https://kotlinlang.org logo
Channels
Powered by
#spek
  • h

    hhariri

    09/27/2016, 12:29 PM
    <!here|@here> We’re considering removing the context function in upcoming versions. We’ll first mark it as deprecated and then eventually remove it. Before doing so, we wanted to see if people are actually using it and would given/describe not cover those particular scenarios? If so, what scenarios are you using it for?
    👎 2
    👍 3
  • h

    hhariri

    09/27/2016, 12:30 PM
    The main reason for removing it is that it’s leading to confusion and most scenarios could actually be covered without it. This is what we’re trying to discuss first though - if we’re missing some special use-case
  • s

    simon.vergauwen

    09/27/2016, 12:31 PM
    Most of the time I find 
    context
    to ambiguous. But if some1 really finds that keyword useful he could just create an alias with an extension function on Dsl, no?
  • b

    bob

    09/27/2016, 12:32 PM
    I’m fine with given/describe
  • h

    hhariri

    09/27/2016, 12:59 PM
    or just use given/describe.
  • h

    hhariri

    09/27/2016, 12:59 PM
    Please feel free to vote for removing or keeping (Thumbs-up remove)
  • s

    simon.vergauwen

    09/27/2016, 1:00 PM
    @hhariri
    The main reason for removing it is that it’s leading to confusion and most scenarios could actually be covered without it.
  • s

    simon.vergauwen

    09/27/2016, 1:01 PM
    Could you argument this a little more? I’m kinda confused to how it could lead to confusion.
  • n

    nstewart

    09/27/2016, 1:02 PM
    We use it, but I agree it could be replaced with 
    given
    . But as long as the team has agreed on which to use, I don’t see how it’s confusing
  • s

    simon.vergauwen

    09/27/2016, 1:03 PM
    RSpec also uses 
    context
    . I’d be fine with just having 
    describe
    like Jasmine or Cedar 2.
  • s

    simon.vergauwen

    09/27/2016, 1:03 PM
    I’m more concerned about the 
    beforeEach
    conversation we had yd 😂
  • n

    nstewart

    09/27/2016, 1:03 PM
    @simon.vergauwen that’s why we chose 
    context
    - because coming from RSpec it seems like the right choice
  • s

    simon.vergauwen

    09/27/2016, 1:04 PM
    I use all 4 mixed together depending on what suits behaviour description the best.
  • s

    simon.vergauwen

    09/27/2016, 1:05 PM
    But I could definitely see why you choose 
    context
    when you come from RSpec like @nstewart mentioned
  • h

    hhariri

    09/27/2016, 2:30 PM
    @simon.vergauwen Because honestly I don’t see what context could be used for that a given or describe can’t. And when you give people too many options to do the same thing, IMHO it adds complexity and reduces discoverability.
  • h

    hhariri

    09/27/2016, 2:32 PM
    Regarding the other conversation, we’re discussing it still, but the solution is to provide a fix that won’t break existing tests. One possible solution is to rename and introduce a new construct.
  • h

    hhariri

    09/27/2016, 2:33 PM
    Spek already has two styles - the original given/on/it, which for instance I personally use, and the describe Mocha/Jasmine style. And we want to keep make sure anything that is introduced works with both.
  • s

    simon.vergauwen

    09/27/2016, 2:35 PM
    I don’t see what context could be used for that a given or describe can’t.
    And when you give people too many options to do the same thing, IMHO it adds complexity and reduces discoverability.
  • s

    simon.vergauwen

    09/27/2016, 2:35 PM
    Spek already has two styles - the original given/on/it, which for instance I personally use, and the describe Mocha/Jasmine style.
  • s

    simon.vergauwen

    09/27/2016, 2:36 PM
    But I find it an odd decision give that statement.
  • s

    simon.vergauwen

    09/27/2016, 2:38 PM
    Then you might just as well kill of 
    describe
    and add a section in the documentation or on the first page. For the people coming from RSpec, Jasmine, Cedar or Mocha you can easily adapt Spek to the syntax your used to.
    Copy code
    inline fun Dsl.describe(description: String, crossinline body: Dsl.() -> Unit) = group("describe $description", body = body)
  • s

    simon.vergauwen

    09/27/2016, 2:40 PM
    I’d be totally fine with that 2. Since Kotlin is flexible enough to allow for extending functionality.
  • h

    hhariri

    09/27/2016, 2:41 PM
    this is about beforeXXX not removing context. Sorry, mixing conversations.
  • s

    simon.vergauwen

    09/27/2016, 2:42 PM
    okey now I’m confused 😂
  • s

    simon.vergauwen

    09/27/2016, 2:47 PM
    What does the given/on/it and describe style have to do with before/after fixtures?
  • h

    hhariri

    09/27/2016, 2:57 PM
    @simon.vergauwen where do you execute the action for which you verify currently when running tests?
  • h

    hhariri

    09/27/2016, 2:57 PM
    if you want we can take this to DM
  • m

    michaelbannister

    09/27/2016, 5:33 PM
    @simon.vergauwen, the junit4 dependency was a leftover from before I upgraded. @verachadw, the spek-junit-platform-engine in conjunction with junit-platform-gradle-plugin seem to cause the junit5 spek engine to run, it just doesn’t find my spec. I see that in your kraph repo you annotate your tests with 
    @RunWith(JUnitPlatform.class)
    – according to the JUnit 5 docs this is a JUnit 4 test runner.
  • m

    michaelbannister

    09/27/2016, 5:34 PM
    Thanks for your replies though!
  • m

    michaelbannister

    09/27/2016, 5:34 PM
    Here’s what I get when I run gradle tests:
1...555657...112Latest