I submitted a tiny pull request lately to enable a test. The code below is completely different than that.
My main purpose of forking was to change the code to be more like Dan North's original BDD definitions in his original 2008 introduction of BDD. Aka, the "purist" vision of BDD with Given/When/Then
that follows the overall scenario:
Given some initial context (the givens),
When an event occurs,
then ensure some outcomes.
I also come from a .NET background where I was an avid user of Machine.Specifications (aka MSpec) BDD framework and his patterns. More specifically, his very nicely formatted TEXT and HTML outputs that was completely human readable.
Therefore, I present the following updates to Go's Zen BDD Framework: Full Contextual Scenarios
// bdd_test.go
//
func TestBddSceneriosUsingContextAndSpecifications(t *testing.T) {
Given(t, "a BDD scenario", func(when When) {
when("an event occurs", func(it It) {
it("should evaluate to 1", func(expect Expect) {
expect(1).ToEqual(1)
})
it("should also evaluate to 3", func(expect Expect) {
expect(3).ToEqual(3)
})
it("should perform another evaluation", func(expect Expect) {
expect(4).ToNotEqual(5)
})
it("should also perform another evaluation", func(expect Expect) {
expect("hellow").ToNotEqual("world")
})
})
// common context
count := 0
before := func() {
count++
}
after := func() {
count--
}
setup := Setup(before, after)
when("using Setup() in extended-style", func(it It) {
it("should increment count to 1", setup(func(expect Expect) {
expect(count).ToEqual(1)
}))
if count != 0 {
t.Error("In BDD-specs, count should have been reset to zero by the teardown func")
}
})
})
}
Which outputs the following:
Given a BDD scenario
When an event occurs
It should evaluate to 1
It should also evaluate to 3
It should perform another evaluation
It should also perform another evaluation
When using Setup() in extended-style
It should increment count to 1
PASS
ok github.com/eduncan911/zen 0.008s
Note my new tests that start at the "Given a BDD scenario". You will also notice that I inserted "Given / When / It" in the output.
If you are interested in a Pull request for these changes, let me know. You can have a look at the changes here:
eduncan911@91df917
In short:
- extended existing code and keeping it backwards compatible with all existing Zen projects,
- the only breaking change is any custom formatter(s) will need to implement two new methods,
- broke apart but also grouped Desc from Titles for all existing,
- enabled grouping of output indentions
It remains completely compatible with all existing Zen users. Just the output is formatted differently. If you run all tests, you will the output has been tweaked to group multiple existing it()
funcs:
eduncanmac1:~/go/src/github.com/eduncan911/zen [master +0 ~5 -0]$ go test
When Equality Specs
It should have an integer equal to itself
It should not have any integer equal to nil
When Setup Specs
It should execute before by incrementing count
Given a BDD scenario
When an event occurs
It should evaluate to 1
It should also evaluate to 3
It should perform another evaluation
It should also perform another evaluation
When using Setup() in extended-style
It should increment count to 1
PASS
ok github.com/eduncan911/zen 0.008s
That runs all existing tests before my changes indicating it remains backwards compatible; then, my new scenario specs are run along side them seamlessly.
Well, there is a breaking change: it is for anyone using custom formatters - I extended the interface to have two additional methods:
type formatter interface {
PrintContext()
PrintWhen()
PrintTitle()
PrintError(string)
}
You'll need to implement a PrintContext()
and PrintWhen()
. If anyone really has overwritten the default formatter, I'm sure this will be a welcome upgrade/tweak.
It's also missing documentation which I am happy to add later in another pull request.