR-URL: https://github.com/nodejs/node/pull/6537
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
James M Snell
Anna Henningsen
Mathias Bynens
Brian White
cjihrig
Pavel Feldman
Ingvar Stepanyan
David Halls
Fedor Indutny
Santiago Gimeno
Rich Trott
Sakthipriyan Vairamani
Jeremiah Senkpiel
Ben Noordhuis
Matt Lavin
Matteo Collina
Justin Sprigg
Dirceu Pereira Tiegs
Eugene Obrezkov
Blake Embrey