Re:Save the code or save the tests? (2020)
by zeroimpl on Wednesday June 09, 2021 @09:53PM
(#27455856)
Something like:
assertEquals(0, add(0,0))
assertEquals(1, add(0,1))
assertEquals(2, add(1,1))
assertEquals(8, add(3,5))
assertEquals(-3, add(2,-5))
PS: This kind of proves why a test is a lot less useful than the algorithm. Try figuring out how to implement "add" from the above test.
Re:Save the code or save the tests? (2020)
by anonytrary on Wednesday June 09, 2021 @11:37PM
(#27456396)
Personally I see this a lot and I'm not a fan:
it('should work', () => {
inputOutputPairs.forEach(([input, output]) => {
expect(testedFunction(input)).equals(output);
})
})
Tests are way more useful when they are written with the intent of acting as documentation. If I had to give this test a grade (from F- to A+), it would get a D+. It catches regressions, yet does an absolutely horrid job of explaining the intent behind testedFunction.
Re:Save the code or save the tests? (2020)
by ramses0 on Thursday June 10, 2021 @12:34PM
(#27462603)
cases = [
{ in: 1, arg: 2, expected: 3, message: "base case" },
{ in: 2, arg: -2, expected: 0, message: "negative numbers" },
...
]
for ( c in cases ) {
actual = obj.do( c.in, c.arg )
assert( actual == c.expected, c.message )
}
see: http://fit.c2.com/ ... table-based testing, "cases" and "drivers"
Re:Save the code or save the tests? (2020)
by cronin101 on Wednesday June 09, 2021 @06:11PM
(#27453961)
Forall A:
add(A, 0) == A
Forall A B C:
add(add(A,B), C) == add(A, add(B, C))
I think that’s sufficient. Rule 101 of useful testing is understanding the actual constraints you rely on.Maybe throw in add(A, B) == add(B, A) if you’re angling for a promotion…
Re:Save the code or save the tests? (2020)
by postalrat on Wednesday June 09, 2021 @08:05PM
(#27455061)
add(a, b) => a == 1 && b == 1 ? 1 : a + b
Kinda disappointing your test would take so long to run and still catch a simple error.
Re:Save the code or save the tests? (2020)
by milkey_mouse on Wednesday June 09, 2021 @10:51PM
(#27456167)
> your test would take so long to run
I think GP's test would probably be implemented with property testing (or even better, symbolic execution). It wouldn't necessarily take super long to run.
How would you verify that algorithm?