stream docs, writable.cork() and writable.uncork() methods

[italoacasas]

Working on #7287 I notice that writable.cork() and writable.uncork() are not together, I guess it's because the alphabetical order, but I think the experience reading the stream documentation is going to be better if they live together.

[cjihrig]

You're right, but I think the overall doc experience is better if everything is predictably in alphabetical order.

[italoacasas]

Yes, alphabetical order I think is the right choose, but on my opinion we can improve the experience if we make an exception on this case.

[cjihrig]

There are many other cases in the documentation where an exception would be warranted. The problem is that if you keep making exceptions, then the documentation loses all consistency.

[jasnell]

I'm -1 on changing the order. Improving the documentation so that the links between the two are clearer is definitely a good idea.

[addaleax]

How about adding See also: links in the places where it would otherwise make sense to reorder, e.g. one from cork() to uncork() and vice versa?

[cjihrig]

Makes sense to me.

[Salakar]

On a side note - it seems these methods are moot for socket connections to non local servers? Using cork / uncork to buffer writes to a local connection practically doubles the op/s. But to a remote server it halves it...? 😭

Is there any different logic for non-local connections? See redis/ioredis#343 that PR I did, I switched to not using cork as it's more performant (~70k ops/s more), but even with a cork/uncork implementation the result on remote servers is the same - worse.

If there is some different logic then is it worth documenting this?

[mcollina]

Fixed in c38b6d2