Trevor Norris [Thu, 14 Nov 2013 23:40:56 +0000 (15:40 -0800)]
buffer: convert values to uint, not int
In many cases values expected to be unsigned were converted to a signed
integer.
Also include some small code cleanup.
Trevor Norris [Thu, 14 Nov 2013 01:19:53 +0000 (17:19 -0800)]
src: make buffer size errors more explicit
Fixes #6490
Trevor Norris [Thu, 14 Nov 2013 00:27:20 +0000 (16:27 -0800)]
buffer: no warning when encoding isn't passed
Buffer#write() was showing the deprecation warning when only
buf.write('string') was passed. This is incorrect since the encoding is
always optional.
Argument order should follow:
Buffer#write(string[, offset[, length]][, encoding])
(yeah, not confusing at all)
Ben Noordhuis [Thu, 14 Nov 2013 00:59:01 +0000 (01:59 +0100)]
src: remove unused import in cares_wrap.cc
Trevor Norris [Wed, 13 Nov 2013 23:14:34 +0000 (15:14 -0800)]
doc: fix few smalloc entries for proper formatting
Trevor Norris [Wed, 13 Nov 2013 22:51:32 +0000 (14:51 -0800)]
smalloc: check if object has external memory
Add HasExternalData API to check if Object has externally allocated
memory, and accompanying tests.
Trevor Norris [Tue, 5 Nov 2013 00:24:36 +0000 (16:24 -0800)]
src: add comments about implicit dependencies
Trevor Norris [Mon, 4 Nov 2013 18:49:55 +0000 (10:49 -0800)]
base-object: add BaseObject
BaseObject is a class that just handles the Persistent handle attached
to the class instance.
This also removed WeakObject. Reordering the inheritance chain helps
prevent unneeded calls on instances that don't call MakeCallback.
Trevor Norris [Fri, 1 Nov 2013 21:45:46 +0000 (14:45 -0700)]
async-wrap: make typenames consistent
Trevor Norris [Tue, 5 Nov 2013 19:56:13 +0000 (11:56 -0800)]
tls_wrap: move members to initialization list
They didn't need to be defined in the body of the constructor.
Ben Noordhuis [Mon, 11 Nov 2013 10:31:36 +0000 (11:31 +0100)]
src: use Context::Scope objects in cares_wrap.cc
Enter the context explicitly, don't rely on the fact that there is a
Context::Scope a few stack frames below because it may be gone someday
Ben Noordhuis [Mon, 11 Nov 2013 09:53:00 +0000 (10:53 +0100)]
src: fix Context::Scope usage
env->context() may or may not create a new Local. It currently does
not but don't depend on that behavior, create a HandleScope first.
Ben Noordhuis [Mon, 11 Nov 2013 09:34:16 +0000 (10:34 +0100)]
crypto: fix up implicit HandleScope abuse
Don't depend on any HandleScope objects that may or may not have been
created in src/stream_wrap.cc, create them explicitly.
Ben Noordhuis [Mon, 11 Nov 2013 20:03:08 +0000 (21:03 +0100)]
src: remove global HandleScope
Make it more difficult to accidentally leak handles by removing the
top-level HandleScope. Now if there's no valid HandleScope now, V8
will complain and, in debug builds, abort.
Ben Noordhuis [Mon, 11 Nov 2013 21:02:03 +0000 (22:02 +0100)]
src: fix Environment::GetCurrent() usage
Create a HandleScope before calling the Environment::GetCurrent() that
takes a v8::Isolate* as an argument because it creates a handle with
the call to v8::Isolate::CurrentContext().
Trevor Norris [Mon, 11 Nov 2013 19:24:31 +0000 (11:24 -0800)]
node: add HandleScope to prevent memory leak
MakeCallback can only be called from a valid v8::HandleScope.
Fixed #6487
Ben Noordhuis [Sun, 10 Nov 2013 01:02:27 +0000 (02:02 +0100)]
v8: upgrade to 3.22.24
This commit removes the simple/test-event-emitter-memory-leak test for
being unreliable with the new garbage collector: the memory pressure
exerted by the test case is too low for the garbage collector to kick
in. It can be made to work again by limiting the heap size with the
--max_old_space_size=x flag but that won't be very reliable across
platforms and architectures.
Ben Noordhuis [Sat, 9 Nov 2013 22:46:05 +0000 (23:46 +0100)]
crypto: update root certificates
Update the list of root certificates in src/node_root_certs.h with
tools/mk-ca-bundle.pl and update src/node_crypto.cc to make use of
the new format.
Fixes #6013.
Ben Noordhuis [Sat, 9 Nov 2013 22:32:12 +0000 (23:32 +0100)]
tools: check in certdata.txt from mozilla NSS
This is a file from the NSS project containing root certificate data.
It can be downloaded from:
https://mxr.mozilla.org/mozilla/source/security/nss/lib/ckfw/builtins/certdata.txt?raw=1
Once downloaded, running `perl tools/mk-ca-bundle.pl` will update
src/node_root_certs.h. Commit, rebuild and go.
Ben Noordhuis [Sat, 9 Nov 2013 22:31:54 +0000 (23:31 +0100)]
tools: customize mk-ca-bundle.pl
Remove unneeded functionality and tweak the generated output so we
can #include it in C++ source code.
Ben Noordhuis [Sat, 9 Nov 2013 22:01:26 +0000 (23:01 +0100)]
tools: bundle mk-ca-bundle.pl from upstream curl
isaacs [Fri, 8 Nov 2013 21:41:44 +0000 (13:41 -0800)]
Merge remote-tracking branch 'ry/v0.10'
isaacs [Fri, 8 Nov 2013 21:41:21 +0000 (13:41 -0800)]
npm: Upgrade to 1.3.14
Ben Noordhuis [Fri, 8 Nov 2013 21:23:59 +0000 (22:23 +0100)]
build: make v8 use random seed for hash tables
Upstream V8 as of commit v8/v8@4bc70e8 uses a fixed seed of
314159265
for hash tables unless instructed otherwise. Tell V8 to keep using a
random seed.
Ben Noordhuis [Fri, 8 Nov 2013 20:44:32 +0000 (21:44 +0100)]
src: remove container_of, use CONTAINER_OF
CONTAINER_OF was introduced a while ago but was not used consistently
everywhere yet. This commit fixes that.
Why CONTAINER_OF instead of container_of? The former makes it crystal
clear that it's a macro, not a function.
Ben Noordhuis [Fri, 8 Nov 2013 16:47:39 +0000 (17:47 +0100)]
crypto: unbreak build with no-NPN shared openssl
Unbreak the build when linking against a shared version of OpenSSL that
doesn't support NPN (Next Protocol Negotiation.)
Fixes the following build error:
../src/node_crypto.cc:140: error: no member function
'AdvertiseNextProtoCallback' declared in
'node::crypto::SSLWrap<node::TLSCallbacks>'
../src/node_crypto.cc:147: error: no member function
'SelectNextProtoCallback' declared in
'node::crypto::SSLWrap<node::TLSCallbacks>'
Fedor Indutny [Thu, 15 Aug 2013 21:30:47 +0000 (01:30 +0400)]
crypto: try moving read head in Peek()
Otherwise it might get stall (`Peek()` will return zero-length chunk)
in following situation:
1. `Write(kBufferLength)`
2. `Read(kBufferLength)`
3. `Write(anything)`
4. `Peek()` => `len=0`
Fedor Indutny [Mon, 21 Oct 2013 12:26:29 +0000 (16:26 +0400)]
doc: encoding is ignored if input is a Buffer
NOTE: it wasn't in 0.8
fixes #6386
isaacs [Sun, 3 Nov 2013 22:32:03 +0000 (14:32 -0800)]
http: force socket encoding to be null
Otherwise the string triggers an assertion error in node_http_parser.c,
line 370:
assert(Buffer::HasInstance(args[0]) == true);
because the first argument is not a Buffer object.
Ben Noordhuis [Thu, 7 Nov 2013 11:34:22 +0000 (12:34 +0100)]
doc: cluster: bump stability level to 'unstable'
As discussed on the mailing list: the module will not go away but the
API will continue to receive updates as the need arises.
Link: https://groups.google.com/forum/#!topic/nodejs/uqyTcQfimAI
Message-ID: <
7384b30e-b64c-4086-b78f-
b5acca9842a9@googlegroups.com>
fengmk2 [Wed, 6 Nov 2013 11:58:03 +0000 (19:58 +0800)]
http: cleanup freeSockets when socket destroyed
If the socket was destroyed, we need to remove it from the agent's
`freeSockets` list, otherwise dead socket could be reused by new
request.
Jackson Tian [Sun, 1 Sep 2013 17:41:00 +0000 (01:41 +0800)]
benchmark: Correct the bufferSize to highWaterMark
The bufferSize has been removed. Use highWaterMark instead of.
Fedor Indutny [Tue, 5 Nov 2013 14:14:28 +0000 (18:14 +0400)]
tls: more accurate wrapping of connecting socket
When socket, passed in `tls.connect()` `options` argument is not yet
connected to the server, `_handle` gets assigned to a `net.Socket`,
instead of `TLSSocket`.
When socket is connecting to the remote server (i.e. not yet connected,
but already past dns resolve phase), derive `_connecting` property from
it, because otherwise `afterConnect()` will throw an assertion.
fix #6443
Ben Noordhuis [Tue, 5 Nov 2013 07:23:58 +0000 (08:23 +0100)]
fs: make fs.watch() non-recursive by default
Fixes a behavioral regression introduced in commit 691b9eb.
Ben Noordhuis [Mon, 4 Nov 2013 21:42:48 +0000 (22:42 +0100)]
src: fix whitespace/braces lint error from 962686b
Silly warning. One day I'll probably remove it altogether, see if I
don't.
Nathan Rajlich [Mon, 4 Nov 2013 17:39:29 +0000 (09:39 -0800)]
http: make DELETE requests set `req.method`
Fixes #6461.
Ben Noordhuis [Sun, 3 Nov 2013 20:00:37 +0000 (21:00 +0100)]
src: squelch unused variable warning
|i| and |j| arent't used when building without crypto support. Hat tip
to Brian White.
Rename |l| to |k| while we're here because it's quite hard to discern
from |i| or |j| with some fonts.
isaacs [Thu, 31 Oct 2013 20:35:54 +0000 (13:35 -0700)]
doc: Correct and add several items
Several names/urls were out of date, and some really
awesome stuff was missing.
Sam Roberts [Wed, 30 Oct 2013 23:19:17 +0000 (16:19 -0700)]
doc: document node signal handling
Partly lifted from uv.h, partly from observation of node.cc.
Timothy J Fontaine [Wed, 30 Oct 2013 16:09:04 +0000 (09:09 -0700)]
blog: Post for v0.11.8
Erik Dubbelboer [Mon, 14 Oct 2013 14:53:59 +0000 (16:53 +0200)]
tls: add ECDH ciphers support
Switch test fixtures to 1024 bit keys.
Trevor Norris [Fri, 1 Nov 2013 20:57:09 +0000 (13:57 -0700)]
test: fix async listener test for windows
Ben Noordhuis [Fri, 1 Nov 2013 14:44:43 +0000 (15:44 +0100)]
benchmark: add http chunked encoding benchmarks
Test both "Content-Length: xxx" and "Transfer-Encoding: chunked" responses in
the http/simple benchmark.
Trevor Norris [Wed, 30 Oct 2013 00:05:40 +0000 (17:05 -0700)]
doc: add docs for AsyncListeners
Documentation has been added on how to use the AsyncListener API.
Jacob Groundwater [Tue, 15 Oct 2013 20:50:00 +0000 (13:50 -0700)]
test: add additional async listener tests
Trevor Norris [Tue, 29 Oct 2013 23:35:32 +0000 (16:35 -0700)]
domain: use AsyncListener API
The domain module has been switched over to use the domain module API as
much as currently possible. There are still some hooks in the
EventEmitter, but hopefully we can remove those in the future.
Trevor Norris [Sat, 28 Sep 2013 00:04:30 +0000 (17:04 -0700)]
cares: add AsyncListener support
Trevor Norris [Tue, 1 Oct 2013 00:24:41 +0000 (17:24 -0700)]
crypto: convert pbkdf2_req to a class
pbkdf2_req has been renamed to PBKDF2Request and converted to a class.
It now uses AsyncWrap::MakeCallback.
Also includes, using env()->ondone_string() instead of "ondone" and
using malloc instead of new char[].
Trevor Norris [Sun, 29 Sep 2013 06:33:29 +0000 (23:33 -0700)]
crypto: convert RandomBytesRequest to a class
Since RandomBytesRequest makes a call to MakeCallback, needed it to be
a class so AsyncWrap could handle any async listeners.
Also added a simple test for an issue had during implementation where
the memory was being released and returned.
Trevor Norris [Mon, 7 Oct 2013 19:39:52 +0000 (12:39 -0700)]
async-wrap: add methods to udp/tcp/pipe/timers
Now it's possible to add/remove an async listener to an individual
handle created by UDP, TCP, Pipe or Timer.
Trevor Norris [Fri, 27 Sep 2013 17:30:02 +0000 (10:30 -0700)]
async-wrap: integrate with WeakObject
Making WeakObject inherit from AsyncWrap allows us to peak into almost
all the MakeCallback calls in Node internals.
Trevor Norris [Tue, 24 Sep 2013 21:12:11 +0000 (14:12 -0700)]
node: add AsyncListener support
AsyncListener is a JS API that works in tandem with the AsyncWrap class
to allow the user to be alerted to key events in the life cycle of an
asynchronous event. The AsyncWrap class has its own MakeCallback
implementation that core will be migrated to use, and uses state sharing
techniques to allow quicker communication between JS and C++ whether the
async event callbacks need to be called.
Fedor Indutny [Thu, 31 Oct 2013 19:39:43 +0000 (23:39 +0400)]
test: fix tls-client-reject after
ba7c9ce96
Ben Noordhuis [Mon, 28 Oct 2013 19:18:59 +0000 (20:18 +0100)]
build: use zero overhead systemtap probes
Profiling suggested that on Linux sometimes over 10% of CPU time was
being spent inside the systemtap probe entry points in the binding
layer, even when the process was not actively being traced with the
`stap` tool.
That's why this commit makes it possible to use the *_ENABLED() macros
and bail out early when we're not being traced, reducing the overhead
of unused probes to (almost) zero.
Said macros were already being generated by `dtrace -h` but were not
usable because they rely on external definitions. To remedy that, we
now generate the accompanying object files with `dtrace -G`.
This commit includes a change to libuv that has been landed upstream in
commit joyent/libuv@3c172ea.
Fedor Indutny [Mon, 28 Oct 2013 12:10:10 +0000 (16:10 +0400)]
tls: do not default to 'localhost' servername
Fedor Indutny [Mon, 28 Oct 2013 10:25:27 +0000 (14:25 +0400)]
tls: reuse hostname from underlying net.Socket
When `tls.connect()` is called with `socket` option, it should try to
reuse hostname previously passed to `net.connect()` and only after that
fall back to `'localhost'`.
fix #6409
Nick Simmons [Tue, 22 Oct 2013 02:08:28 +0000 (22:08 -0400)]
fs: add recursive subdirectory support to fs.watch
Currently fs.watch does not have an option to specify if a directory
should be recursively watched for events across all subdirectories.
Several file watcher APIs support this. FSEvents on OS X > 10.5 is
one example. libuv has added support for FSEvents, but fs.watch had
no way to specify that a recursive watch was required.
fs.watch now has an additional boolean option 'recursive'. When set
to true, and when supported, fs.watch will return notifications for
the entire directory tree hierarchy rooted at the specified path.
Timothy J Fontaine [Wed, 30 Oct 2013 15:55:06 +0000 (08:55 -0700)]
Now working on 0.11.9
Timothy J Fontaine [Wed, 30 Oct 2013 15:54:48 +0000 (08:54 -0700)]
Merge branch 'v0.11.8-release'
Timothy J Fontaine [Wed, 30 Oct 2013 00:08:07 +0000 (17:08 -0700)]
2013.10.30, Version 0.11.8 (Unstable)
* uv: Upgrade to v0.11.14
* v8: upgrade 3.21.18.3
* assert: indicate if exception message is generated (Glen Mailer)
* buffer: add buf.toArrayBuffer() API (Trevor Norris)
* cluster: fix premature 'disconnect' event (Ben Noordhuis)
* crypto: add SPKAC support (Jason Gerfen)
* debugger: count space for line numbers correctly (Alex Kocharin)
* debugger: make busy loops SIGUSR1-interruptible (Ben Noordhuis)
* debugger: repeat last command (Alex Kocharin)
* debugger: show current line, fix for #6150 (Alex Kocharin)
* dgram: send() can accept strings (Trevor Norris)
* dns: rename domain to hostname (Ben Noordhuis)
* dns: set hostname property on error object (Ben Noordhuis)
* dtrace, mdb_v8: support more string, frame types (Dave Pacheco)
* http: add statusMessage (Patrik Stutz)
* http: expose supported methods (Ben Noordhuis)
* http: provide backpressure for pipeline flood (isaacs)
* process: Add exitCode property (isaacs)
* tls: socket.renegotiate(options, callback) (Fedor Indutny)
* util: format as Error if instanceof Error (Rod Vagg)
Timothy J Fontaine [Tue, 29 Oct 2013 23:46:15 +0000 (16:46 -0700)]
fs_event_wrap: update to new libuv api
Timothy J Fontaine [Tue, 29 Oct 2013 23:33:17 +0000 (16:33 -0700)]
uv: Upgrade to v0.11.14
Trevor Norris [Tue, 29 Oct 2013 20:23:58 +0000 (13:23 -0700)]
src: shorten Object{Wrap,Unwrap}
Going back to the original names of Wrap/Unwrap now that most all the
class members that duplicate the name and functionality have been
removed.
Trevor Norris [Tue, 29 Oct 2013 20:09:52 +0000 (13:09 -0700)]
src: don't use class specific Unwrap methods
Instead use the template functions in util.h.
Trevor Norris [Tue, 29 Oct 2013 19:49:53 +0000 (12:49 -0700)]
src: don't use WeakObject::Unwrap
Switch out to use UnwrapObject from util.h.
Trevor Norris [Tue, 29 Oct 2013 19:27:24 +0000 (12:27 -0700)]
src: use function to get internal pointer
Remove the NODE_{WRAP,UNWRAP} macros and instead use template functions.
Trevor Norris [Tue, 27 Aug 2013 18:30:06 +0000 (11:30 -0700)]
node: don't share state with in_tick/last_threw
There was no need to share state between C++ and JS for these two
values. So they have been moved to their respective locations. This will
help performance only a tiny bit, but it does help code complexity much
more.
Scott Blomquist [Tue, 29 Oct 2013 19:49:41 +0000 (12:49 -0700)]
doc: Update documentation to reflect ObjectWrap changes
Scott Blomquist [Tue, 29 Oct 2013 18:43:31 +0000 (11:43 -0700)]
win: fix ObjectWrap for latest v8
We need to keep ObjectWrap around for module authors (we think), but
v8 3.21 broke node_object_wrap.h with respect to MSVC. Coincidentally,
we no longer use ObjectWrap at all in core, and native modules might
as well use their own entirely internal implementation if they need it.
Scott Blomquist [Thu, 24 Oct 2013 20:10:43 +0000 (13:10 -0700)]
src: Remove unused refs to node_object_wrap.h
Turns out that we don't use node_object_wrap.h any more in core,
and, with v8 3.21, it's breaking our Windows build. Removing refs
to it everywhere (and adding node.h in one case where it was the
only way node.h was being included), we have restored the Windows
build.
Thom Seddon [Fri, 4 Oct 2013 11:59:38 +0000 (12:59 +0100)]
src: accept passphrase when crypto signing with private key
Previous behaviour was to drop to an openssl prompt
("Enter PEM pass phrase:") when supplying a private key with a
passphrase. This change adds a fourth, optional, paramter that
will be used as the passphrase.
To include this parameter in a backwards compatible way it was
necessary to expose the previously undocumented (and unexposed)
feature of being able to explitly setting the output encoding.
Ben Noordhuis [Tue, 29 Oct 2013 09:31:14 +0000 (10:31 +0100)]
doc: dgram: reword dgram.Socket#send() docs
Make it clear that the address argument is not really optional and fix
some Engrish and long lines while we're here.
Fixes #6433.
Trevor Norris [Mon, 28 Oct 2013 21:58:37 +0000 (14:58 -0700)]
dgram: send() can accept strings
Strings passed to Socket#send() will be passed to Buffer and parsed as
UTF8.
isaacs [Mon, 28 Oct 2013 21:12:00 +0000 (14:12 -0700)]
Merge remote-tracking branch 'ry/v0.10'
isaacs [Mon, 28 Oct 2013 21:10:47 +0000 (14:10 -0700)]
npm@1.3.13
isaacs [Sat, 26 Oct 2013 01:56:25 +0000 (18:56 -0700)]
vm: Copy missing properties from context
This addresses a current shortcoming of the V8 SetNamedPropertyHandler
function.
It does not provide a way to intercept Object.defineProperty(..) calls.
As a result, these properties are not copied onto the contextified
sandbox when a new global property is added via either a function
declaration or a Object.defineProperty(global, ...) call.
Note that any function declarations or Object.defineProperty() globals
that are created asynchronously (in a setTimeout, callback, etc.) will
happen AFTER the call to copy properties, and thus not be caught.
The way to properly fix this is to add some sort of a
Object::SetNamedDefinePropertyHandler() function that takes a callback,
which receives the property name and property descriptor as arguments.
Luckily, such situations are rare, and asynchronously-added globals
weren't supported by Node's VM module until 0.12 anyway. But, this
should be fixed properly in V8, and this copy function should be removed
once there is a better way.
Fix #6416
Ben Noordhuis [Mon, 28 Oct 2013 12:44:41 +0000 (13:44 +0100)]
http: remove MethodToString()
The list of supported HTTP methods is available in JS land now so there
is no longer any need to pass a stringified version of the method to the
parser callback, it can look up the method name for itself.
Saves a call to v8::Eternal::Get() in the common case and a costly
v8::String::NewFromOneByte() in the uncommon case.
Ben Noordhuis [Mon, 28 Oct 2013 12:35:33 +0000 (13:35 +0100)]
http: expose supported methods
Expose the list of supported HTTP methods as a property on the 'http'
module object.
Fixes #6422.
Phillip Alexander [Sat, 26 Oct 2013 20:13:56 +0000 (13:13 -0700)]
doc: fs: clarify fs.symlink Windows specific args
Ryan Graham [Sat, 26 Oct 2013 05:03:02 +0000 (22:03 -0700)]
doc: improve module documentation
Expands on when to use module.exports vs. exports. This is a recurring
question on mailing list and continues to confuse new devs.
Sam Roberts [Fri, 25 Oct 2013 20:16:50 +0000 (13:16 -0700)]
doc: fix missing backtick in 2e16037
Timothy J Fontaine [Fri, 25 Oct 2013 18:26:05 +0000 (11:26 -0700)]
Merge remote-tracking branch 'upstream/v0.10'
Conflicts:
ChangeLog
deps/uv/ChangeLog
deps/uv/include/uv-darwin.h
deps/uv/src/unix/fsevents.c
deps/uv/src/unix/process.c
deps/uv/src/version.c
doc/api/addons.markdown
doc/api/cluster.markdown
doc/api/http.markdown
lib/http.js
lib/tls.js
src/node_crypto.cc
src/node_http_parser.cc
src/node_version.h
src/pipe_wrap.cc
src/v8abbr.h
src/v8ustack.d
test/simple/test-http-pipeline-flood.js
Brian White [Thu, 24 Oct 2013 02:43:23 +0000 (22:43 -0400)]
doc: streams: document default objectMode setting
Maxim Bogushevich [Tue, 22 Oct 2013 07:27:02 +0000 (17:27 +1000)]
debugger: Fix bug in sb() with unnamed script
setBreakpoint() cause error when unnamed script is loaded
isaacs [Thu, 24 Oct 2013 16:21:59 +0000 (09:21 -0700)]
npm@1.3.12
Ben Noordhuis [Wed, 23 Oct 2013 11:42:53 +0000 (13:42 +0200)]
build: fix test-gc weakref build rule
Make the build rule depend on the build artifact (weakref.node) itself
rather than the directory it's built in. Depending on the directory
means that a build failure won't trigger a rebuild on the next
invocation because the directory's timestamp has been updated.
This is a back-port of commit 1189571 from the master branch that
hopefully fixes the following CI error:
executing: make test/gc/node_modules/weak/build/
make: *** No rule to make target `test/gc/node_modules/weak/build/'.
Command exited with non-zero: make test/gc/node_modules/weak/build/
Build step 'Execute NodeJS script' marked build as failure
Zarko Stankovic [Thu, 12 Sep 2013 18:13:32 +0000 (20:13 +0200)]
doc: add nodejs.rs to the community page
isaacs [Wed, 23 Oct 2013 20:08:06 +0000 (13:08 -0700)]
http client: pull last chunk on socket close
When the socket closes, the client's http incoming message object was
emitting an 'aborted' event if it had not yet been ended.
However, it's possible, when a response is being repeatedly paused and
resumed (eg, if piped to a slow FS write stream), that there will be a
final chunk remaining in the js-land buffer when the socket is torn
down.
When that happens, the socketCloseListener function detects that we have
not yet reached the end of the response message data, and treats this as
an abrupt abort, immediately (and forcibly) ending the incoming message
data stream, and discarding that final chunk of data.
The result is that, for example, npm will have problems because tarballs
are missing a few bytes off the end, every time.
Closes GH-6402
Ben Noordhuis [Wed, 23 Oct 2013 11:45:57 +0000 (13:45 +0200)]
test: fix up weakref.cc after v8 api change
Ben Noordhuis [Wed, 23 Oct 2013 11:42:53 +0000 (13:42 +0200)]
build: fix test-gc weakref build rule
Make the build rule depend on the build artifact (weakref.node) itself
rather than the directory it's built in. Depending on the directory
means that a build failure won't trigger a rebuild on the next
invocation because the directory's timestamp has been updated.
Ben Noordhuis [Tue, 22 Oct 2013 22:17:45 +0000 (15:17 -0700)]
src: update after v8 api changes
Timothy J Fontaine [Tue, 22 Oct 2013 22:14:25 +0000 (15:14 -0700)]
v8: upgrade 3.21.18.3
Ben Noordhuis [Wed, 23 Oct 2013 11:25:35 +0000 (13:25 +0200)]
src: IsInt64() should return bool, not int
isaacs [Tue, 22 Oct 2013 17:56:03 +0000 (10:56 -0700)]
blog: HTTP server DoS vulnerability details
CVE-2013-4450
Timothy J Fontaine [Fri, 18 Oct 2013 22:46:02 +0000 (15:46 -0700)]
blog: Post for v0.10.21
Timothy J Fontaine [Fri, 18 Oct 2013 22:43:55 +0000 (15:43 -0700)]
blog: Post for v0.8.26
Timothy J Fontaine [Fri, 18 Oct 2013 22:40:32 +0000 (15:40 -0700)]
Now working on 0.10.22
Timothy J Fontaine [Fri, 18 Oct 2013 22:39:56 +0000 (15:39 -0700)]
Merge branch 'v0.10.21-release' into v0.10
Timothy J Fontaine [Fri, 18 Oct 2013 21:26:27 +0000 (14:26 -0700)]
2013.10.18, Version 0.10.21 (Stable)
* uv: Upgrade to v0.10.18
* crypto: clear errors from verify failure (Timothy J Fontaine)
* dtrace: interpret two byte strings (Dave Pacheco)
* fs: fix fs.truncate() file content zeroing bug (Ben Noordhuis)
* http: provide backpressure for pipeline flood (isaacs)
* tls: fix premature connection termination (Ben Noordhuis)
Timothy J Fontaine [Fri, 18 Oct 2013 21:14:21 +0000 (14:14 -0700)]
crypto: clear errors from verify failure
OpenSSL will push errors onto the stack when a verify fails, which can
disrupt TLS and other routines if we don't clear the error stack
Fixes #6304