lib: add an alias at addListener on Server connection socket

[himself65]

about #27199

socket.addListener should be the same behavior with socket.on

Checklist

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

[apapirovski]

This isn't a complete change. Code around line 748 also needs to be modified to add this line: this.addListener = net.Socket.prototype.addListener;

Personally I would also modify the relevant test (test-http-server-unconsume.js) to account for this change.

[himself65]

@apapirovski fixed

[Antonius-S]

Isn't more gentle and logical to add newlistener listener instead? And how about prependListener?

[himself65]

Isn't more gentle and logical to add newlistener listener instead? And how about prependListener?

why?

all this extend EventEmitter class

and all behavior should be the same as EventEmitter, or It will cause the disorganized behavior and namespace problem

node/lib/events.js

Lines 269 to 273 in 2161690

	EventEmitter.prototype.addListener = function addListener(type, listener) {
	return _addListener(this, type, listener, false);
	};
	EventEmitter.prototype.on = EventEmitter.prototype.addListener;

[apapirovski]

Thanks! LGTM

[ZYSzys]

@Antonius-S @SimonSchick Hi, did you have any other thought on this ? Otherwise we'll land this.

[SimonSchick]

@ZYSzys Imo this is not a fix but just a band-aid/work-around, this should be fixed using the new listener event.
Otherwise the same issue will occur for prependListener etc.

Also doesn't that break off and removeEventListener?

I don't really have any authority here and therefor won't block anything but I don't think this should be merged.

[himself65]

@SimonSchick
emmm, I have a new idea that use Object​.define​Property() to bind functions on EventEmitter
And I will create a new pull request in the few days

[himself65]

@SimonSchick
I did further modifications at #27427, How would you think of it?

[apapirovski]

@himself65 imo that PR is a step back. Just add prependListener here if needed. I don’t see what other edge cases there are tbh.

Otherwise, yes we could use an event for newListener but that’s a lot more work... I dunno. Seems like semantics.

Also as far as I can tell the main reason to not use events here would be performance.

[Trott]

I removed author ready pending the prependListener() fix, but if someone else thinks that shouldn't block this, by all means, go ahead and land it.

[himself65]

@Trott added

[Trott]

LGTM. Could add Fixes: https://github.com/nodejs/node/issues/27199 to the commit message?

[himself65]

LGTM. Could add Fixes: https://github.com/nodejs/node/issues/27199 to the commit message?

sure

[himself65]

no one to merge this?

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/23199/

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/23209/

[addaleax]

Landed in be26f6e, thanks for the PR! 🎉

(Changed the ommit message changed to begin with http: while landing)