Backport: fs.mkdtemp to v4.x

[indutny]

• Version: v4
• Platform: OS X
• Subsystem: fs

I suggest that we should backport fs.mkdtemp to v4 LTS. These are commits that introduced it (and fixed it several times):

• e5f8a6a
• 2ccba1f
• acc1142
• 2ef8359
• dcbf246
• 79a5eb1
• c4fadbc

cc @nodejs/lts

[indutny]

cc @Fishrock123

[MylesBorins]

@indutny would this be semver minor?

[indutny]

Yep.

[MylesBorins]

@nodejs/lts what do people think about this? If we can get decent consensus we may be able to avoid having to do this in an official meeting

[rvagg]

@indutny why? can you help us justify why this is important enough? We were just yesterday discussing how v4.5 may be our last semver-minor bump on v4. We still have time left on Active LTS so it's not out of the question but we need pretty strong justification for each semver-minor I think.

[indutny]

@rvagg this is not necessary, but it is hard to see why it hasn't been done before. This API was provided by libuv since the initial release of v4, so there is no reason to not expose it.

Right now, I had to disable v4 on travis tests for some of my modules that use this.

[rvagg]

OK, worth a discussion anyway, as I said in nodejs/Release#121 I don't think we even now have a clear idea of what the rules are for what gets into LTS, I was hoping for principles to emerge a bit quicker than this through what we've been doing but that hasn't been the case.

[ChALkeR]

Hm. This change (with all the commits merged) looks purely additive and rather small.
The actual change looks good.

The timing is bad, though — 4.5.0-rc.1 is already out. Is this targeted to 4.6?

[indutny]

@ChALkeR this is just a suggestion, there is no decision on target version.

[jasnell]

The @nodejs/lts WG discussed the possibility of backporting this to v4.x and decided against it due to lack of a pressing justification.

[jasnell]

@indutny ... going to go ahead and close this given the LTS discussion yesterday. Can revisit if necessary.