lib: remove unused modules

[Trott]

Some files in lib were using require to load modules that were
subsequently not used in the file. This removes those require
statements.

[jbergstroem]

CI: https://ci.nodejs.org/job/node-test-pull-request/1053/

[Trott]

ppcle-ubuntu1404 failure looks like Jenkins/build problem.

[Trott]

CI looks good other than aforementioned ppcle-ubuntu1404 issue that seems unrelated.

[jbergstroem]

It's 99.999% my bad; currently rewriting init scripts in upstart so they respawn if they go down. Not sure why the iojs user would have permission issues in ~/ though.

Edit: DIAGFILE=${HOME}/jenkins_diagnostics.txt -- guessing HOME is unset.
Edit2: fixed.

[bnoordhuis]

LGTM

[cjihrig]

LGTM. Any reason we aren't linting for unused variables?

[targos]

@cjihrig we would need the varsIgnorePattern option to ignore common in tests. Unfortunately it doesn't exist yet in our current version of ESLint.

[Trott]

@cjihrig Linting for unused variables would require us fixing 700+ instances of the issue in the code. The ones I've submitted are uncontroversial ones, but some might result in some debate and discussion, so I'd like to do all the straightforward ones like this one first so the discussion is around the 70 or 80 (guessing) potentially-controversial ones and doesn't hold up the hundreds of ones that should sail through.

Here's an example of one that might result in a prolonged discussion: There are three unused vars in src/node.js including id at

node/src/node.js

Line 953 in d847a74

NativeModule.isInternal=function(id){

.
On the one hand, id isn't used in that function. On the other hand, removing it from the function signature may make it less clear that it is an analogue to the function with the same signature at

node/src/node.js

Line 961 in d847a74

NativeModule.isInternal=function(id){

. I personally think we should remove the unused variable and name the function if we want to make the analogous nature clear. But that's not quite as straightforward as "the variable in this var statement is never used, so I'm removing it".

[cjihrig]

The no-unused-vars rule also has an option to not check function arguments. That would help with at least some of the cases. But, yea, not being able to upgrade ESLint is a pain.

[Trott]

For common, we can just change var common = require(...) to just require(...). In other words, require it without assigning it to a variable.

[mscdex]

Maybe I missed it but, what is the reason why we can't upgrade our copy of eslint?

[Trott]

@mscdex See #2286.

[silverwind]

@mscdex Basically the way we did indentation is not compatible with the latest versions and to fix that we'd have to introduce around 174 lines of "churn" in lib which some people opposed for git blame reasons.

[Trott]

Landed in fdeb862

[MylesBorins]

@Trott it looks like these changes are breaking v4.x-staging. Would you be able to backport the changes and open a PR against v4.x-staging?

[Trott]

@thealphanerd#4683