fixing failing test on Ternary Search Tree

[pbodsk]

The problem was that the testdata generator could sometimes (but not on every test run...) generate keys that was not unique...which caused the find method to return the wrong values for keys when more keys existed in the tree with the same value.

[vincentngo]

I was wondering why the test were sometimes passing and sometimes failing randomly :D

[vincentngo]

@pbodsk would it be possible to build the TST once and use that tree to perform your test? Instead of creating it on every test?

[vincentngo]

You could extract testStrings and treeOfStrings into properties of XCTestCase, and perform a setup to generate the tree. That way each test focuses on what it's testing.

[pbodsk]

I see you point but...

The first test: testCanFindStringInTree, generates a tree containing String elements so therefore the testStrings and treeOfStrings are created there, like so:

    var testStrings: [(key: String, data: String)] = []
    let treeOfStrings = TernarySearchTree<String>()

The second test: testCanFindNumberInTree, generates a tree containing Int elements, so here the variables used looks like this:

    var testNums: [(key: String, data: Int)] = []
    let treeOfInts = TernarySearchTree<Int>()

(notice that one tree works on String elements and the other works on Int elements).

If all tests were using the same tree structure then I agree that it would be way better to generate/populate the treeOfStrings and the testStrings in setup(), but now the two tests are using different trees for their testing, and I think adding them as two different properties on the testcase would clutter the testclass.

If we add more tests at a later point then perhaps we should consider extracting treeOfStrings into a property and then do tests only with String elements.

Hope this makes sense to you (not trying to dodge the task here 😄)

[vincentngo]

Ah ok fair enough XD