net: Socket.prototype.connect should accept args like net.connect #11761 #11762

vhain · 2017-03-09T03:38:30Z

Checklist

make -j4 test (UNIX), or vcbuild test (Windows) passes
tests and/or benchmarks are included
commit message follows commit guidelines

Affected core subsystem(s)

net.Socket.prototype.connect

#11761

sam-github

needs test coverage

vhain · 2017-03-09T04:55:36Z

@sam-github just added test

mscdex · 2017-03-09T08:34:03Z

lib/net.js

+  const args = new Array(arguments.length);
+  for (var i = 0; i < arguments.length; i++)
+    args[i] = arguments[i];
+  // TODO(joyeecheung): use destructuring when V8 is fast enough


@joyeecheung Are you ok with this comment being added?

this code is same code as original line number 62 to 68.
and while doing this, i also considered exports.createConnection this function to just create new Socket and call connect by just forwarding arguments (for higher code coverage, if it is okay to move options.timeout also within Socket.prototype.connect.

e.g.

... exports.connect = exports.createConnection = function(options) { const socket = new Socket(options); return Socket.prototype.connect.apply(socket, arguments); }; ... Socket.prototype.connect = function() { const args = new Array(arguments.length); for (var i = 0; i < arguments.length; i++) args[i] = arguments[i]; // TODO(joyeecheung): use destructuring when V8 is fast enough const normalized = normalizeArgs(args); const options = normalized[0]; const cb = normalized[1]; debug('Socket.connect', normalized); if (options.timeout) this.setTimeout(options.timeout); if (this.write !== Socket.prototype.write) this.write = Socket.prototype.write; ...

what do you think @joyeecheung @mscdex

never mind. i think this code cannot be merged easily since new Socket requires normalized options.

I'll be OK with that if it's just copy-pasting.

The timeouts are added in #8101, I can't see any harm adding them to Socket.prototype.connect, so +1 to reduce code duplicate.

Yup. but anyway we need to normalizeArgs on exports.createConnection since arguments of new Socket needs normalized arguments.

I will just add one more commit to move timeout from exports.createConnection to Socket.prototype.connect.

After that, maybe it should be good to merge?

@vhain I think it's already good to merge since it fixes #11761 without moving the timeout option handling.

@vhain (hit the button too soon)..but yeah you can push it here(should be semver-patch) or add it in a separate PR(semver-minor).

@joyeecheung oops.. was too fast. already added commit to this PR. do you want me to revert that commit only?

@vhain I don't think it needs to be reverted if no one objects to / spots a bug caused by moving it down though.

@joyeecheung right. it needs to be different branch and PR. i'm on it.

it's done ccb807e

joyeecheung · 2017-03-09T09:05:52Z

lib/net.js

+  const args = new Array(arguments.length);
+  for (var i = 0; i < arguments.length; i++)
+    args[i] = arguments[i];
+  // TODO(joyeecheung): use destructuring when V8 is fast enough


I'll be OK with that if it's just copy-pasting.

The timeouts are added in #8101, I can't see any harm adding them to Socket.prototype.connect, so +1 to reduce code duplicate.

joyeecheung · 2017-03-09T09:27:20Z

CI: https://ci.nodejs.org/job/node-test-pull-request/6760/

joyeecheung · 2017-03-09T09:35:50Z

New CI after revert: https://ci.nodejs.org/job/node-test-pull-request/6761/

bajtos · 2017-03-09T12:15:54Z

FWIW, to make this pull request easier to find via google/github search - here is the stack trace we observed in LoopBack tests when calling net.createConnection(port) (see strongloop/loopback#3262)

Uncaught TypeError: "listener" argument must be a function
    at Socket.connect (net.js:943:10)
    at Socket.connect (node_modules/async-listener/index.js:76:27)
    at Object.exports.connect.exports.createConnection (net.js:76:35)
    at sendHttpRequestInOnePacket (test/integration.test.js:77:24)
    at Server.<anonymous> (test/integration.test.js:19:7)
    at emitListeningNT (net.js:1302:10)
    at node_modules/async-listener/glue.js:188:31
    at _combinedTickCallback (internal/process/next_tick.js:77:11)
    at process._tickDomainCallback [as _tickCallback] (internal/process/next_tick.js:128:9)

cjihrig

LGTM with a nit.

cjihrig · 2017-03-09T12:15:32Z

test/parallel/test-net-socket-connect-without-cb.js

+const server = net.createServer((conn) => {
+  conn.end();
+  server.close();
+}).listen(0, () => {


The callback here needs to be wrapped in common.mustCall().

evanlucas · 2017-03-09T12:52:19Z

test/parallel/test-net-socket-connect-without-cb.js

+}).listen(0, () => {
+  const client = new net.Socket();
+
+  client.on('connect', () => {


The callback here needs to be wrapped in common.mustCall() as well

I am on it. but then test/parallel/test-net-socket-connecting.js also need to be changed in future?

anyway for this PR, i will only wrap test-net-socket-connect-without-cb.js callback with common.mustCall

@evanlucas commit is added. Thanks for the comment.

sam-github

other than the missing comment in the test file, LGTM. It also needs to be rebased and squashed.

sam-github · 2017-03-09T19:06:33Z

test/parallel/test-net-socket-connect-without-cb.js

+'use strict';
+const common = require('../common');
+const net = require('net');
+


https://github.com/nodejs/node/blob/master/doc/guides/writing-tests.md#lines-4-5

sam-github · 2017-03-09T19:11:38Z

https://github.com/nodejs/node/blob/master/CONTRIBUTING.md#step-3-commit <--after squashing, commit message should be reworked to follow guidelines (it shouldn't have the PR number in the subject, and doesn't need PR-URL in the description, but it should have a Fixes: https://github.com/nodejs/node/pull/11761 in the description)

vhain · 2017-03-09T19:42:49Z

@sam-github comment added to test to match guideline, as well as squashing.
Again, thank you for the comments.

mcollina

LGTM

sam-github · 2017-03-11T00:33:13Z

@nodejs/ctc this fixes a breaking bug in recent 7.x, it would be good to get a fix out

addaleax · 2017-03-11T23:06:43Z

@nodejs/release ^^^

evanlucas · 2017-03-11T23:49:00Z

One of us will be able to cut one next week. We try to release on Tuesday or Wednesday generally.

joyeecheung · 2017-03-13T03:53:39Z

@evanlucas The callbacks are wrapped in common.mustCall now, LGTY?

CI: https://ci.nodejs.org/job/node-test-pull-request/6817/

Arguments of Socket.prototype.connect should be also normalized, causing error when called without callback. Changed Socket.prototype.connect's code same as net.connect and added test. Fixes: #11761 PR-URL: #11762 Reviewed-By: Sam Roberts <vieuxtech@gmail.com> Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com> Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Evan Lucas <evanlucas@me.com>

joyeecheung · 2017-03-13T12:10:02Z

ARM and FreeBSD failures are false positives. Landed in c6cbbf9, again, sorry for the inconvenience and thanks for the fix!

Arguments of Socket.prototype.connect should be also normalized, causing error when called without callback. Changed Socket.prototype.connect's code same as net.connect and added test. Fixes: nodejs#11761 PR-URL: nodejs#11762 Reviewed-By: Sam Roberts <vieuxtech@gmail.com> Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com> Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Evan Lucas <evanlucas@me.com>

Notable changes: * module: fix loading from global folders on Windows (Richard Lau) [nodejs#9283](nodejs#9283) * net: allow missing callback for Socket.connect (Juwan Yoo) [nodejs#11762](nodejs#11762) PR-URL: nodejs#11831

Notable changes: * module: The [module loading global fallback] (https://nodejs.org/dist/latest-v6.x/docs/api/modules.html#modules_loading_from_the_global_folders) to the Node executable's directory now works correctly on Windows. (Richard Lau) [nodejs#9283](nodejs#9283) * net: `Socket.prototype.connect` now once again functions without a callback. (Juwan Yoo) [nodejs#11762](nodejs#11762) * url: `URL.prototype.origin` now properly specified an opaque return of `'null'` for `file://` URLs. (Brian White) [nodejs#11691](nodejs#11691) PR-URL: nodejs#11831

Notable changes: * module: The [module loading global fallback] (https://nodejs.org/dist/latest-v6.x/docs/api/modules.html#modules_loading_from_the_global_folders) to the Node executable's directory now works correctly on Windows. (Richard Lau) [#9283](nodejs/node#9283) * net: `Socket.prototype.connect` now once again functions without a callback. (Juwan Yoo) [#11762](nodejs/node#11762) * url: `URL.prototype.origin` now properly specified an opaque return of `'null'` for `file://` URLs. (Brian White) [#11691](nodejs/node#11691) PR-URL: nodejs/node#11831 Signed-off-by: Ilkka Myller <ilkka.myller@nodefield.com>

Arguments of Socket.prototype.connect should be also normalized, causing error when called without callback. Changed Socket.prototype.connect's code same as net.connect and added test. Fixes: nodejs#11761 PR-URL: nodejs#11762 Reviewed-By: Sam Roberts <vieuxtech@gmail.com> Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com> Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Evan Lucas <evanlucas@me.com>

Notable changes: * module: The [module loading global fallback] (https://nodejs.org/dist/latest-v6.x/docs/api/modules.html#modules_loading_from_the_global_folders) to the Node executable's directory now works correctly on Windows. (Richard Lau) [nodejs#9283](nodejs#9283) * net: `Socket.prototype.connect` now once again functions without a callback. (Juwan Yoo) [nodejs#11762](nodejs#11762) * url: `URL.prototype.origin` now properly specified an opaque return of `'null'` for `file://` URLs. (Brian White) [nodejs#11691](nodejs#11691) PR-URL: nodejs#11831

MylesBorins · 2017-04-17T22:38:19Z

should this be backported?

mcollina · 2017-05-02T17:27:46Z

Only if we decide to backport #11667, as this is a fix for that.

joyeecheung · 2017-05-10T12:05:06Z

#11667 depends on a 7.x semver major patch so this one should not be backported either.

nodejs-github-bot added the net Issues and PRs related to the net subsystem. label Mar 9, 2017

vhain force-pushed the net-socket-connect-cb-fix branch from 6305ab4 to 3d2a8b3 Compare March 9, 2017 03:54

sam-github suggested changes Mar 9, 2017

View reviewed changes

mscdex reviewed Mar 9, 2017

View reviewed changes

joyeecheung approved these changes Mar 9, 2017

View reviewed changes

joyeecheung mentioned this pull request Mar 9, 2017

Fix CI on Node 7.7.2 strongloop/loopback#3262

Merged

mcollina approved these changes Mar 9, 2017

View reviewed changes

cjihrig approved these changes Mar 9, 2017

View reviewed changes

evanlucas suggested changes Mar 9, 2017

View reviewed changes

sam-github suggested changes Mar 9, 2017

View reviewed changes

vhain force-pushed the net-socket-connect-cb-fix branch from 7e8a552 to 48924fc Compare March 9, 2017 19:39

mcollina approved these changes Mar 9, 2017

View reviewed changes

This was referenced Mar 10, 2017

TypeError: "listener" argument must be a function keymetrics/pm2-logrotate#48

Closed

TypeError: "listener" argument must be a function opbeat/opbeat-node#129

Closed

Gambit is crashing when New Relic is enabled DoSomethingArchive/gambit-content#811

Closed

evanlucas approved these changes Mar 13, 2017

View reviewed changes

joyeecheung closed this Mar 13, 2017

italoacasas mentioned this pull request Mar 13, 2017

7.7.3 Release Proposal #11831

Merged

BarryCarlyon mentioned this pull request Mar 14, 2017

Can't create a redis Client under Node v7.7+ redis/node-redis#1210

Closed

joyeecheung mentioned this pull request Mar 14, 2017

cb may be undefined as well #11848

Closed

2 tasks

bajtos mentioned this pull request Mar 15, 2017

Run the latest Node.js 7 version on Travis again strongloop/loopback#3285

Merged

lindehoff mentioned this pull request Mar 15, 2017

TypeError('"listener" argument must be a function'); mscdex/node-imap#599

Closed

MylesBorins added dont-land-on-v6.x and removed dont-land-on-v6.x labels Apr 17, 2017

joyeecheung mentioned this pull request May 9, 2017

net: refactor overloaded argument handling #11667

Closed

3 tasks

joyeecheung added dont-land-on-v6.x and removed lts-watch-v6.x labels May 10, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

net: Socket.prototype.connect should accept args like net.connect #11761 #11762

net: Socket.prototype.connect should accept args like net.connect #11761 #11762

vhain commented Mar 9, 2017 •

edited

Loading

sam-github left a comment

vhain commented Mar 9, 2017

mscdex Mar 9, 2017

vhain Mar 9, 2017 •

edited

Loading

vhain Mar 9, 2017

joyeecheung Mar 9, 2017 •

edited

Loading

vhain Mar 9, 2017

joyeecheung Mar 9, 2017

joyeecheung Mar 9, 2017

vhain Mar 9, 2017

joyeecheung Mar 9, 2017

vhain Mar 9, 2017 •

edited

Loading

joyeecheung Mar 9, 2017 •

edited

Loading

joyeecheung commented Mar 9, 2017

joyeecheung commented Mar 9, 2017

bajtos commented Mar 9, 2017 •

edited

Loading

cjihrig left a comment

cjihrig Mar 9, 2017

evanlucas Mar 9, 2017

vhain Mar 9, 2017

vhain Mar 9, 2017

sam-github left a comment

sam-github Mar 9, 2017

sam-github commented Mar 9, 2017

vhain commented Mar 9, 2017

mcollina left a comment

sam-github commented Mar 11, 2017

addaleax commented Mar 11, 2017

evanlucas commented Mar 11, 2017

joyeecheung commented Mar 13, 2017

joyeecheung commented Mar 13, 2017 •

edited

Loading

MylesBorins commented Apr 17, 2017

mcollina commented May 2, 2017

joyeecheung commented May 10, 2017

net: Socket.prototype.connect should accept args like net.connect #11761 #11762

net: Socket.prototype.connect should accept args like net.connect #11761 #11762

Conversation

vhain commented Mar 9, 2017 • edited Loading

Checklist

Affected core subsystem(s)

sam-github left a comment

Choose a reason for hiding this comment

vhain commented Mar 9, 2017

Choose a reason for hiding this comment

vhain Mar 9, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

joyeecheung Mar 9, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

vhain Mar 9, 2017 • edited Loading

Choose a reason for hiding this comment

joyeecheung Mar 9, 2017 • edited Loading

Choose a reason for hiding this comment

joyeecheung commented Mar 9, 2017

joyeecheung commented Mar 9, 2017

bajtos commented Mar 9, 2017 • edited Loading

cjihrig left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

sam-github left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

sam-github commented Mar 9, 2017

vhain commented Mar 9, 2017

mcollina left a comment

Choose a reason for hiding this comment

sam-github commented Mar 11, 2017

addaleax commented Mar 11, 2017

evanlucas commented Mar 11, 2017

joyeecheung commented Mar 13, 2017

joyeecheung commented Mar 13, 2017 • edited Loading

MylesBorins commented Apr 17, 2017

mcollina commented May 2, 2017

joyeecheung commented May 10, 2017

vhain commented Mar 9, 2017 •

edited

Loading

vhain Mar 9, 2017 •

edited

Loading

joyeecheung Mar 9, 2017 •

edited

Loading

vhain Mar 9, 2017 •

edited

Loading

joyeecheung Mar 9, 2017 •

edited

Loading

bajtos commented Mar 9, 2017 •

edited

Loading

joyeecheung commented Mar 13, 2017 •

edited

Loading