Add bucket#makePublic and bucket#makePrivate

[ryanseys]

Do not merge! No tests yet!

This is the implementation to make a bucket public and private. In my preliminary manually testing, it works great!

Optionally a recursive option can be set to true to set all the files in the bucket to public as well. QUESTION: Should this be default true? Also, I'm returning either a single error after the first error occurs if a file cannot be made public for some reason, or if they set force to true then it will add those errors to an array and return that array of errors at the end.

Additionally, I'm returning a list of the file objects that were successfully publicized/privatized.

I'm looking for feedback on general approach and big issues you can see.

Issues: bucket#makePrivate with strict: true always throws an error forbidden before it finishes, because it somehow doesn't recognize my account using service account key json to authenticate as an owner. Any idea how I can become the root owner of the bucket?

TODO: Add options.auto to apply a predefinedObjectACL on the bucket to allow new files to be created with the same permissions.

Depends on #390.

[stephenplusplus]

I know the worst thing someone reviewing your PR can do is be like "here's a whole other bunch of code that renders yours useless!" I didn't intend to let it get to that point- I just wanted to drop the PR in my editor, which makes it easier for me to make sense of. I found some things I thought I could improve by demonstrating as opposed to trying to explain. And, before you know it... https://gist.github.com/stephenplusplus/66bfc7ba04ba97a8a15a

I think either approach is perfectly fine, I just wanted to hold less state outside of each function, and make it more modular in general. As an example, I let bucket#makePrivate() and bucket#makePublic() handle each of their metadata calls, then changed the behavior of _makePublicPrivate to a more focused intent: _makeAllFilesPublicPrivate.

It's not tested (didn't even run it), but if it's something that doesn't look hideous to you, I can send a PR with it to evolve from this one, or you can send me a meme that explains what a jerk I am, and I'll delete the gist and pretend it never happened!

[ryanseys]

Hey this is great! I don't mind one bit because mine was more POC. One thing that is still TODO that needs to be thought of is the options.auto which sets the default object ACL for the bucket as well. I've got a midterm that requires my focus today so can't really care for this PR so I'm 100% ok with abandoning it and letting you run with your gist. We can close this PR, and you can reopen a new one if you want to keep running with this. That's the easiest way forward.

[ryanseys]

Upon closer inspection your async.eachLimit doesn't really make sense. You don't want to run with the nextQuery on each file, only after all files are processed.

[stephenplusplus]

Ooh, good catch. I'll have to think my way out of that one!

[ryanseys]

Actually you're right, I was messing up the callback count.

[stephenplusplus]

Wait:

async.eachLimit(files, MAX_PARALLEL_LIMIT, processFile, function(err) {

Won't that process files (max 10 at a time) until all files are processed, then call processFiles again for the nextQuery?

[stephenplusplus]

We're a mess

[ryanseys]

Yeah my bad... :( I think you'll be good. Or you'll figure it out! 😄

[stephenplusplus]

Going to close this to let #429 carry it on. If I missed anything in the transfer, please update the issue pr.