isaacs [Thu, 28 Mar 2013 20:06:18 +0000 (13:06 -0700)]
Node working on 0.10.3
isaacs [Thu, 28 Mar 2013 20:05:37 +0000 (13:05 -0700)]
Merge branch 'v0.10.2-release' into v0.10
Ben Noordhuis [Thu, 28 Mar 2013 19:12:02 +0000 (20:12 +0100)]
src: tie process.versions.uv to uv_version_string()
isaacs [Thu, 28 Mar 2013 18:46:21 +0000 (11:46 -0700)]
2013.03.28, Version 0.10.2 (Stable)
* npm: Upgrade to 1.2.15
* uv: Upgrade to 0.10.3
* tls: handle SSL_ERROR_ZERO_RETURN (Fedor Indutny)
* tls: handle errors before calling C++ methods (Fedor Indutny)
* tls: remove harmful unnecessary bounds checking (Marcel Laverdet)
* crypto: make getCiphers() return non-SSL ciphers (Ben Noordhuis)
* crypto: check randomBytes() size argument (Ben Noordhuis)
* timers: do not calculate Timeout._when property (Alexey Kupershtokh)
* timers: fix off-by-one ms error (Alexey Kupershtokh)
* timers: handle signed int32 overflow in enroll() (Fedor Indutny)
* stream: Fix stall in Transform under very specific conditions (Gil Pedersen)
* stream: Handle late 'readable' event listeners (isaacs)
* stream: Fix early end in Writables on zero-length writes (isaacs)
* domain: fix domain callback from MakeCallback (Trevor Norris)
* child_process: don't emit same handle twice (Ben Noordhuis)
* child_process: fix sending utf-8 to child process (Ben Noordhuis)
Bert Belder [Thu, 28 Mar 2013 19:05:33 +0000 (20:05 +0100)]
deps: upgrade libuv to v0.10.3
isaacs [Thu, 28 Mar 2013 18:35:12 +0000 (11:35 -0700)]
npm: Upgrade to v1.2.15
Fedor Indutny [Thu, 28 Mar 2013 16:55:51 +0000 (20:55 +0400)]
tls: handle SSL_ERROR_ZERO_RETURN
see #5004
wicked [Thu, 25 Oct 2012 04:53:35 +0000 (11:53 +0700)]
setTimeout: do not calculate Timeout._when property
Dramatically improves Timer performance.
isaacs [Wed, 27 Mar 2013 05:43:53 +0000 (22:43 -0700)]
stream: Emit readable on ended streams via read(0)
cc: @mjijackson
isaacs [Tue, 26 Mar 2013 21:42:56 +0000 (14:42 -0700)]
stream: Handle late 'readable' event listeners
In cases where a stream may have data added to the read queue before the
user adds a 'readable' event, there is never any indication that it's
time to start reading.
True, there's already data there, which the user would get if they
checked However, as we use 'readable' event listening as the signal to
start the flow of data with a read(0) call internally, we ought to
trigger the same effect (ie, emitting a 'readable' event) even if the
'readable' listener is added after the first emission.
To avoid confusing weirdness, only the *first* 'readable' event listener
is granted this privileged status. After we've started the flow (or,
alerted the consumer that the flow has started) we don't need to start
it again. At that point, it's the consumer's responsibility to consume
the stream.
Closes #5141
isaacs [Thu, 28 Mar 2013 01:41:41 +0000 (18:41 -0700)]
doc: Add 'don't ignore errors' section to domain
Also, an example program of using cluster and domain to handle errors
safely, with zero downtime, using process isolation.
Benjamin Ruston [Thu, 28 Mar 2013 14:19:10 +0000 (14:19 +0000)]
doc: debugger, dns, http: fix grammar
Ben Noordhuis [Wed, 27 Mar 2013 23:28:45 +0000 (00:28 +0100)]
deps: upgrade libuv to
7514149
isaacs [Wed, 27 Mar 2013 16:47:16 +0000 (09:47 -0700)]
test: Accept either kind of NaN
A llvm/clang bug on Darwin ia32 makes these tests fail 100% of
the time. Since no one really seems to mind overly much, and we
can't reasonably fix this in node anyway, just accept both types
of NaN for now.
Fedor Indutny [Wed, 27 Mar 2013 08:18:43 +0000 (12:18 +0400)]
tls: handle errors before calling C++ methods
Calling `this.pair.encrypted._internallyPendingBytes()` before
handling/resetting error will result in assertion failure:
../src/node_crypto.cc:962: void node::crypto::Connection::ClearError():
Assertion `handle_->Get(String::New("error"))->BooleanValue() == false'
failed.
see #5058
Benjamin Ruston [Wed, 27 Mar 2013 10:09:37 +0000 (10:09 +0000)]
doc: addon: fix grammar
Fedor Indutny [Tue, 26 Mar 2013 18:33:14 +0000 (22:33 +0400)]
openssl: disable HEARTBEAT TLS extension
Microsoft's IIS doesn't support it, and is not replying with ServerHello
after receiving ClientHello which contains it.
The good way might be allowing to opt-out this at runtime from
javascript-land, but unfortunately OpenSSL doesn't support it right now.
see #5119
Trevor Norris [Tue, 26 Mar 2013 05:32:41 +0000 (22:32 -0700)]
domain: fix domain callback from MakeCallback
Since _tickCallback and _tickDomainCallback were both called from
MakeCallback, it was possible for a callback to be called that required
a domain directly to _tickCallback.
The fix was to implement process.usingDomains(). This will set all
applicable functions to their domain counterparts, and set a flag in cc
to let MakeCallback know domain callbacks always need to be checked.
Added test in own file. It's important that the test remains isolated.
Ben Noordhuis [Tue, 26 Mar 2013 15:34:42 +0000 (16:34 +0100)]
doc: child_process: document 'error' event
Fixes #5130.
Ben Noordhuis [Tue, 26 Mar 2013 15:09:48 +0000 (16:09 +0100)]
doc: fix formatting in tty.markdown
Fixes #5135.
Timothy J Fontaine [Mon, 25 Mar 2013 23:16:44 +0000 (16:16 -0700)]
test: test name is the last elem, not second
When a test requires node to have parameters passed (--expose-gc) the
test name will be the last element in the command array, not the second.
Ben Noordhuis [Sun, 24 Mar 2013 23:29:55 +0000 (00:29 +0100)]
child_process: don't emit same handle twice
It's possible to read multiple messages off the parent/child channel.
When that happens, make sure that recvHandle is cleared after emitting
the first message so it doesn't get emitted twice.
Ben Noordhuis [Mon, 18 Mar 2013 23:16:55 +0000 (00:16 +0100)]
crypto: make getCiphers() return non-SSL ciphers
Commit
f53441a added crypto.getCiphers() as a function that returns the
names of SSL ciphers.
Commit
14a6c4e then added crypto.getHashes(), which returns the names of
digest algorithms, but that creates a subtle inconsistency: the return
values of crypto.getHashes() are valid arguments to crypto.createHash()
but that is not true for crypto.getCiphers() - the returned values are
only valid for SSL/TLS functions.
Rectify that by adding tls.getCiphers() and making crypto.getCiphers()
return proper cipher names.
Mathias Bynens [Tue, 19 Mar 2013 08:10:23 +0000 (09:10 +0100)]
doc: mention `process.*.isTTY` under `process`
Ben Noordhuis [Thu, 14 Mar 2013 15:59:07 +0000 (16:59 +0100)]
child_process: fix sending utf-8 to child process
In process#send() and child_process.ChildProcess#send(), use 'utf8' as
the encoding instead of 'ascii' because 'ascii' mutilates non-ASCII
input. Correctly handle partial character sequences by introducing
a StringDecoder.
Sending over UTF-8 no longer works in v0.10 because the high bit of
each byte is now cleared when converting a Buffer to ASCII. See
commit
96a314b for details.
Fixes #4999 and #5011.
Ben Noordhuis [Thu, 21 Mar 2013 12:12:41 +0000 (13:12 +0100)]
bench: add child process read perf benchmark
Ben Noordhuis [Sun, 24 Mar 2013 21:53:07 +0000 (22:53 +0100)]
deps: fix openssl build on windows
Commit
8632af3 ("tools: update gyp to r1601") broke the Windows build.
Older versions of GYP link to kernel32.lib, user32.lib, etc. but that
was changed in r1584. See https://codereview.chromium.org/
12256017
Fix the build by explicitly linking to the required libraries.
isaacs [Fri, 22 Mar 2013 01:18:55 +0000 (18:18 -0700)]
stream: Fix early end in Writables on zero-length writes
Doing this causes problems:
z.write(Buffer(0));
z.end();
Fix by not ending Writable streams while they're still in the process of
writing something.
Ben Noordhuis [Sun, 24 Mar 2013 13:29:17 +0000 (14:29 +0100)]
tools: update gyp to r1601
Among other things, this should make it easier for people to build
node.js on openbsd.
Ben Noordhuis [Sun, 24 Mar 2013 12:22:01 +0000 (13:22 +0100)]
doc: update CONTRIBUTING.md
* Latest stable is v0.10 now.
* Add example of what the first line of the commit log should look like.
Alexey Kupershtokh [Fri, 26 Oct 2012 05:49:22 +0000 (12:49 +0700)]
timer: fix off-by-one ms error
Fix #5103
Marcel Laverdet [Thu, 21 Mar 2013 21:56:02 +0000 (16:56 -0500)]
tls: remove harmful unnecessary bounds checking
The EncIn, EncOut, ClearIn & ClearOut functions are victims of some code
copy + pasting. A common line copied to all of them is:
`if (off >= buffer_length) { ...`
448e0f43 corrected ClearIn's check from `>=` to `>`, but left the others
unchanged (with an incorrect bounds check). However, if you look down at
the next very next bounds check you'll see:
`if (off + len > buffer_length) { ...`
So the check is actually obviated by the next line, and should be
removed.
This fixes an issue where writing a zero-length buffer to an encrypted
pair's *encrypted* stream you would get a crash.
verwaest@chromium.org [Mon, 18 Mar 2013 13:35:17 +0000 (13:35 +0000)]
v8: Unify kMaxArguments with number of bits used to encode it.
Increase the number of bits by 1 by making Flags unsigned.
BUG=chromium:211741
Review URL: https://chromiumcodereview.appspot.com/
12886008
This is a back-port of commits 13964 and 13988 addressing CVE-2013-2632.
Ben Noordhuis [Sat, 23 Mar 2013 14:48:56 +0000 (15:48 +0100)]
crypto: check randomBytes() size argument
Throw a TypeError if size > 0x3fffffff. Avoids the following V8 fatal
error:
FATAL ERROR: v8::Object::SetIndexedPropertiesToExternalArrayData()
length exceeds max acceptable value
Fixes #5126.
Ben Noordhuis [Sat, 23 Mar 2013 14:38:17 +0000 (15:38 +0100)]
doc: document that stdio is usually blocking
Gil Pedersen [Sun, 17 Mar 2013 14:04:01 +0000 (15:04 +0100)]
stream: Fix stall in Transform under very specific conditions
The stall is exposed in the test, though the test itself asserts before
it stalls.
The test is constructed to replicate the stalling state of a complex
Passthrough usecase since I was not able to reliable trigger the stall.
Some of the preconditions for triggering the stall are:
* rs.length >= rs.highWaterMark
* !rs.needReadable
* _transform() handler that can return empty transforms
* multiple sync write() calls
Combined this can trigger a case where rs.reading is not cleared when
further progress requires this. The fix is to always clear rs.reading.
Fedor Indutny [Thu, 21 Mar 2013 17:38:30 +0000 (21:38 +0400)]
timers: handle signed int32 overflow in enroll()
Before this patch calling `socket.setTimeout(0xffffffff)` will result in
signed int32 overflow in C++ which resulted in assertion error:
Assertion failed: (timeout >= -1), function uv__io_poll, file
../deps/uv/src/unix/kqueue.c, line 121.
see #5101
isaacs [Thu, 21 Mar 2013 16:14:01 +0000 (09:14 -0700)]
blog: Post for v0.10.1
isaacs [Thu, 21 Mar 2013 16:14:23 +0000 (09:14 -0700)]
Now working on v0.10.2
isaacs [Thu, 21 Mar 2013 16:13:45 +0000 (09:13 -0700)]
Merge branch 'v0.10.1-release' into v0.10
Ben Noordhuis [Thu, 21 Mar 2013 13:59:16 +0000 (14:59 +0100)]
deps: upgrade libuv to
9b61939
isaacs [Thu, 21 Mar 2013 01:15:04 +0000 (18:15 -0700)]
2013.03.21, Version 0.10.1 (Stable)
* npm: upgrade to 1.2.15
* crypto: Improve performance of non-stream APIs (Fedor Indutny)
* tls: always reset this.ssl.error after handling (Fedor Indutny)
* tls: Prevent mid-stream hangs (Fedor Indutny, isaacs)
* net: improve arbitrary tcp socket support (Ben Noordhuis)
* net: handle 'finish' event only after 'connect' (Fedor Indutny)
* http: Don't hot-path end() for large buffers (isaacs)
* fs: Missing cb errors are deprecated, not a throw (isaacs)
* fs: make write/appendFileSync correctly set file mode (Raymond Feng)
* stream: Return self from readable.wrap (isaacs)
* stream: Never call decoder.end() multiple times (Gil Pedersen)
* windows: enable watching signals with process.on('SIGXYZ') (Bert Belder)
* node: revert removal of MakeCallback (Trevor Norris)
* node: Unwrap without aborting in handle fd getter (isaacs)
isaacs [Thu, 21 Mar 2013 00:49:57 +0000 (17:49 -0700)]
npm: upgrade to 1.2.15
Fedor Indutny [Fri, 15 Mar 2013 08:59:30 +0000 (12:59 +0400)]
crypto: initialize transform lazily
isaacs [Wed, 20 Mar 2013 23:14:36 +0000 (16:14 -0700)]
tls: Prevent hang in readStart
This is not a great fix, and it's a bug that's very tricky to reproduce.
Occasionally, while downloading a file, especially on Linux for some
reason, the pause/resume timing will be just right such that the
CryptoStream is in a 'reading' state, but actually has no data, so it
ought to pull more in. Because there's no reads happening, it just sits
there, and the process will exit
This is, fundamentally, a factor of how the HTTP implementation sits
atop CryptoStreams and TCP Socket objects, which is utterly horrible,
and needs to be rewritten. However, in the meantime, npm downloads are
prematurely exiting, causing hard-to-debug "cb() never called!" errors.
Trevor Norris [Wed, 20 Mar 2013 08:58:11 +0000 (01:58 -0700)]
bench: compare binaries equal times
The benchmark compare would drop the last run of the binary pairs. So
when they were only run once an error would arise because no data was
generated for the second binary.
Ben Noordhuis [Wed, 20 Mar 2013 15:53:23 +0000 (16:53 +0100)]
bench: add dgram send/recv benchmark
Fedor Indutny [Wed, 20 Mar 2013 13:35:38 +0000 (17:35 +0400)]
tls: always reset this.ssl.error after handling
Otherwise assertion may happen:
src/node_crypto.cc:962: void node::crypto::Connection::ClearError():
Assertion `handle_->Get(String::New("error"))->BooleanValue() == false'
failed.
See #5058
Raymond Feng [Wed, 20 Mar 2013 00:37:03 +0000 (01:37 +0100)]
fs: make write/appendFileSync correctly set file mode
Iskren Ivov Chernev [Tue, 19 Mar 2013 17:58:18 +0000 (10:58 -0700)]
doc: fix streams2 SimpleProtocol example
A non-existing variable `b` was used to queue data for reading.
Bert Belder [Tue, 19 Mar 2013 23:16:02 +0000 (00:16 +0100)]
windows: enable watching signals with process.on('SIGXYZ')
This reverts commit
ea1cba6246a8b1784e22d076139b9244a9ff42f8.
The offending commit was intended to land on the v0.8 branch only, but
it accidentally got merged at some point.
Closes #5054.
Bert Belder [Tue, 19 Mar 2013 19:07:38 +0000 (20:07 +0100)]
Update .mailmap and AUTHORS
Ben Noordhuis [Mon, 18 Mar 2013 13:40:03 +0000 (14:40 +0100)]
doc: fix broken links in blog footer
The blog lives at blog.nodejs.org while the main website lives at
nodejs.org. Ergo, use absolute URLs for links to the main website.
Fixes #5062.
Fedor Indutny [Fri, 15 Mar 2013 18:49:31 +0000 (22:49 +0400)]
tls: write pending data of opposite side
Fix stucked CryptoStream behaviour, happening when one of the sides
locks-up in queued state.
fix #5023
Trevor Norris [Sun, 17 Mar 2013 04:59:47 +0000 (21:59 -0700)]
node: revert removal of MakeCallback
In
0168109 an implementation of MakeCallback was accidently removed. It
has been re-added.
JeongHoon Byun [Tue, 12 Mar 2013 12:21:43 +0000 (21:21 +0900)]
doc: fix typo in crypto docs
Yi EungJun [Fri, 15 Mar 2013 16:31:48 +0000 (01:31 +0900)]
doc: https: Fix the link to tls.connect
Ben Noordhuis [Sat, 16 Mar 2013 22:32:42 +0000 (23:32 +0100)]
deps: upgrade libuv to
b45a74f
Ben Noordhuis [Sat, 16 Mar 2013 22:18:50 +0000 (23:18 +0100)]
install: don't install man page twice
Looks like a merge conflict in
77ed12f left in the old, unconditional
install rule. Remove it, the new and improved rule is a few lines down.
Fixes #5044.
Timothy J Fontaine [Thu, 14 Mar 2013 23:06:59 +0000 (16:06 -0700)]
v8: move 32 bit heap hint on sunos
Setting the V8 heap at or near 0x20000000 on 32bit sunos only allows
512 MB of heap space, instead on sunos move this to 0x80000000.
Fixes #4010.
isaacs [Thu, 14 Mar 2013 23:43:19 +0000 (16:43 -0700)]
stream: Return self from readable.wrap
Also, set paused=false *before* calling resume(). Otherwise,
there's an edge case where an immediately-emitted chunk might make
it call pause() again incorrectly.
isaacs [Thu, 14 Mar 2013 23:18:42 +0000 (16:18 -0700)]
test: No need for kicking in streams2 test
This was necessary when we weren't auto-starting when a 'readable'
listener is added.
Gil Pedersen [Thu, 14 Mar 2013 13:01:14 +0000 (14:01 +0100)]
stream: Never call decoder.end() multiple times
Updated version that does what it says without assigning state.decoder.
koichik [Thu, 14 Mar 2013 14:57:13 +0000 (23:57 +0900)]
doc: don't mark fs callbacks as optional
Refs #5005, #5008
isaacs [Thu, 14 Mar 2013 14:48:18 +0000 (07:48 -0700)]
http: Don't hot-path end() for large buffers
The benefits of the hot-path optimization below start to fall off when
the buffer size gets up near 128KB, because the cost of the copy is more
than the cost of the extra write() call. Switch to the write/end method
at that point.
Heuristics and magic numbers are awful, but slow http responses are
worse.
Fix #4975
Ben Noordhuis [Thu, 14 Mar 2013 14:13:58 +0000 (15:13 +0100)]
net: improve arbitrary tcp socket support
Consider this example:
// fd 3 is a bound tcp socket
var s = net.createServer(cb);
s.listen({ fd: 3 });
console.log(s.address()); // prints null
This commit makes net.Server#address() print the actual address.
Ditto for non-listen sockets; properties like net.Socket#localAddress
and net.Socket#remoteAddress now return the correct value.
Fixes #5009.
Ben Noordhuis [Thu, 14 Mar 2013 12:40:27 +0000 (13:40 +0100)]
deps: upgrade libuv to
7b66ea1
Adam Malcontenti-Wilson [Thu, 14 Mar 2013 02:17:07 +0000 (13:17 +1100)]
doc: change dgram to socket for properties of dgram.Socket
Fixes #4919.
isaacs [Wed, 13 Mar 2013 22:48:56 +0000 (15:48 -0700)]
Revert "stream: Never call decoder.end() multiple times"
This reverts commit
615d809ac684a7d2cc7ee5e1aa58f0a921b529a0.
isaacs [Wed, 13 Mar 2013 21:59:42 +0000 (14:59 -0700)]
fs: Missing cb errors are deprecated, not a throw
Commit
a804347 makes fs function rethrow errors when the callback is
omitted. While the right thing to do, it's a change from the old v0.8
behavior where such errors were silently ignored.
To give users time to upgrade, temporarily disable that and replace it
with a function that warns once about the deprecated behavior.
Close #5005
Ben Noordhuis [Wed, 13 Mar 2013 21:52:33 +0000 (22:52 +0100)]
doc: implicit fs callbacks don't throw in v0.10
But they will in v0.12.
Re #5005.
Ben Noordhuis [Wed, 13 Mar 2013 22:22:55 +0000 (23:22 +0100)]
doc: add note on process.title max length
Fixes #5006.
Gil Pedersen [Tue, 12 Mar 2013 14:56:30 +0000 (15:56 +0100)]
stream: Never call decoder.end() multiple times
Fixes decoder.end() being called on every push(null). As the tls module
does this, corrupt stream data could potentially be added to the end.
isaacs [Wed, 13 Mar 2013 18:22:33 +0000 (11:22 -0700)]
core: Move UNWRAP_NO_ABORT to handle_wrap.h
Otherwise it cannot be used in StreamWrap.
Forgot to include in last patch, broke the build.
Ben Noordhuis [Wed, 13 Mar 2013 17:27:09 +0000 (18:27 +0100)]
doc: path.join() arguments must be strings
In v0.8, non-string arguments were ignored. v0.10 throws an exception.
Fedor Indutny [Wed, 13 Mar 2013 12:53:27 +0000 (16:53 +0400)]
net: handle 'finish' event only after 'connect'
isaacs [Wed, 13 Mar 2013 17:16:26 +0000 (10:16 -0700)]
core: Unwrap without aborting in handle fd getter
isaacs [Wed, 13 Mar 2013 17:15:30 +0000 (10:15 -0700)]
blog: Fix typo in typo fix
Nathan Rajlich [Wed, 13 Mar 2013 02:00:48 +0000 (19:00 -0700)]
blog: fix small typo in v0.10.0 release article
Ben Noordhuis [Mon, 11 Mar 2013 23:03:56 +0000 (00:03 +0100)]
doc: events: add 'removeListener' event section
Amends commit
84221fd by (also) documenting the 'removeListener' event
in a dedicated section, like the 'newListener' event.
Fixes #4977.
isaacs [Mon, 11 Mar 2013 15:56:47 +0000 (08:56 -0700)]
blog: Post about v0.10.0
isaacs [Mon, 11 Mar 2013 15:49:33 +0000 (08:49 -0700)]
Now working on 0.10.1
isaacs [Mon, 11 Mar 2013 15:49:20 +0000 (08:49 -0700)]
Merge branch 'v0.10.0-release' into v0.10
isaacs [Sat, 9 Mar 2013 17:24:56 +0000 (09:24 -0800)]
2013.03.11, Version 0.10.0 (Stable)
* npm: Upgrade to 1.2.14
* core: Append filename properly in dlopen on windows (isaacs)
* zlib: Manage flush flags appropriately (isaacs)
* domains: Handle errors thrown in nested error handlers (isaacs)
* buffer: Strip high bits when converting to ascii (Ben Noordhuis)
* win/msi: Enable modify and repair (Bert Belder)
* win/msi: Add feature selection for various node parts (Bert Belder)
* win/msi: use consistent registry key paths (Bert Belder)
* child_process: support sending dgram socket (Andreas Madsen)
* fs: Raise EISDIR on Windows when calling fs.read/write on a dir (isaacs)
* unix: fix strict aliasing warnings, macro-ify functions (Ben Noordhuis)
* unix: honor UV_THREADPOOL_SIZE environment var (Ben Noordhuis)
* win/tty: fix typo in color attributes enumeration (Bert Belder)
* win/tty: don't touch insert mode or quick edit mode (Bert Belder)
isaacs [Sun, 10 Mar 2013 02:05:39 +0000 (18:05 -0800)]
stream: Don't emit 'end' unless read() called
This solves the problem of calling `readable.pipe(writable)` after the
readable stream has already emitted 'end', as often is the case when
writing simple HTTP proxies.
The spirit of streams2 is that things will work properly, even if you
don't set them up right away on the first tick.
This approach breaks down, however, because pipe()ing from an ended
readable will just do nothing. No more data will ever arrive, and the
writable will hang open forever never being ended.
However, that does not solve the case of adding a `on('end')` listener
after the stream has received the EOF chunk, if it was the first chunk
received (and thus, length was 0, and 'end' got emitted). So, with
this, we defer the 'end' event emission until the read() function is
called.
Also, in pipe(), if the source has emitted 'end' already, we call the
cleanup/onend function on nextTick. Piping from an already-ended stream
is thus the same as piping from a stream that is in the process of
ending.
Updates many tests that were relying on 'end' coming immediately, even
though they never read() from the req.
Fix #4942
isaacs [Sat, 9 Mar 2013 18:56:17 +0000 (10:56 -0800)]
stream: Avoid nextTick warning filling read buffer
In the function that pre-emptively fills the Readable queue, it relies
on a recursion through:
stream.push(chunk) ->
maybeReadMore(stream, state) ->
if (not reading more and < hwm) stream.read(0) ->
stream._read() ->
stream.push(chunk) -> repeat.
Since this was only calling read() a single time, and then relying on a
future nextTick to collect more data, it ends up causing a nextTick
recursion error (and potentially a RangeError, even) if you have a very
high highWaterMark, and are getting very small chunks pushed
synchronously in _read (as happens with TLS, or many simple test
streams).
This change implements a new approach, so that read(0) is called
repeatedly as long as it is effective (that is, the length keeps
increasing), and thus quickly fills up the buffer for streams such as
these, without any stacks overflowing.
Julian Gruber [Fri, 8 Mar 2013 18:41:28 +0000 (19:41 +0100)]
events: Handle missing error obj when domains in use
so `ee.emit('error')` doesn't throw when domains are active
create an empty error only when handled by a domain
test for when no error is provided to an error event
koichik [Sun, 10 Mar 2013 11:10:19 +0000 (20:10 +0900)]
http: ServerRequest does not timeout after 'end'
Fixes #4967
isaacs [Sun, 10 Mar 2013 02:46:39 +0000 (18:46 -0800)]
http: Do not setTimeout a not-yet-existent socket
Fixes #4967
isaacs [Sun, 10 Mar 2013 02:34:59 +0000 (18:34 -0800)]
isaacs [Sat, 9 Mar 2013 17:22:00 +0000 (09:22 -0800)]
win/msi: Fix typos
isaacs [Sat, 9 Mar 2013 17:11:22 +0000 (09:11 -0800)]
npm: Upgrade to 1.2.14 (fixed)
hc [Fri, 8 Mar 2013 11:07:27 +0000 (06:07 -0500)]
http: check if incoming parser has already been freed
Fix #4948
This adds a check before setting the incoming parser
to null. Under certain circumstances it'll already be set to
null by freeParser().
Otherwise this will cause node to crash as it tries to set
null on something that is already null.
Andreas Madsen [Tue, 26 Feb 2013 17:26:09 +0000 (18:26 +0100)]
timers: consistent this keyword in setImmediate
When calling setImmediate with extra arguments the this keyword in the
callback would refer to the global object, but when not calling
setImmediate with extra arguments this would refer to the returned
handle object.
This commit fixes that inconsistency so its always set handle object.
The handle object was chosen for performance reasons.
isaacs [Sat, 9 Mar 2013 15:00:21 +0000 (07:00 -0800)]
domain: Fix double-exit on nested domains
Minor oversight in fix for #4953.
isaacs [Sat, 9 Mar 2013 05:52:09 +0000 (21:52 -0800)]
Now working on 0.10.0
isaacs [Sat, 9 Mar 2013 04:26:26 +0000 (20:26 -0800)]
lint
Gil Pedersen [Fri, 8 Mar 2013 08:26:53 +0000 (09:26 +0100)]
stream: Always defer preemptive reading to improve latency
isaacs [Sat, 9 Mar 2013 02:59:44 +0000 (18:59 -0800)]
Merge remote-tracking branch 'ry/v0.8' into master
Conflicts:
AUTHORS
ChangeLog
src/node_version.h
isaacs [Sat, 9 Mar 2013 02:31:21 +0000 (18:31 -0800)]
test: Fail faster in simple/test-cluster-bind-twice-v2
Crashing on windows, but at least now it's a crash rathert han a timeout.