NH-3905 - Implement Async operations

[fredericDelaporte]

NH-3905 This is a rebase of #588.

For eliminating the different directions async future have taken before (and for easing the rebase), most future related changes have been squashed into first @maca88 commit.
For allowing to generates async files and compile most commits individually, most async generation setup commits have been moved prior to first maca88 commit. This means most upgrade async generator commits are squashed and the generator version is 0.3.6 from almost "the beginning" of this rebase.
Some additional squashing have been made (especially changes reverted some commits later).
Most "regen async" commits have been regenerated with 0.3.6 async generator version.

[hazzik]

@fredericDelaporte I think we need to move "Regenerate Async" to the latest commit. Made in one step.

[hazzik]

I did further combination of commits and moved async generation to the top.

[fredericDelaporte]

I think we need to move "Regenerate Async" to the latest commit. Made in one step.

Thought the same at first but this means previous commits cannot compile without first generating async files, modifying the tree. It may not be needed to be able to compile previous commits but I still rather avoid letting many non compilable commits in history. Maybe just one regen right after remaining maca88 commit would have been enough for this purpose though.

[hazzik]

Then we, probably, need to clash all these prep commits into one.

[fredericDelaporte]

Moved most prep prior to async generation commit, done some additional squashing, dropped syntax workaround for async gen no more needed, cleaned some undue added namespace, split maca88 commit between prep and what is requiring to already have async files generating (Linq & Future).

This yield an intial async generation, followed by changes for exploiting it for linq & futures, followed by a regen.

All commits compile separately.

Some of your dropped partial on the two test files under Async folder are back: I do not know why there were removed in your rebase, because when I generate async on your rebase, it does add them back. Anyway, undue partial in some generated code is not a concern I think.

[hazzik]

All good. I think we might need to move Future refactoring into the future ticket (we had one for 5.0). Or at least mention this refactoring at the ticket.

[fredericDelaporte]

I think we might need to move Future refactoring into the future ticket (we had one for 5.0). Or at least mention this refactoring at the ticket.

Going to mention that in NH-3905. This refactoring (IFutureEnumerable instead of IEnumerable) was really needed for async case, a bit less without it. And the previous refactoring (NH-4030) was mainly about Linq Future implementation, not its overall implementation.

[fredericDelaporte]

Whitespace miss probably introduced during rebasing. I am anyway rebasing again, while rewording commits for including Jira number.

[hazzik]

I'm going to merge as soon as tests pass. Any objections?

[fredericDelaporte]

Ok for me, going to do that myself in a couple of minutes if you do not seem online.