Replace calls to fs.readFileSync() with an internal variant that does
not create Error objects on failure and is a bit speedier in general.
A secondary benefit is that it improves start-up times in the debugger
because it no longer emits thousands of exception debug events.
On a medium-sized application[0], this commit and its predecessor reduce
start-up times from about 1.5s to 0.5s and reduce the number of start-up
exceptions from ~6100 to 32, half of them internal to the application.
[0] https://github.com/strongloop/loopback-sample-app
PR-URL: https://github.com/nodejs/io.js/pull/1801
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Replace calls to fs.statSync() with an internal variant that does not
create Error or Stat objects that put strain on the garbage collector.
A secondary benefit is that it improves start-up times in the debugger
because it no longer emits thousands of exception debug events.
PR-URL: https://github.com/nodejs/io.js/pull/1801
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Modifies the Socket.setNoDelay and Socket.setKeepAlive methods to return
the socket instance instead of undefined, to allow for chaining.
PR-URL: https://github.com/nodejs/io.js/pull/1779
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Modifies the following methods to return the instance instead
of undefined, to allow for chaining these methods:
- net.Server.ref
- net.Server.unref
- net.Socket.ref
- net.Socket.unref
- dgram.Socket.ref
- dgram.Socket.unref
PR-URL: https://github.com/nodejs/io.js/pull/1768
Reviewed-By: Evan Lucas <evanlucas@me.com>
Buffer.byteLength is important for speed because it is called whenever a
new Buffer is created from a string.
This commit optimizes Buffer.byteLength execution by:
- moving base64 length calculation into JS-land, which is now much
faster
- remove redundant code and streamline the UTF8 length calculation
It also adds a benchmark and better tests.
PR-URL: https://github.com/nodejs/io.js/pull/1713
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Emit errors using `.destroy(err)` instead of `.destroy()` and
`.emit('error', err)`. Otherwise `close` event is emitted with the
`error` argument set to `false`, even if the connection was torn down
because of the error.
See: https://github.com/nodejs/io.js/issues/1119
PR-URL: https://github.com/nodejs/io.js/pull/1711
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
There is no need to check for globals and also check if it is
an array.
PR-URL: https://github.com/nodejs/io.js/pull/1722
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
REPL evaluate `.scope` when it needs to get a list of the variable names
available in the current scope. Do not throw if the output of such
evaluation is not array, just ignore it.
PR-URL: https://github.com/nodejs/io.js/pull/1682
Reviewed-By: Fedor Indutny <fedor@indutny.com>
In ssl.onerror event, `this` refers `ssl` so that
`this._secureEstablished` is always undefined. Fix it to refer
TLSSocket.
PR-URL: https://github.com/nodejs/io.js/pull/1661
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Modifies the setTimeout methods for the following prototypes:
- http.ClientRequest
- http.IncomingMessage
- http.OutgoingMessage
- http.Server
- https.Server
- net.Socket
- tls.TLSSocket
Previously, the above functions returned undefined. They now return
`this`. This is useful for chaining function calls.
PR-URL: https://github.com/nodejs/io.js/pull/1699
Reviewed-By: Roman Reiss <me@silverwind.io>
os.tmpdir() began stripping trailing slashes in
b57cc51d8d. This causes problems if
the temp directory is simply '/'. It also stripped trailing
slashes without first determining which slash type is used by
the current operating system. This commit only strips trailing
slashes if another character precedes the slash. On Windows, it
checks for ':', as not to strip slashes from something like 'C:\'.
It also only strips slashes that are appropriate for the user's
operating system.
Fixes: https://github.com/iojs/io.js/issues/1669
PR-URL: https://github.com/iojs/io.js/pull/1673
Reviewed-By: Rod Vagg <rod@vagg.org>
Reviewed-By: Christian Tellnes <christian@tellnes.no>
Previously, in the event of an unhandled error event, if the error is a
not an actual Error, then a default error is thrown. Now, the argument
is appended to the error message and added as the `context` property
of the error.
PR-URL: https://github.com/iojs/io.js/pull/1654
Reviewed-By: Rod Vagg <rod@vagg.org>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
AES-GCM or CHACHA20_POLY1305 ciphers must be used in current version of
Chrome to avoid an 'obsolete cryptography' warning.
Prefer 128 bit AES over 192 and 256 bit AES considering attacks that
specifically affect the larger key sizes but do not affect AES 128.
PR-URL: https://github.com/iojs/io.js/pull/1660
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Reviewed-By: Shigeki Ohtsu <ohtsu@iij.ad.jp>
Reviewed-By: Roman Reiss <me@silverwind.io>
In certain environments escape sequences could be splitted into
multiple chunks. For example, when user presses left arrow,
`\x1b[D` sequence could appear as two keypresses (`\x1b` + `[D`).
PR-URL: https://github.com/iojs/io.js/pull/1601
Fixes: https://github.com/iojs/io.js/issues/1403
Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com>
Reviewed-By: Roman Reiss <me@silverwind.io>
The second step of augmenting the internal REPL with persistent
history was to re-open the history file with a 'w' handle. This
truncated the file. If a user did not enter a new line before
closing the REPL, their history would be deleted.
PR-URL: https://github.com/iojs/io.js/pull/1605
Reviewed-By: Roman Reiss <me@silverwind.io>
The _debugger module uses the internal REPL module, but expects
to receive the userland REPL module. This fixes the breakage that
occurs by proxying the userland REPL module through the internal
module.
It also fixes an unintended in-REPL bug, where require(node-module)
was not resolving correctly.
PR-URL: https://github.com/iojs/io.js/pull/1605
Reviewed-By: Roman Reiss <me@silverwind.io>
This makes `TLSWrap` and `SecureContext` objects collectable by the
incremental gc.
`res = null` destroys the cyclic reference in the `reading` property.
`this.ssl = null` removes the remaining reference to the `TLSWrap`.
`this.ssl._secureContext.context = null` removes the reference to
the `SecureContext` object, even though there might be references
to `this.ssl._secureContext` somewhere.
The `reading` property will now throw an error if accessed after the
socket is closed, but that should not happen.
PR-URL: https://github.com/iojs/io.js/pull/1580
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Reviewed-By: Shigeki Ohtsu <ohtsu@iij.ad.jp>
This reverts commit 3fd7fc429c.
It was agreed that this change contained too much potential ecosystem
breakage, particularly around the inability to `delete` properties off a
`Url` object. It may be re-introduced for a later release, along with
better work on ecosystem compatibility.
PR-URL: https://github.com/iojs/io.js/pull/1602
Reviewed-By: Mikeal Rogers <mikeal.rogers@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Forrest L Norvell <forrest@npmjs.com>
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Reviewed-By: Isaac Z. Schlueter <i@izs.me>
Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com>
This reverts commit dbdd81a91b.
It was agreed that this change contained too much potential ecosystem
breakage, particularly around the inability to `delete` properties off a
`Url` object. It may be re-introduced for a later release, along with
better work on ecosystem compatibility.
PR-URL: #1602
Reviewed-By: Mikeal Rogers <mikeal.rogers@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Forrest L Norvell <forrest@npmjs.com>
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Reviewed-By: Isaac Z. Schlueter <i@izs.me>
Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com>
This reverts commit 66877216bd.
It was agreed that this change contained too much potential ecosystem
breakage, particularly around the inability to `delete` properties off a
`Url` object. It may be re-introduced for a later release, along with
better work on ecosystem compatibility.
PR-URL: #1602
Reviewed-By: Mikeal Rogers <mikeal.rogers@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Forrest L Norvell <forrest@npmjs.com>
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Reviewed-By: Isaac Z. Schlueter <i@izs.me>
Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com>
In addition to null, undefined and the empty string
are treated as empty (removing the component from the url).
The string '#' is treated same as empty values when
setting .hash.
The string '?' is treated same as empty values when
setting .search.
PR-URL: https://github.com/iojs/io.js/pull/1589
Fixes: https://github.com/iojs/io.js/issues/1588
Reviewed-By: Roman Reiss <me@silverwind.io>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
`StreamBase::AfterWrite` is passing handle as an argument to the
`afterWrite` function in net.js. Thus GC should not collect the handle
and the request separately and assume that they are tied together.
With this commit - request will always outlive the StreamBase instance,
helping us survive the GC pass.
Same applies to the ShutdownWrap instances, they should never be
collected after the StreamBase instance.
Fix: https://github.com/iojs/io.js/pull/1580
PR-URL: https://github.com/iojs/io.js/pull/1590
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Issue #1575 did introduce a check for options.terminal but this variable
wasn't able to get truthy, which in turn broke persistent history
completely. This changes the variable to get truthy on true terminals.
Additionally, the docs and the code did differ on which environment
variable was used for history. This changes the code to use
NODE_REPL_HISTORY_FILE.
PR-URL: https://github.com/iojs/io.js/pull/1593
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Do not enable ClientHello parser for async SNI/OCSP. Use new
OpenSSL-1.0.2's API `SSL_set_cert_cb` to pause the handshake process and
load the cert/OCSP response asynchronously. Hopefuly this will make
whole async SNI/OCSP process much faster and will eventually let us
remove the ClientHello parser itself (which is currently used only for
async session, see #1462 for the discussion of removing it).
NOTE: Ported our code to `SSL_CTX_add1_chain_cert` to use
`SSL_CTX_get0_chain_certs` in `CertCbDone`. Test provided for this
feature.
Fix: https://github.com/iojs/io.js/issues/1423
PR-URL: https://github.com/iojs/io.js/pull/1464
Reviewed-By: Shigeki Ohtsu <ohtsu@iij.ad.jp>
this creates a new internal module responsible for providing
the repl created via "iojs" or "iojs -i," and adds the following
options to the readline and repl subsystems:
* "repl mode" - determine whether a repl is strict mode, sloppy mode,
or auto-detect mode.
* historySize - determine the maximum number of lines a repl will store
as history.
The built-in repl gains persistent history support when the
NODE_REPL_HISTORY_FILE environment variable is set. This functionality
is not exposed to userland repl instances.
PR-URL: https://github.com/iojs/io.js/pull/1513
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Upcoming V8 changes will make it impossible to keep supporting the
smalloc module so deprecate it now and tell people to switch to
typed arrays.
The module is used in core in a few places so this commit makes the
public module private and replaces the public part with wrappers that
print deprecation notices.
PR-URL: https://github.com/iojs/io.js/pull/1564
Reviewed-By: Domenic Denicola <domenic@domenicdenicola.com>
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Allows customization of the lookup function used when
Socket.prototype.connect is called using a hostname.
PR-URL: https://github.com/iojs/io.js/pull/1505
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Reviewed-By: Yosuke Furukawa <yosuke.furukawa@gmail.com>
Separates out the lookup logic for net.Socket. In the event
the `host` property is an IP address, the lookup is skipped.
PR-URL: https://github.com/iojs/io.js/pull/1505
Reviewed-By: Chris Dickinson <christopher.s.dickinson@gmail.com>
Reviewed-By: Yosuke Furukawa <yosuke.furukawa@gmail.com>
When buffer list less than 2, no need to calculate the length.
The change's benchmark result is here:
https://gist.github.com/JacksonTian/2c9e2bdec00018e010e6
It improve 15% ~ 25% speed when list only have one buffer,
to other cases no effect.
PR-URL: https://github.com/iojs/io.js/pull/1437
Reviewed-By: Trevor Norris <trev.norris@gmail.com>
Reviewed-By: Brendan Ashworth <brendan.ashworth@me.com>
If `$NODE_PATH` contains trailing separators, `Module.globalPaths` will
contains empty strings. When `Module` try to resolve a module's path,
`path.resolve('', 'index.js')` will boil down to `$PWD/index.js`, which
makes sub modules can access global modules and get unexpected result.
PR-URL: https://github.com/iojs/io.js/pull/1488
Reviewed-By: Roman Reiss <me@silverwind.io>
tls.connect(options) with no options.host should accept a certificate
with CN: 'localhost'. Fix Error: Hostname/IP doesn't match
certificate's altnames: "Host: undefined. is not cert's CN: localhost"
'localhost' is not added directly to defaults because that is not
always desired (for example, when using options.socket)
PR-URL: https://github.com/iojs/io.js/pull/1493
Fixes: https://github.com/iojs/io.js/issues/1489
Reviewed-By: Brendan Ashworth <brendan.ashworth@me.com>
Reviewed-By: Roman Reiss <me@silverwind.io>
This reverts commit 06cfff9350.
Reverted because it introduced a regression where (because options were
modified in the later functionality) options.host and options.port would
be overridden with values provided in other, supported ways.
PR-URL: https://github.com/iojs/io.js/pull/1467
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>