feat: value class support

[qoomon]

No description provided.

[Raibaz]

Looks like all the builds running on 1.3.72 and 1.4.30 are failing because maxOrNull has been introduced in 1.5.0.

I'd like to maintain backwards compatibility with both of them for a little while now - and i think reverting the apiVersion/languageVersion changes should fix this.

I know it gives a warning in 1.5 mentioning that apiVersion 1.3 is going to be removed, but I'd rather bear with the warning for now instead of breaking backwards compatibility.

What do you think?

[qoomon]

I'll try this. However how to run test with api version 1.3 and the new value class only available in api version 1.5?
I could use inline class however this is a kinda bad test than. What do you think?

[codecov-commenter]

Codecov Report

Merging #633 (fe07e44) into master (bd0a0c8) will decrease coverage by 0.48%.
The diff coverage is 4.76%.

❗ Current head fe07e44 differs from pull request most recent head c4d1bd7. Consider uploading reports for the commit c4d1bd7 to get more accurate results

@@             Coverage Diff              @@
##             master     #633      +/-   ##
============================================
- Coverage     75.82%   75.33%   -0.49%     
+ Complexity      819      817       -2     
============================================
  Files           100      101       +1     
  Lines          2825     2846      +21     
  Branches        445      448       +3     
============================================
+ Hits           2142     2144       +2     
- Misses          503      522      +19     
  Partials        180      180              

Impacted Files	Coverage Δ
...lin/io/mockk/proxy/jvm/advice/ValueClassSupport.kt	0.00% <0.00%> (ø)
...n/io/mockk/impl/recording/states/RecordingState.kt	91.39% <ø> (ø)
...kotlin/io/mockk/impl/verify/VerificationHelpers.kt	80.00% <ø> (ø)
...in/kotlin/io/mockk/proxy/jvm/advice/Interceptor.kt	85.71% <100.00%> (-14.29%)	⬇️
...c/main/kotlin/io/mockk/impl/stub/StubRepository.kt	66.66% <0.00%> (-8.34%)	⬇️
...oxy/jvm/transformation/FixParameterNamesVisitor.kt	77.77% <0.00%> (+3.70%)	⬆️
...lin/io/mockk/impl/platform/JvmWeakConcurrentMap.kt	56.00% <0.00%> (+4.00%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update bd0a0c8...c4d1bd7. Read the comment docs.

[qoomon]

It's working. I've revered maxOrNull() to max() and suppress the compile errors by using @Suppress("DEPRECATION_ERROR")

[peterfigure]

Thanks for this PR, I've been trying it out locally and it's working well, really hope this gets merged soon! 🙏

[peterfigure]

Any way to move this PR forward? Very hard to release projects based on Kotlin 1.5 without this 😅

[Raibaz]

It's currently still failing because the mockk-android module is lacking the ValueClassSupport class, which in turn is making a lot of tests fail.

@qoomon, do you think we could move it to mockk-common, so that it is also visible to the android module?

[qoomon]

@Raibaz it is not that easy cause ValueClassSupport depends on kotlin.reflect.jvm.isAccessible and java.declaredMethods (this is just needed for kotlin <1.5 polyfill isValue method)

[qoomon]

@Raibaz why is mockk-agent-jvm excluded within mokk-android build.gradle

   api(mockKProject, {
        exclude group: 'io.mockk', module: 'mockk-agent-jvm'
    })

[Raibaz]

It's excluded because android has its own agent.

As a possible workaround for this, how about providing a dummy implementation of ValueClassSupport for mockk-android, that basically behaves as the polyfilled version for kotlin <1.5?

This will not provide support for value classes in Android, but it would still allow us to merge this.

What do you think?

[qoomon]

Sounds reasonable. Is there an easy way to test if it works without the need of a full android environment?

[qoomon]

I did some adjustments so it should work now, however I'm not sure if the changes within mockk-agent-android are working because I don't have the right environment

[Raibaz]

Thanks a lot for taking care of this!

Unfortunately, Github requires manual approval for first-time contributors, so I have to manually trigger the builds which are, effectively, the fastest way to make sure everything works on Android.

Sorry about this :(

[qoomon]

it seems all test are passing. :-D

[Raibaz]

Thanks a lot again!

[pawelsmagala]

@Raibaz when will you make new mockk release with this PR included?

[Raibaz]

I have a couple of PRs pending that I'd like to merge to release them all together, hopefully by the end of this week or in the next week.

[peterfigure]

@Raibaz any news on the release including this PR?

[antondudakov]

ship it 🚀

[hrach]

Guys, if you wanna help you may use prebuilts via jitpack.io. I am doing so and just discovered an issue, see latest PR. These comment won't speed this up.

[schernetsky]

With this change, I'm seeing unboxing enter an infinite loop (inside every block) when mocking methods returning nullable value classes.

On Kotlin 1.5.10 with MockK 1.12.0

Opened issue: #656