feat(http): add transfer cache module from nguniversal/common

[CaerusKaru]

• Adds TransferHttpCacheModule from nguniversal/common to avoid
  duplicate HTTP calls between server and browser
• Bump RxJS dep to 5.5.5 to avoid issues with lower patch versions
• Clean up certain areas of code, remove redundant imports

PR close #509

[CaerusKaru]

cc @vikerman @Toxicable

[Toxicable]

Looks good aside from the linting changes, revert those for now.
If you want to fix that up another PR adding a tslint would be great, the default one produce by ng new will do

[CaerusKaru]

@Toxicable on it!

[CaerusKaru]

#511 created to deal with linting

[m98]

Nice job 👍

I think it would be good if you can provide an example of it too
Maybe a new component for HTTP request representing and how does cache work.

[CaerusKaru]

@Toxicable thoughts on @m98's comments? Personally I don't see the usefulness of giving an example for TransferHttpCache since it's pretty self-explanatory, and is mostly here to give a leg-up to developers getting started. But I'm open to adding something if you think it would be useful.

@m98 I still think it would be invaluable if you could provide another example to the repo for demonstrating usage of the TransferState feature, just without HTTP.

[m98]

I did not mean adding an example to demonstrating usage of the TransferState

But since this pull request adds ability to transfer state from server to browser, it might be helpful to use a simple HTTP request just let users know there is already something which is supporting transfer state.

They might write their own services to support this if they don't know such thing added to the repo

[Toxicable]

I think adding an example for each would be the better path for this one, so that we can compare and show the different situations you'd use either

[Toxicable]

uhhhh we shouldn't be ignoring this file.
We should either use Yarn or Npm and stick to just one.
I vote Yarn :P

[CaerusKaru]

Fixed

[Toxicable]

this change is unneeded

[CaerusKaru]

Fixed