added stencil template files for mocking and randoms generation

[vmalakhovskiy]

No description provided.

[a-voronov]

awesome 👍

but please remove this dependency

btw is it possible to know in stencil what module we're in? like in xcode templates?

[vmalakhovskiy]

unfortunately I didn't found solution for that yet

[a-voronov]

same

[vmalakhovskiy]

unfortunately I didn't found solution for that yet

[PaulTaykalo]

You can pass parameter to sourcery script

[a-voronov]

ah, just wanted to answer the same :)

there's args that you can pass and read inside

or better - setup yml file with all configs and pass path to it via --config flag

so that it'll contain

args:
  testable:
    - Betterme

and read as

<%_ if let testable = arguments["testable"] as? String { -%>
        <%_ %>@testable import <%= testable %>
<%_ } -%>

[a-voronov]

and guess we don't need Result here as it's imported from ReactiveSwift

[PaulTaykalo]

@vmalakhovskiy Can you please write down how are you using these stencils, may be add sourcery script right to the example itself?
Thanks

[vmalakhovskiy]

Actually yeah, I'm using sourcery with config file running in watch mode, so it can detect changes in source code or templates and automatically regenerate code.

sourcery --config .sourcery.yml --watch

my config looks like this:

sources:
  - MyProject
  - MyProjectTests
templates:
  - MyProjectTests/Templates
output:
  path: MyProjectTests/Generated

[a-voronov]

do we need random implementations?
I'd better not make it as a part of this tool, as it doesn't have anything to do with mocking

and there're many good tools that are specifically dedicated to randomization

[adamkaplan]

I tend to agree Random is useful but orthogonal. However, one important note is that it is not a great idea to use pure random tests. Why? Because if a test fails, you cannot reliably reproduce the failure.

Tests should be predictable and reproducible. Proper randomized testing re-uses the same values in the same order each time after (re)generation. This way CI pipeline will always be consistent. You can do that by either saving the actual values used, or by seeding the RNG with the same value (however, not arc4random). Anyway this is a side note.

[a-voronov]

@adamkaplan agree, I tend to use third-party libraries for randomization, i.e. RandomKit - it can be seeded