The install prefix gets written to config.gypi and config.mk.  Tildes
were expanded in the first file but not in the second one, causing the
`make install` target to install files to a directory named `~` in
the current working directory.

Fixes: #75
PR-URL: #2307
Reviewed-By: Johan Bergström <[email protected]>
Reviewed-By: Sakthipriyan Vairamani <[email protected]>

 * doc: improve http.abort description
 * doc: mention that mode is ignored if file exists
 * docs: Fix default options for fs.createWriteStream()
 * Documentation update about Buffer initialization
 * doc: add a note about readable in flowing mode
 * doc: Document http.request protocol option
 * doc, comments: Grammar and spelling fixes
 * updated documentation for fs.createReadStream
 * Update child_process.markdown, spelling
 * doc: Clarified read method with specified size argument.
 * docs:events clarify emitter.listener() behavior
 * doc: two minor stream doc improvements
 * doc: clarify Readable._read and Readable.push
 * doc: stream.unshift does not reset reading state
 * doc: readable event clarification
 * doc: additional refinement to readable event

Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Ben Noorduis <[email protected]>
PR-URL: #2302

simple tests test-http-request-end.js, test-http-default-encoding.js
hangs in AIX. The root cause for both the failures is related to the
timing with which packets are sent between the client and server.
On the client side, one factor that affects the timing is Nagle's
algorithm. With Nagle enabled there may be a delay between two packets
as the stack may wait until either:
  a. An acknowledgement for the first packet is received, or
  b. 200 ms elapses.
before sending the second packet.

Similarly at the server side 2 sequential packages can be delivered to
the application either together or separatly.

On AIX we see that they are delivered separately to the server, while on
Linux delivered together. If we change the timing, for example disabling
Nagle on AIX we see the 2 packets delivered together and the tests pass.

In the test case simple/test-http-request-end.js, the client request
handler of the server receives and stores the data in a data callback,
closes the server in a request end callback, and writes to the client
and ends the response, in-line with the request receipt. An HTTP parser
module parses the incoming message, and invokes callback routines which
are registered for HTTP events (such as header, body, end etc.)

Because the termination sequence arrive in a separate packet, there is a
delay in parsing that message and identify that the client request ended
(and thereby invoke the request end call backhandler). Due to this delay,
the response close happens first, which in-turn destroys the server
socket leading to the fd and watcher removal from the uv loop abandoning
further events on this connection, and end call back never being called,
causing the reported hang.  simple/test-http-default-encoding.js suffers
from the same problem.

Also, remove the timer logic from the test case. Test harness anyways
contain a timer which controls the individual tests so remove such
controls from the test case, as suggested by @tjfontaine

Reviewed-By: Michael Dawson <[email protected]>
PR-URL: nodejs/node-v0.x-archive#9432

PORT-FROM: joyent/node @ 13e1131
PR-URL: #2294
Reviewed-By: Fedor Indutny <[email protected]>
Reviewed-By: Ben Noordhuis <[email protected]>

Any time the connection state or the underlying handle itself changes,
the socket's name (aka, local address) can change.

To deal with this we need to reset the cached sockname any time we
set or unset the internal handle or an existing handle establishes a
connection.

PR-URL: #2095
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Jeremiah Senkpiel <[email protected]>

- eliminate unnecessary intermediate process ("parent")
- children exit if runner dies unexpectedly (killed on a test timeout,
  for example)
- use explicit messaging from children to parents to indicate when
  worker is ready to accept http requests, rather than racing to see
  whether runner will make request before worker is listening

PR-URL: #1944
Reviewed-By: Johan Bergstrom <[email protected]>
Reviewed-By: Ben Noordhuis <[email protected]>

`enableTicketKeyCallback` and `onticketkeycallback` could be potentially
used to renew the TLS Session Tickets before they expire. However this
commit will introduce it only for private use yet, because we are not
sure about the API, and already need this feature for testing.

See: #2304
PR-URL: #2312
Reviewed-By: Shigeki Ohtsu <[email protected]>

When TLS Session Ticket is renewed by server - no Certificate record is
to the client. We are prepared for empty certificate in this case, but
this relies on the session reuse check, which was implemented
incorrectly and was returning false when the TLS Session Ticket was
renewed.

Use session reuse check provided by OpenSSL instead.

Fix: #2304
PR-URL: #2312
Reviewed-By: Shigeki Ohtsu <[email protected]>

This commit modifies util.inspect(obj) to additionally show the name of
the function that constructed the object. This often reveals useful
information about the object's prototype. In other words, instead of

> new Cls
{}

we have

> new Cls
Cls {}

This also works with exotic objects:

> class ArrayCls extends Array {}
> new ArrayCls(1, 2, 3)
ArrayCls [ 1, 2, 3 ]

The names of "trivial" constructors like Object and Array are not shown,
unless there is a mismatch between the object representation and the
prototype:

> Object.create([])
Array {}

This feature is inspired by browser devtools.

PR-URL: #1935
Reviewed-By: Roman Reiss <[email protected]>
Reviewed-By: Jeremiah Senkpiel <[email protected]>

Windows 10 wasn't listed in the executable manifest.
This caused problems with trying to detect Windows 10
via `os.release()`.

PR-URL: #2332
Reviewed-By: Roman Reiss <[email protected]>

Remove dead code paths that are created by assertions that will never
trigger. They may only trigger if either the `splitDeviceRe` or
`splitPathRe` regular expressions are modified. If at some point they
are modified, current unit tests will catch most of the resulting
errors and this commit adds extra tests to catch the remaining errors.

PR-URL: #2282
Reviewed-By: Roman Reiss <[email protected]>

According to docs, dropMembership() is automatically called
by the kernel when the socket is closed, and most apps will
never need to call it. It's called here as a sanity check
only so let's note that with a comment.

Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Ben Noordhuis <[email protected]>
PR-URL: #2062

v8 introduced the new flag `total_available_size` in version 4.4
and upwards. This flag is now available on `v8.getHeapStatistics`
with the name `total_available_size`. It contains the total
available heap size of v8.

Introduced with commit: v8/v8@0a1352a7

PR-URL: #2348
Reviewed-By: Ben Noordhuis <[email protected]>

With this option, tests can be divided in groups and only one is run
from each group. This enables us to parallelize the tests across
multiple machines.

PR-URL: #2260
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Rod Vagg <[email protected]>
Reviewed-By: Alexis Campailla <[email protected]>

node was using ICU4C 54 ( released 2014-oct-06 )

* Bump to ICU4C 55 ( released 2015-apr-08)
 * Timezone fixes, translation fixes, bug/performance fixes.
 * Speed improvements in date formatting
 * ICU4C 55 changelog: http://site.icu-project.org/download/55
 * CLDR 27 changelog: http://cldr.unicode.org/index/downloads/cldr-27

* Retarget the file exclusions from ICU4C 54 to ICU4C 55
 * (improves on-disk footprint on some platforms)

Reviewed-By: João Reis <[email protected]>
Reviewed-By: Johan Bergström <[email protected]>
Fixes: #2292
PR-URL: #2293

* turn on small-icu by default for builds (Makefile+Windows)
* add license info from ICU
  http://source.icu-project.org/repos/icu/icu/trunk/license.html
  All text pasted. Long lines wrapped. (original is HTML.)

Port from joyent/node of:
* #26
 * port of joyent/node 67f87a7
* nodejs/node-v0.x-archive#9038
 * Merge from joyent/node 70d04e7
 * Merge from joyent/node 6168fe6
 * merge from joyent/node e670732

PR-URL: #2264
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Rod Vagg <[email protected]>
Fixes: #238

In a few places dynamic memory was passed to the Buffer::New() overload
that makes a copy of the input, not the one that takes ownership.

This commit is a band-aid to fix the memory leaks.  Longer term, we
should look into using C++11 move semantics more effectively.

Fixes: #2308
PR-URL: #2352
Reviewed-By: Fedor Indutny <[email protected]>
Reviewed-By: Trevor Norris <[email protected]>

The circular dependency problem that put them there in the first place
is no longer an issue.  Move them out of the public node_buffer.h header
and into the private node_internals.h header.

Fixes: #2308
PR-URL: #2352
Reviewed-By: Fedor Indutny <[email protected]>
Reviewed-By: Trevor Norris <[email protected]>

Rename the three argument overload of Buffer::New() to Buffer::Copy()
and update the code base accordingly.  The reason for renaming is to
make it impossible to miss a call site.

This coincidentally plugs a small memory leak in crypto.getAuthTag().

Fixes: #2308
PR-URL: #2352
Reviewed-By: Fedor Indutny <[email protected]>
Reviewed-By: Trevor Norris <[email protected]>

Fixes: #2308
PR-URL: #2352
Reviewed-By: Fedor Indutny <[email protected]>
Reviewed-By: Trevor Norris <[email protected]>

The particular ufile.c is from
http://bugs.icu-project.org/trac/changeset/37704
and should be OK for ICU 54 and 55.

Also, adds general mechanism for floating patches on top of ICU.

Fixes: #2279
PR-URL: #2283
Reviewed-By: João Reis <[email protected]>

Fix a regression that was introduced in commit 1bbf8d0 ("lib: speed up
require(), phase 2") where file paths with Unicode characters fail to
load on Windows.

Fixes: #2236
PR-URL: #2377
Reviewed-By: Bert Belder <[email protected]>

Per: nodejs/node-v0.x-archive#7554

Originally submitted by @stites

PR-URL: #2388
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Roman Reiss <[email protected]>

Per: nodejs/node-v0.x-archive#8715

Cleans up a minor spelling error in two tests.

PR-URL: #2389
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Roman Reiss <[email protected]>

PR-URL: #2375
Reviewed-By: Ben Noordhuis <[email protected]>

PR-URL: #2375
Reviewed-By: Ben Noordhuis <[email protected]>

PR-URL: #2375
Reviewed-By: Ben Noordhuis <[email protected]>

PR-URL: #2375
Reviewed-By: Ben Noordhuis <[email protected]>

PR-URL: #2375
Reviewed-By: Ben Noordhuis <[email protected]>

PR-URL: #2375
Reviewed-By: Ben Noordhuis <[email protected]>

Currently, v8 native deps must be built in order to run the log
processor on node profiling output. These scripts use node instead
of d8 to remove this dependency.

This change was originally proposed to the v8 team but since the
changes are not v8 specific, we have moved the proposal here. See:
https://codereview.chromium.org/1179173009/

PR-URL: #2090
Reviewed-By: Ali Ijaz Sheikh <[email protected]>
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Fedor Indutny <[email protected]>
Reviewed-By: Jeremiah Senkpiel <[email protected]>
Reviewed-By: Trevor Norris <[email protected]>

1e066e4 was done incorrectly and has overwritten an important change
in: c17449d. Using bigger output buffer increases performance in 3-4
times.

PR-URL: #2381
Fix: nodejs/node-v0.x-archive#25803
Reviewed-By: Trevor Norris <[email protected]>
Reviewed-By: Sakthipriyan Vairamani <[email protected]>

Errors might be injected into OpenSSL's error stack
without the return value of `PEM_read_bio_PrivateKey` being set to
`nullptr`. See the test of `test_bad_rsa_privkey.pem` for an
example.

PR-URL: #2342
Reviewed-By: Fedor Indutny <[email protected]>

PR-URL: #2422
Reviewed-By: Ben Noordhuis <[email protected]>

The test no longer waits about 5 seconds between callback invocations.

It now writes to the tmp directory rather than the fixtures directory.

Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Sakthipriyan Vairamani <[email protected]>
PR-URL: #2393

The explanation for the `options` argument to `fs.watchFile()` had
missing punctuation. I took the opportunity to try to rewrite the
paragraph for greater clarity.

PR-URL: #2425
Reviewed-By: Sakthipriyan Vairamani <[email protected]>
Reviewed-By: Brendan Ashworth <[email protected]>

Add files in test/addon to the `make cpplint` rule and fix up existing
style issues.  Tests scraped from doc/api/addon.md are filtered out
because those are predominantly for illustrative purposes.

PR-URL: #2427
Reviewed-By: Trevor Norris <[email protected]>

v8::Isolate::GetCurrent() is not exactly deprecated at this point but
its use is strongly discouraged.  Update the addon tests so they no
longer use it.

PR-URL: #2427
Reviewed-By: Trevor Norris <[email protected]>

Update the instructions to follow the requirements
in the security policy and user guide

PR-URL: #2278
Fixes: #2242
Reviewed-By: Fedor Indutny <[email protected]>
Reviewed-By: Rod Vagg <[email protected]>

Notable changes:

* buffer: Fixed a couple large memory leaks (Ben Noordhuis)
#2352.
* crypto:
  - Fixed a couple of minor memory leaks (Karl Skomski)
#2375.
  - Signing now checks for OpenSSL errors (P.S.V.R)
#2342. **Note that this may expose
previously hidden errors in user code.**
* intl: Intl using small-icu is now enabled by default in builds
(Steven R. Loomis) #2264.
  - `String#normalize()` can now be used for unicode
normalization.
  - The `Intl` object and various `String` and `Number` methods are
present, but only support the English locale.
  - For support of all locales, node must be built with full-icu.
* tls: Fixed tls throughput being much lower after an incorrect merge
(Fedor Indutny) #2381.
* tools: The v8 tick processor now comes bundled with node (Matt
Loring) #2090.
  - This can be used by producing performance profiling output by
running node with `--perf`, then running your appropriate platform's
script on the output as found in
https://github.com/nodejs/node/tree/master/tools/v8-prof.
* util: `util.inspect(obj)` now prints the constructor name of the
object if there is one (Christopher Monsanto)
#1935.

PR-URL: #2347

PR-URL: #2347

As per the discussion in #734, this patch deprecates the usage of
`EventEmitter.listenerCount` static function in the docs, and introduces
the `listenerCount` function in the prototype of `EventEmitter` itself.

PR-URL: #2349
Reviewed-By: Trevor Norris <[email protected]>
Reviewed-By: Brian White <[email protected]>

This includes the following changes:
- a more strict data check rather than a regex
- reduced number of annoying log calls

The most important of the changes is the annoying log calls, which
speeds up the test execution from about 0m1.130s to 0m0.481s on my
machine.

PR-URL: #2429
Reviewed-By: Rich Trott <[email protected]>

Previously the test had a massive timeout (3s!), reduce this to a
platform specific timeout of 50ms.

This test runs two servers at the same time in an attempt to compare
behaviour. I've added a check to make sure one event fires before the
other event, as is expected, but that is a possible race condition.

Improves test run time on my machine from 0m3.141s to 0m0.356s.

PR-URL: #2429
Reviewed-By: Rich Trott <[email protected]>

Previously 1000-1200ms, they're now (platform dependent) 50-100ms.

Improves test run time on my machine from 0m1.335s to 0m0.236s.

PR-URL: #2429
Reviewed-By: Rich Trott <[email protected]>

Expose and use in TLSWrap an `v8::External` wrap of the
`StreamBase*` pointer instead of guessing the ancestor C++ class in
`node_wrap.h`.

Make use of `StreamBase::Callback` structure for storing/passing both
callback and context in a single object.

Introduce `GetObject()` for future user-land usage, when a child class
is not going to be inherited from AsyncWrap.

PR-URL: #2351
Reviewed-By: Trevor Norris <[email protected]>

If you have no history file written to disk, but the environment
variable set, `fs.readFileSync` will throw an ENOENT error,
but there's nothing to convert. The converter should ignore
ENOENT on that `fs.readFileSync` call.

Fixes: #2449
PR-URL: #2451
Reviewed-By: Jeremiah Senkpiel <[email protected]>
Reviewed-By: Sakthipriyan Vairamani <[email protected]>

PR-URL: #2289
Reviewed-By: Sakthipriyan Vairamani <[email protected]>
Reviewed-By: Roman Reiss <[email protected]>
Reviewed-By: Brendan Ashworth <[email protected]>

There is no way a line can be called after throwing an exception.

PR-URL: #2289
Reviewed-By: Sakthipriyan Vairamani <[email protected]>
Reviewed-By: Roman Reiss <[email protected]>
Reviewed-By: Brendan Ashworth <[email protected]>

Per: nodejs/node-v0.x-archive#8876

PR-URL: #2391
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Roman Reiss <[email protected]>

This resolves nodejs/node-v0.x-archive#9272. `tlsSocket.getPeerCertificate` will
return an empty object when the peer does not provide a certificate,
but, prior to this, when the certificate is empty, `checkServerIdentity`
would throw because the `subject` wasn't present on the cert.
`checkServerIdentity` must return an error, not throw one, so this
returns an error when the cert is empty instead of throwing
a `TypeError`.

PR-URL: #2343
Reviewed-By: Fedor Indutny <[email protected]>
Reviewed-By: Shigeki Ohtsu <[email protected]>

Use correct deallocator for returned buffera.
Don't free internal structure via ASN1_STRING_data.
Deallocate NETSCAPE_SPKI.

PR-URL: #2359
Reviewed-By: Fedor Indutny <[email protected]>

PR-URL: #2436
Reviewed-By: Ben Noordhuis <[email protected]>

Previously, V8:SetFlagsFromCommandLine was being called even if v8_argc
was 0. This change prevents that from being called unless v8 arguments
are actually passed.

Improves startup time by about 5%.

PR-URL: #2483
Reviewed-By: Rod Vagg <[email protected]>
Reviewed-By: Sakthipriyan Vairamani <[email protected]>

Support unaligned output buffer when writing out UCS2 in
`StringBytes::Write`.

Fix: #2457
PR-URL: #2480
Reviewed-By: Trevor Norris <[email protected]>

Original commit message:

    buffer: align chunks on 8-byte boundary

    When slicing global pool - ensure that the underlying buffer's data
    ptr is 8-byte alignment to do not ruin expectations of 3rd party C++
    addons.

    NOTE: 0.10 node.js always returned aligned pointers and io.js should
    do this too for compatibility.

PR-URL: #2487
Reviewed-By: Trevor Norris <[email protected]>

Refactored version of nodejs/node-v0.x-archive#25819

Removes integer keys (and keys starting with numbers) from
candidate list on repl tab complete. Refactored the originally
submitted change to simplify and ensure that the integer keys
do not show up on objects either.

Reviewed By: Sakthipriyan Vairamani <[email protected]>
PR-URL: #2409

PR-URL: #2454
Reviewed-By: James M Snell <[email protected]>

nodejs/node -> nodejs/node-convergence-archive

PR-URL: #2454
Reviewed-By: James M Snell <[email protected]>

nodejs/node -> nodejs/node-convergence-archive

PR-URL: #2454
Reviewed-By: James M Snell <[email protected]>

PR-URL: #2453
Reviewed-By: James M Snell <[email protected]>

PR-URL: #2460
Reviewed-By: James M Snell <[email protected]>

PR-URL: #2437
Reviewed-By: Ben Noordhuis <[email protected]>

Update AUTHORS list using tools/update-authors.sh

PR-URL: #2318
Reviewed-By: Ben Noordhuis <[email protected]>

When an anchor tag is used within a pre tag, the link is not
distinguishable. This patch makes sure that the links are highlighted
by underlining them.

PR-URL: #2491
Reviewed-By: Roman Reiss <[email protected]>

Instead of throwing an error with `Bad arguments` the fs methods
return a concrete error message. If an argument is missing, the
methods throw an error with the information, what is missing.
In case of a type mismatch, they throw an error with some hints,
what datatype is expected.

PR-URL: #2495
Reviewed-By: Trevor Norris <[email protected]>
Reviewed-By: Sakthipriyan Vairamani <[email protected]>

eg changes:

  if(x) { ... }

to:

  if (x) { ... }

Requires that you do:

  if (x) { ... }

Rather than:

  if(x) { ... }

Update AUTHORS list using tools/update-authors.sh
Fix @muddletoes' .mailmap entry

PR-URL: #2505
Reviewed-By: Ben Noordhuis <[email protected]>

Adds mention of response.finished to http.markdown
Originally submitted by @HACKERJS. The original
commit needed a bit of cleanup on grammar.

Reviewed By: James M Snell <[email protected]>
Reviewed By: Michaël Zasso <[email protected]>
PR-URL: #2414

This adds a new `--tls-cipher-list` command line switch
that can be used to override the built-in default cipher
list. The intent of this is to make it possible to enforce
an alternative default cipher list at the process level.
Overriding the default cipher list is still permitted at
the application level by changing the value of
`require('tls').DEFAULT_CIPHERS`.

As part of the change, the built in default list is moved
out of tls.js and into node_constants.h and node_constants.cc.
Two new constants are added to require('constants'):

  * defaultCipherList (the active default cipher list)
  * defaultCoreCipherList (the built-in default cipher list)

A test case and doc changes are included.

A new NODE_DEFINE_STRING_CONSTANT macro is also created in
node_internals.h

When node_constants is initialized, it will pick up either
the passed in command line switch or fallback to the default
built-in suite.

Within joyent/node, this change had originaly been wrapped
up with a number of other related commits involving the
removal of the RC4 cipher. This breaks out this isolated
change.

/cc @mhdawson, @misterdjules, @trevnorris, @indutny, @rvagg

Reviewed By: Ben Noordhuis <[email protected]>
PR-URL: #2412

Per: nodejs/node-v0.x-archive#6962

This applies only the first commit from nodejs/node-v0.x-archive#6962.

/cc @bnoorduis

Reviewed By: Ben Noordhuis <[email protected]>
PR-URL: #2386

fixed grammatically wrong expressions to make it clear

Reviewed By: James M Snell <[email protected]>
PR-URL: #2490
PR-URL: #2489

Use reserved IP in test-net-connect-timeout.js rather than arbitrary
IP.

PR-URL: #2257
Fixes: #2469
Reviewed By: Ben Noordhuis <[email protected]>

Notable changes:

...