deps: upgrade to npm 2.14.15

[zkat]

Hi! This PR includes three releases!

• https://github.com/npm/npm/releases/tag/v2.14.13
• https://github.com/npm/npm/releases/tag/v2.14.14
• https://github.com/npm/npm/releases/tag/v2.14.15

Relevant changes:

• Lots of doc updates/fixes
• 6051a69 #10685 Fix npm public registry URL in notices.
• #10629 Node can be compiled FIPS-compliant now due to MD5 extraction (it was previously used for non-security purposes)
• nodejs/node-gyp#831 New node-gyp with better *BSD support
• #10532 Clarify that registry.npmjs.org is the default, but that you're free to use the npm CLI with whatever registry you wish. This goes along with previous licensing-related improvements, too.
• 333e118 Better AIX support for node-gyp

Those are the highlights! And to repeat: This does include some of the previously-talked-about licensing changes.

r: @Fishrock123
r: @iarna

[MylesBorins]

CI: https://ci.nodejs.org/job/node-test-pull-request/1382/
CITGM: https://ci.nodejs.org/job/thealphanerd-smoker/49/

[MylesBorins]

will manually test tomorrow

[zkat]

I may be able to slurp a fix into this that includes the relevant CI fixes -- but this still needs that quick pre-test patch to make it stop being a prerelease :(

[MylesBorins]

@zkat the CI failures appear to be unrelated... looks like windows fanned was down last night, and the single failure on arm is also a CI failure not code. I'm going to re run those two systems again

ARM: https://ci.nodejs.org/job/node-test-commit-arm/1931/

WINDOWS: https://ci.nodejs.org/job/node-test-commit-windows-fanned/1115/

[MylesBorins]

Failure on PPC for citgm is expected

[Fishrock123]

@thealphanerd Keep in mind node-test-commit is irrelevant for npm since it doesn't actually test that code. :)

[MylesBorins]

Oh yeah... Womp.

At least citgm tests a bunch of npm functionality
On Jan 26, 2016 11:53 AM, "Jeremiah Senkpiel" notifications@github.com
wrote:

@thealphanerd https://github.com/TheAlphaNerd Keep in mind
node-test-commit is irrelevant for npm since it doesn't actually test
that code. :)

—
Reply to this email directly or view it on GitHub
#4872 (comment).

[zkat]

Also our test suite still isn't passing on Windows. That'll stay red until we're done with our current project.

[joaocgreis]

I'm trying to get the Windows CI stable, and noticed this failing.

C:\workspace\node-test-binary-windows\RUN_SUBSET\0\VS_VERSION\vs2013\label\win2008r2>git fetch https://github.com/nodejs/node.git +refs/pull/4872/head:refs/remotes/origin/_t

C:\workspace\node-test-binary-windows\RUN_SUBSET\0\VS_VERSION\vs2013\label\win2008r2>git checkout refs/remotes/origin/_t~1
warning: unable to access 'deps/npm/node_modules/node-gyp/node_modules/npmlog/node_modules/gauge/node_modules/lodash.padright/node_modules/lodash._createpadding/node_modules/lodash.repeat/.gitattributes': Filename too long
Checking out files: 100% (9189/9189), done.
Previous HEAD position was af7313c... added binaries
HEAD is now at 39eeca6... test: add test for debugging one line files

C:\workspace\node-test-binary-windows\RUN_SUBSET\0\VS_VERSION\vs2013\label\win2008r2>git checkout refs/remotes/origin/_t
warning: unable to access 'deps/npm/node_modules/node-gyp/node_modules/npmlog/node_modules/gauge/node_modules/lodash.padright/node_modules/lodash._createpadding/node_modules/lodash.repeat/.gitattributes': Filename too long
fatal: cannot create directory at 'deps/npm/node_modules/npm-install-checks/node_modules/npmlog/node_modules/gauge/node_modules/lodash.pad/node_modules/lodash._createpadding/node_modules/lodash.repeat': Filename too long

So, C:\workspace\node-test-binary-windows\RUN_SUBSET\0\VS_VERSION\vs2013\label\win2008r2\'deps/npm/node_modules/npm-install-checks/node_modules/npmlog/node_modules/gauge/node_modules/lodash.pad/node_modules/lodash._createpadding/node_modules/lodash.repeat is the problem... Is there a way to make this filename smaller? I could try to make the test directory name smaller in CI, but not without breaking the parallel tests in Windows, and that would make CI noticeably slower.

[Fishrock123]

@joaocgreis Maybe we should run npm3's dedupe on the npm2 source tree for this, that might help?

[joaocgreis]

@Fishrock123 if that works and makes the tree less deep, it should solve this.

[Fishrock123]

Usual failures with pre-release builds. None with a "release" build. LGTM.

[zkat]

Please don't flatten things -- last we checked, it won't work, and this is all part of the work we're doing right now to get Windows CI stable for npm. It is, in fact, our main focus right now, so you should be seeing a green Windows soon enough.

I've pushed a small patch that -should- get things working on non-windows CI again (it worked locally, so here's hoping! fingers crossed).

[richardlau]

@zkat If npm's LICENSE has been updated please could you keep Node's LICENSE in step? I think this will have to be done manually as the license-builder tool hasn't been backported to v4.x.

[orangemocha]

@zkat any reason not to include 2.14.16 in this PR? It has a couple of important Windows fixes and we were hoping to get it merged into Node soon.

[iarna]

@orangemocha Not to speak for @zkat, but 2.14.16 hasn't spent it's week in pre-release testing (as next-2) yet. It'd be eligible tomorrow though.

[jasnell]

LGTM on cursory scan. At least nothing significant caught my eye.

[zkat]

@richardlau Ahh! Thanks for the heads up. Change is pushed.

@orangemocha As @iarna said -- we have delayed releases. I'll be putting up another PR for 2.14.16 on Friday or Monday, most likely. :)

I'm gonna go ahead and merge this once the LICENSE thing gets a +1

[iarna]

LGTM

[joaocgreis]

This cannot land without doing something about CI, or it'll break. @zkat isn't there any way to make the tree less deep, even if just for that package?