Working on 1.0.0

[rvagg]

Someone has to do it!

[Fishrock123]

🎉

[dshaw]

👏👏👏

[bnoordhuis]

LGTM :-)

[fundon]

⭐

[cjihrig]

Thanks Rod! Landed in 8a0e7d6

[jaw187]

[dougwilson]

So process.versions.node === 'v1.0.0'? Any way to tell the runtimes apart?

[rvagg]

I was thinking that we should add a process.versions.iojs property but perhaps it's best to take the same approach as in browser-land where you feature detect rather than platform detect?

[dougwilson]

I'm not saying I would be using it (though presumably npm would use it for engines matching in the package.json file, but that's another thing), but of course if it's there, people will use it. Can't image what would then happen if there was a Node.js 1.0.0 and it's wasn't the same ;)

[cjihrig]

I am in favor of adding process.versions.iojs if for no other reason than to identify node vs. io.js when someone opens an issue and doesn't know what they are running (I'm sure it will happen eventually, even with an iojs binary). Feature detection would be preferred as @rvagg said. If no one is strongly opposed, I'll open a PR.

[rvagg]

the only ones with strong opinions on this I can imagine are @isaacs and perhaps @mikeal

[isaacs]

\o/

[rvagg]

@isaacs is that a +1 for an process.versions.iojs or just general excitement about 1.0.0?

[isaacs]

@rvagg Just general excitement about 1.0.0.

Adding process.versions.iojs is unnecessary and I am opposed to it. Would that version ever !== process.version?

When do you need to feature-detect apart from the version number? Does anyone actually think that there will be a release of joyent/node that is 1.0.0, and not composed of this exact code and this exact set of people doing the release?

[algesten]

Does anyone actually think that there will be a release of joyent/node that is 1.0.0, and not composed of this exact code and this exact set of people doing the release?

OT: @isaacs dropping the bomb? sounds almost like you don't believe in the work you do on the NAB? ;)

[quantizor]

process.runtime === 'iojs' ?

[yorkie]

btw, in process.config it has many keys that prefixes with node like node_install_npm, we maybe remove the prefix completely?

[yorkie]

and i think this should seem to be applied in joyent/node as well :(

[ruimarinho]

My only concern is how npm will be able to test the engines property as pointed out by @dougwilson. We can't expect joyent to keep 0.x forever.

[isaacs]

@algesten I do believe that the JNAB is doing good work, and will get to a good place. But the place it'll get to is merging io.js into node, not creating a competing 1.0.0 release.

The post-io.js node will likely be 1.next or 2.0.0.

The version alone will thus be sufficient to differentiate.

[othiym23]

@isaacs I think process.versions.iojs is useful for tooling that needs to know whether it's working with node or iojs, like node-gyp (see nodejs/node-gyp#564, which you may have other opinions about as well).

As a result, I put together PR #491. As I'm sure you know, I share your desire to see Node and io.js converge again ere long, but until then, having a simple way to tell them apart from inside the runtime is useful.