Add methods to console as per spec

[aks-]

• added noop implementions of methods required as per below spec
• added aliases for common console methods
• https://github.com/DeveloperToolsWG/console-object/blob/master/api.md

[Fishrock123]

Related: #1799 (comment)

cc @silverwind & @yosuke-furukawa

[Fishrock123]

Perhaps we can define a noop function once in the file, and then assign them all to it.

[jasnell]

+1 on defining a single noop function.

[mscdex]

+1 to that too

[aks-]

aha, changed the assignments to noop

[jasnell]

@aks- ... probably the next thing I'd say is that a note should be added to the docs that these are noops and while they are defined, they are not expected to actually do anything.

[aks-]

@jasnell Hmm +1. I will add them tomorrow the first thing

[bnoordhuis]

This should probably be semver-major because it's going to break polyfills that do feature detection, e.g.

if (!console.profile) polyfill();

Stubbing methods is also something of a questionable practice IMO. If you're not supporting something, don't implement it. -1 from me although I don't feel really strongly about it.

[silverwind]

Strong -1 here, this would break feature detection as Ben mentions. We have #1799 which takes care of the group* methods, which I plan to land eventually. A noop isn't helping here, and these methods should be taken a look on a case-by-case basis, if something isn't feasable to do in a terminal, we should just throw imho.

[Fishrock123]

@silverwind If we aren't going to follow the console spec there is no reason to land features like console.group() though. Then those belong in user-land.

[silverwind]

@Fishrock123 I'm pretty sure the spec does not dictate to "implement" a noop. I'd certainly prefer to have unimplented stuff as undefined, which is also the strategy browsers follow regarding new API.

[cjihrig]

Also -1 for adding a bunch of noops. Seems like it would be confusing for users to call functions and have nothing happen.

[evanlucas]

-1 from me.

[silverwind]

I think we'd be fine with this PR if it's reduced to just adding the aliases.

[aks-]

I have made a PR doing just that then #3486 And we can discuss it there.

[Fishrock123]

Implementations MUST use a proxy implementation to ensure that calling unimplemented methods will never throw.

@silverwind If this doesn't mean implement no-ops, I don't know what it does mean?

See the top of https://github.com/DeveloperToolsWG/console-object/blob/master/api.md

[targos]

And the message of the commit that introduced this sentence doesn't help much:
DeveloperToolsWG/console-object@cb3d52e

[Fishrock123]

cc @bkardell who appears to have written it.

[silverwind]

I don't think it means doing a noop, but aliasing to console.log.

[Fishrock123]

@silverwind not everything logs there, though. I.e. What should console.clear() do in that case?

[aks-]

Also the signature in our doc doesn't follow the the signature given in console spec. We also have to change that?

[jasnell]

Given the -1's, the lack of activity, and precedence on similar PRs, closing this one.