Laziness in action!

[ALJCepeda]

Lazy chained methods:
step() - Consumes current method in chain
stepBackward() - Decrements then consumes method in chain
walk() - Invoke current method then increment chain
walkBackward() - Decrements then invokes current method
resetChain() - Goes to beginning of chain
endChain() - Goes to end of chain
invokeAll() - Evaluates result of entire chain
invokeRest() - Evaluates result of remaining methods on chain relative to current position in chain
hasStep() - Checks if any methods can be consumed from current position in chain
countSteps() - Returns how many methods can be consumed from current position
hasStepBackward() -> Checks if any methods can be consumed by decrementing chain
countBackwardSteps() -> Returns how many methods can be consumed by decrementing chain
hasChain() - Checks if any methods are available to be evaluated.
countChain() - Returns how many methods are available to be evaluated

Lazy iterator:
Type: Dollar.Iterator
next() - Invokes action on next element in collection (current then increment)
previous() - Invoked action on previous element in collection (decrement then current)
cycle() - Invokes action while cycling through elements in collection
cycleBackwards() - Invokes continuously while decrementing through collection
toFirstObject() - Moves position of iterator to first object
toLastObject() - Moves position of iterator to last object
endIterator() - Moves position of iterator to final position (next will return nil from there)
hasNext()
countNext()
hasPrevious()
countPrevious()

Made changes to reflect new syntax in Beta3 release
Made changes to flatten() and initial() as they didn't work as intended for me.

[ALJCepeda]

I was unable to do create or perform any tests on my beta. I've done all my tests in playground and committed so you can see a demo of my laziness.

[ALJCepeda]

Edit: Added tests to project

[ankurp]

Thanks for the Pull Request. Code looks good but had few comments.

Let's remove the playground file and just have examples in test and README.

Also was there a reason to change T[] to [T]?

Will check the compile issue as it was in working condition before.

[ALJCepeda]

As I said in my second comment, I could not get the tests working on my beta. The rare times when it DID compile, it was riddled with errors. Maybe due to the changes in beta3? I chose to implement my tests as a demo in playground in the meantime.

Changing T[] to [T] is to reflect the syntax changes made in beta3. Also for for i in..myArray now must be for i in...myArray as .. is now reserved for half range specifiers.

[ankurp]

I see what you mean. I downloaded Beta 3 and fixed the compile error and warnings and merged that. #73

You can pull from master again so only your lazy eval changes show up in this PR. Lets remove the playground files for now from this PR as the test not being able to run is a Xcode Beta 3 issue. I will report this bug to Apple so its fixed by Beta 4.

[ALJCepeda]

I only see syntax changes made to the latest commit. If you run the tests you'll see many of them fail, and crash with errors such as BAD_ACCESS and SIGABRT. In my latest commits I've commented out the tests that fail/crash. It was about of the tests if I remember.

[ALJCepeda]

After parsing the tests I've been able to implement my own tests. Indexing in Xcode is very slow, so anybody opening the project for the first time will probably have to wait a while before it'll compile, but my current commit has all the offending code commented out. Playground's been deleted as well, so this should be good to complete the pull request I hope.

[ankurp]

There are some playground files I still see in the code to be merged.

For me compiling takes a long time for Tests but I get no error. Can you pull all of the code from master one more time if you haven't to make sure your changes are in sync.

I have noticed compile errors if you just use Xcode project so maybe compile using the Xcode workspace.

[ankurp]

After you remove those I can pull your changes on my machine and confirm that everything is good before merging.

[ALJCepeda]

Ok I think that should do it. More tests were able to pass but not all of them.

[ALJCepeda]

So it turns out the issue I'm having has to do with AnyObjects and Generics. Put it simply, having chained methods that store AnyObject doesn't play nice with complex Swift Types. I changed the class to use generics exclusively, which solved that issue but some methods like flatten() still remain dysfunctional i think because of it's reliance on NSArrays to get the job done.

Changing generics also broke the iterator, and it'll have to be redone. All my tests involved in chaining was done through an iterator so...none of my tests will work now :(. Still this commit allows you to chain complex swift types, which I think is more desirable than an iterator at the moment.

[ALJCepeda]

Also compiling and running tests is faster now.

[ankurp]

Please remove this file

[ankurp]

@ALJCepeda Thanks for contributing but it seems you are trying to do a lot of things in this pull requests. I would suggest we only make changes to add an array where we push the lazy eval closure function and add iterator methods. Lets not try to change a lot of things as it is leading to existing tests to fails and we do not know the exact cause.

Also you are changing the way these helper methods get called where type needs to be specified. For example, Dollar<Any>.first([1, 2, 3]) This is just a lot of unnecessary typing and type should be inferred.

[ankurp]

The compiler is also very buggy that I am having hard time trying to run the test as swift and SourceKit processes are using up 100% CPU and hogging all of the memory.

[ALJCepeda]

You're right I'm trying too many things at once and I've made a mess of this pull request. I apologize. I've gone ahead and taken a step back and just added lazy evaluation of chained methods. I've deleted out all the commented code, but have left the tests that won't work on my system commented (such as flatten())

I had to take chaining methods out and put it into its own class called Bank. The reason I did this was because I could not get the Dollar to work with with complex Swift types. It's very possible there's something I'm misunderstanding so if that's the case then I apologize again. but using Generics instead of AnyObject was the only way to get it to work.

Please let me know what you think

[ALJCepeda]

If you think this is something you can use, I'll merge and commit with your Ok.

[ankurp]

This causes an error in Beta 3. You will need to use ... or ..< notation

[ankurp]

I implemented the changes in this PR #76