When using require to load a native addon the path must be converted
into a long path, otherwise the addon will fail to be loaded on
windows if the path is longer than 260 characters.
PR-URL: https://github.com/nodejs/node/pull/2965
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Benjamin Gruenbaum <inglor@gmail.com>
Account pending response data to decide whether pause the socket or
not. Writable stream state is a not reliable measure, because it just
says how much data is pending on a **current** request, thus not helping
much with problem we are trying to solve here.
PR-URL: https://github.com/nodejs/node/pull/3128
PR-URL: https://github.com/nodejs/node/pull/2411
Reviewed-By: Rod Vagg <rod@vagg.org>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Evan Lucas <evanlucas@me.com>
If the constructor can't assign a class id then the heap snapshot will
not be able to report the object. So ensure that all AsyncWrap instances
use a FunctionTemplate instance with an internal field count >= 1.
PR-URL: https://github.com/nodejs/node/pull/3139
Reviewed-By: Fedor Indutny <fedor.indutny@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-by: Stephen Belanger <admin@stephenbelanger.com>
Use V8's builtin ObjectIsPromise() to check that the value is a promise
before creating the promise mirror. Reduces garbage collector strain
in the (common) non-promise case, which is beneficial when inspecting
deep object graphs.
PR-URL: https://github.com/nodejs/node/pull/3130
Reviewed-By: Evan Lucas <evanlucas@me.com>
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Bye bye, smalloc. I'm not sure why this was still here; it was removed
in 70d1f32 and hasn't worked since. It wasn't packaged in the
binary, either.
Reviewed-By: Michaël Zasso <mic.besace@gmail.com>
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Reviewed-By: Сковорода Никита Андреевич <chalkerx@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
PR-URL: https://github.com/nodejs/node/pull/3099
Adds the documented but missing DNS error exports.BADNAME. This export
has been there before but got lost in a 2012 commit that added more
error codes. https://github.com/nodejs/node/pull/3076 will remove the
wrong error code exports.ADNAME.
PR-URL: https://github.com/nodejs/node/pull/3051
Fixes: https://github.com/nodejs/node/issues/3050
Reviewed-By: Сковорода Никита Андреевич <chalkerx@gmail.com>
Reviewed-By: Michaël Zasso <mic.besace@gmail.com>
Reviewed-By: Evan Lucas <evanlucas@me.com>
Ref: https://github.com/nodejs/node-convergence-archive/issues/13
This adds a new check for header and trailer fields names and method
names to ensure that they conform to the HTTP token rule. If they do
not, a `TypeError` is thrown.
Previously this had an additional `strictMode` option that has been
removed in favor of making the strict check the default (and only)
behavior.
Doc and test case are included.
On the client-side
```javascript
var http = require('http');
var url = require('url');
var p = url.parse('http://localhost:8888');
p.headers = {'testing 123': 123};
http.client(p, function(res) { }); // throws
```
On the server-side
```javascript
var http = require('http');
var server = http.createServer(function(req,res) {
res.setHeader('testing 123', 123); // throws
res.end('...');
});
```
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Reviewed-By: Trevor Norris <trevnorris@nodejs.org>
Reviewed-By: Сковорода Никита Андреевич <chalkerx@gmail.com>
Reviewed-By: Rod Vagg <rod@vagg.org>
PR-URL: https://github.com/nodejs/node/pull/2526
Instead of slicing the first element off of the matches, shift and then
return. This improves performance of the following path functions:
- basename: 18-20%
- extname: 60-70%
- dirname: 18-20%
- parse: 20-25%
PR-URL: https://github.com/nodejs/node/pull/3034
Reviewed-By: Brian White <mscdex@mscdex.net>
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
`HandleWrap::OnClose` destroys the underlying C++ object and null's the
internal field pointer to it. Therefore there should be no references to
the wrapping JavaScript object.
`null` the process' `_channel` field right after closing it, to ensure
no crashes will happen.
Fix: https://github.com/nodejs/node/issues/2847
PR-URL: https://github.com/nodejs/node/pull/3041
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
The `events` module already exports `EventEmitter` constructor function
So, we don't have to use `events.EventEmitter` to access it.
Refer: https://github.com/nodejs/node/pull/2896
PR-URL: https://github.com/nodejs/node/pull/2921
Reviewed-By: Roman Reiss <me@silverwind.io>
Reviewed-By: Michaël Zasso <mic.besace@gmail.com>
Logic for calling the passed in socket method and/or callback
was duplicated. This commit refactors the relevant code to
remove the redundancy.
PR-URL: https://github.com/nodejs/node/pull/2769
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Brendan Ashworth <brendan.ashworth@me.com>
The actual problem was with the line parsing logic for string literals.
When we use backslash in the string literals, it used to remember the
`\` as the previous character even after we parsed the character next
to it. This leads to REPL thinking that the end of string literals is
not reached.
This patch replaces the previous character with `null`, so that it will
properly skip the character next to it.
Previous Discussion: https://github.com/nodejs/node/pull/2952
Fixes: https://github.com/nodejs/node/issues/2749
PR-URL: https://github.com/nodejs/node/pull/2968
Reviewed-By: Roman Reiss <me@silverwind.io>
Commit 3c293ba ("http: protect against response splitting attacks")
filters out newline characters from HTTP headers but forgot to apply
the same logic to trailing HTTP headers, i.e., headers that come after
the response body. This commit rectifies that.
The expected security impact is low because approximately no one uses
trailing headers. Some HTTP clients can't even parse them.
PR-URL: https://github.com/nodejs/node/pull/2945
Reviewed-By: Сковорода Никита Андреевич <chalkerx@gmail.com>
Reviewed-By: Rod Vagg <r@va.gg>
Instantiating a Buffer of length zero would set the kNoZeroFill flag to
true but never actually call ArrayBuffer::Allocator(). Which means the
flag was never set back to false. The result was that the next
allocation would unconditionally not be zero filled.
Add test to ensure Uint8Array's are zero-filled after creating a Buffer
of length zero. This test may falsely succeed, but will not falsely fail.
Fix: https://github.com/nodejs/node/issues/2930
PR-URL: https://github.com/nodejs/node/pull/2931
Reviewed-By: Rod Vagg <rod@vagg.org>
As `raw` and `raws` encodings are deprecated for such a long time, and
they both are undocumented, this patch removes the support for those
encodings completely.
Previous discussion: https://github.com/nodejs/node/pull/2829
PR-URL: https://github.com/nodejs/node/pull/2859
Reviewed-By: Brian White <mscdex@mscdex.net>
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Most calls to ref() and unref() are chainable, timers should be
chainable, too.
Typical use:
var to = setTimeout(ontimeout, 123).unref();
PR-URL: https://github.com/nodejs/node/pull/2905
Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Trevor Norris <trevnorris@nodejs.org>
Overall construction time of Typed Arrays is faster in JS, but the
problem with using it normally is zero-fill of memory. Get around this
by using a flag in the ArrayBuffer::Allocator to trigger when memory
should or shouldn't be zero-filled.
Remove Buffer::Create() as it is no longer called.
The creation of the Uint8Array() was done at each callsite because at
the time of this patch there was a performance penalty for centralizing
the call in a single function.
PR-URL: https://github.com/nodejs/node/pull/2866
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Refactor out the if/else statement checking for option.host.
Add whitespace to make concatenation chunks more readable and
consistent with the https version of Agent.getName().
PR-URL: https://github.com/nodejs/node/pull/2825
Reviewed-By: Julian Duque <julianduquej@gmail.com>
Reviewed-By: Rich Trott <rtrott@gmail.com>
This commit refactors LazyTransform from the crypto implementation
(lib/crypto.js) into an internal module (not publicy accessible) in
internal/streams/lazy_transform.js. This promotes a more modular core
design and removes code bloat in crypto, as LazyTransform didn't
specifically have anything to do with cryptography, but rather a fast
way to support two APIs on a stream.
PR-URL: https://github.com/nodejs/node/pull/2566
Reviewed-By: Fedor Indutny <fedor.indutny@gmail.com>
Socket resume may happen on a next tick, and in following scenario:
1. `socket.resume()`
2. `socket._handle.close()`
3. `socket._handle = null;`
The `_resume` will be invoked with empty `._handle` property. There is
nothing bad about it, and we should just ignore the `resume`/`pause`
events in this case.
Same applies to the unconsuming of socket on adding `data` and/or
`readable` event listeners.
Fix: https://github.com/nodejs/node/issues/2821
PR-URL: https://github.com/nodejs/node/pull/2824
Reviewed-By: Сковорода Никита Андреевич <chalkerx@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com>
Streams with writev allow many buffers to be pushed to underlying OS
APIs in one batch, in this case improving write throughput by an order
of magnitude. This is especially noticeable when writing many (small)
buffers.
PR-URL: https://github.com/nodejs/node/pull/2167
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Previously, in _validateStdio we were using stdio.fd || stdio. If
stdio.fd was falsy (or 0 in the case of stdin), then the entire stdio
object would be passed which could cause a crash.
Fixes: https://github.com/nodejs/node/issues/2721
PR-URL: https://github.com/nodejs/node/pull/2727
Reviewed-By: silverwind - Roman Reiss <me@silverwind.io>
Reviewed-By: cjihrig - Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: indutny - Fedor Indutny <fedor.indutny@gmail.com>
Fixes a regression that appeared with the new Buffer implementation in v3.
Without this change, calling the SlowBuffer constructor with something else
than a number would abort on the C++ side. This makes sure that the length
argument is coerced to number or is 0.
Fixes: https://github.com/nodejs/node/issues/2634
PR-URL: https://github.com/nodejs/node/pull/2635
Reviewed-By: trevnorris - Trevor Norris <trev.norris@gmail.com>
Reviewed-By: cjihrig - Colin Ihrig <cjihrig@gmail.com>
Now parts of our public and public-ish APIs fall back to old-style
listenerCount() if the emitter does not have a listenerCount function.
Fixes: https://github.com/nodejs/node/issues/2655
Refs: 8f58fb92ff
PR-URL: https://github.com/nodejs/node/pull/2661
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Port of joyent/node commits:
* e17c5a72b2
* 70dafa7b62
Pull over test-child-process-spawn-typeerror.js from v0.12, replacing
the existing test in master. The new test includes a broader set of
tests on the various arg choices and throws.
Reviewed-By: trevnorris - Trevor Norris <trevnorris@nodejs.org>
Reviewed-By: cjihrig - Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: thefourtheye - Sakthipriyan Vairamani
PR-URL: https://github.com/nodejs/node/pull/2667
Fixes: https://github.com/nodejs/node/issues/2515
Add an optional callback parameter to `ChildProcess.prototype.send()`
that is invoked when the message has been sent.
Juggle the control channel's reference count so that in-flight messages
keep the event loop (and therefore the process) alive until they have
been sent.
`ChildProcess.prototype.send()` and `process.send()` used to operate
synchronously but became asynchronous in commit libuv/libuv@393c1c5
("unix: set non-block mode in uv_{pipe,tcp,udp}_open"), which landed
in io.js in commit 07bd05b ("deps: update libuv to 1.2.1").
Fixes: https://github.com/nodejs/node/issues/760
PR-URL: https://github.com/nodejs/node/pull/2620
Reviewed-By: trevnorris - Trevor Norris <trev.norris@gmail.com>
Reviewed-By: jasnell - James M Snell <jasnell@gmail.com>
v8 will silently return an empty handle
which doesn't delete our data if string length is
above String::kMaxLength
Fixes: https://github.com/nodejs/node/issues/1374
PR-URL: https://github.com/nodejs/node/pull/2402
Reviewed-By: trevnorris - Trevor Norris <trev.norris@gmail.com>
Reviewed-By: indutny - Fedor Indutny <fedor.indutny@gmail.com>
Reviewed-By: bnoordhuis - Ben Noordhuis <info@bnoordhuis.nl>
Amended by @rvagg to change author date from
"1970-08-16 16:09:02 +0200"
to
"2015-08-16 16:09:02 +0200"
as per discussion @ https://github.com/nodejs/node/issues/2713
This commit addresses most of the review comments in
https://github.com/nodejs/node/pull/2540, which are kept in this
separate commit so as to better preserve the prior two patches as they
landed in 0.12.
This commit:
- Fixes a bug with unrefActive timers and disposed domains.
- Fixes a bug with unrolling an unrefActive timer from another.
- Adds a test for both above bugs.
- Improves check logic, making it stricter, simpler, or both.
- Optimizes nicer with a smaller, separate function for the try/catch.
Fixes: https://github.com/nodejs/node-convergence-archive/issues/23
Ref: https://github.com/nodejs/node/issues/268
PR-URL: https://github.com/nodejs/node/pull/2540
Reviewed-By: bnoordhuis - Ben Noordhuis <info@bnoordhuis.nl>
Before this change, _unrefActive would keep the unrefList sorted when
adding a new timer.
Because _unrefActive is called extremely frequently, this linear scan
(O(n) at worse) would make _unrefActive show high in the list of
contributors when profiling CPU usage.
This commit changes _unrefActive so that it doesn't try to keep the
unrefList sorted. The insertion thus happens in constant time.
However, when a timer expires, unrefTimeout has to go through the whole
unrefList because it's not ordered anymore.
It is usually not large enough to have a significant impact on
performance because:
- Most of the time, the timers will be removed before unrefTimeout is
called because their users (sockets mainly) cancel them when an I/O
operation takes place.
- If they're not, it means that some I/O took a long time to happen, and
the initiator of subsequents I/O operations that would add more timers
has to wait for them to complete.
With this change, _unrefActive does not show as a significant
contributor in CPU profiling reports anymore.
Fixes: https://github.com/joyent/node/issues/8160
Signed-off-by: Timothy J Fontaine <tjfontaine@gmail.com>
Conflicts:
lib/timers.js
Fixes: https://github.com/nodejs/node-convergence-archive/issues/23
Ref: https://github.com/nodejs/node/issues/268
PR-URL: https://github.com/nodejs/node/pull/2540
Reviewed-By: bnoordhuis - Ben Noordhuis <info@bnoordhuis.nl>
This reverts commit 6cd0e2664b.
This reverts commit 7a999a1376.
This reverts commit f337595441.
It turns out that on Windows, uv_pipe_getsockname() is a no-op for
client sockets. It slipped through testing because of a CI snafu.
PR-URL: https://github.com/nodejs/node/pull/2584
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Complements the existing net.Socket#remoteFamily property.
PR-URL: https://github.com/nodejs/node/pull/956
Reviewed-By: Sam Roberts <vieuxtech@gmail.com>
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
The implementation is a minor API change in that socket.address() now
returns a `{ address: '/path/to/socket' }` object, like it does for TCP
and UDP sockets. Before this commit, it returned `socket._pipeName`,
which is a string when present.
Change common.PIPE on Windows from '\\\\.\\pipe\\libuv-test' to
'\\\\?\\pipe\\libuv-test'. Windows converts the '.' to a '?' when
creating a named pipe, meaning that common.PIPE didn't match the
result from NtQueryInformationFile().
Fixes: https://github.com/nodejs/node/issues/954
PR-URL: https://github.com/nodejs/node/pull/956
Reviewed-By: Sam Roberts <vieuxtech@gmail.com>
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
There is no type-conversion to be done. Therefore, use the === operator.
PR-URL: https://github.com/nodejs/node/pull/2558
Reviewed-By: Rich Trott <rtrott@gmail.com>
Explicitly note that these modules should not be used outside of core.
PR-URL: https://github.com/nodejs/node/pull/2523
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Reviewed-By: Michaël Zasso <mic.besace@gmail.com>
roundUpToNextPowerOf2() does more than just rounding up to the next
power of two. Rename it to computeNewHighWaterMark().
PR-URL: https://github.com/nodejs/node/pull/2479
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Don't iterate over all 32 bits, use some hacker's delight bit twiddling
to compute the next power of two.
The logic can be reduced to `n = 1 << 32 - Math.clz32(n)` but then it
can't easily be backported to v2.x; Math.clz32() was added in V8 4.3.
PR-URL: https://github.com/nodejs/node/pull/2479
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
The high watermark is capped at 8 MB, not 128 MB like the comment
in lib/_stream_readable.js said.
PR-URL: https://github.com/nodejs/node/pull/2479
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
