[CLOSED] Implementation of beforeAll/afterAll

[core-ai-bot]

Issue by TomMalbran
Friday Jul 26, 2013 at 00:30 GMT
Originally opened as adobe/brackets#4581

---

This pull request adds the functions beforeAll and afterAll to be used inside any test suite. With these functions we can better create a test window for a complete test suite and close it at the end. Hopefully, if jasmine finally implements them, we won't need to rewrite anything besides removing the additional code added on this PR.

I then updated the test suites that used a single window for all the tests to use this new functions.

---

TomMalbran included the following code: https://github.com/adobe/brackets/pull/4581/commits

[core-ai-bot]

Comment by peterflynn
Friday Jul 26, 2013 at 05:29 GMT

---

Nice! One quick comment: "before/afterAll()" sound easily confusable with "before/afterEach()" to me. What about "beforeFirst()" and "afterLast()"?

[core-ai-bot]

Comment by TomMalbran
Friday Jul 26, 2013 at 05:46 GMT

---

"beforeFirst()" and "afterLast()" sound better. I went with "before/afterAll()" because I looked like the names jasmine could use to name the APIs, but it doesn't seem that they will add this functions to jasmine, so we could go with different names.

[core-ai-bot]

Comment by jasonsanjose
Tuesday Jul 30, 2013 at 20:15 GMT

---

Reviewing

[core-ai-bot]

Comment by jasonsanjose
Tuesday Jul 30, 2013 at 21:02 GMT

---

Initial review complete. I agree with@peterflynn on the naming. Thanks for working on this cleanup@TomMalbran!

[core-ai-bot]

Comment by TomMalbran
Wednesday Jul 31, 2013 at 02:26 GMT

---

@jasonsanjose Fixes pushed.

One think to notice is that when to chained suites have before or after all functions, the child suite functions will be called first, since we iterate from the spec parent to i'ts parents. Maybe this behavior could be reversed for the before first functions, having the outer most being called first and inner after.

[core-ai-bot]

Comment by jasonsanjose
Wednesday Jul 31, 2013 at 18:44 GMT

---

@TomMalbran It's not an issue with the current usage, but I agree I think that running the highest ancestor beforeFirst callbacks before the lower ones is a good change to make. I'll merge this now and we can follow up later.