add tests for type validation

[Vankog]

See #1012 "[Bug] withPrefabValuesForField() throws for fields declaring interfaces and super-types".

I was a little too slow to also provide the fix, but I can still provide my tests for it.

[Vankog]

Unfortunately, I had issues setting up the project structure with maven. Do you use a specific version?

Prettifier or Spotless (as in PR template) are also not working.

PS D:\ext_code\equalsverifier> mvn -version
Apache Maven 3.9.9 (8e8579a9e76f7d015ee5ec7bfcdc97d260186937)
Maven home: C:\ProgramData\chocolatey\lib\maven\apache-maven-3.9.9
Java version: 1.8.0_322, vendor: Temurin, runtime: C:\Program Files\Temurin\jdk8-hotspot-8.322.6\jre
Default locale: en_US, platform encoding: Cp1252
OS name: "windows 10", version: "10.0", arch: "amd64", family: "windows"

[jqno]

I've never seen this use of blocks before. Is this a JUnit 5 thing? Will it run the second block even if the first one fails?

[Vankog]

Glad you ask. :-)
No it is simple Java syntax.
While it makes not much sense to use them in production code, I found benefits in using it in test code.

The benefits are the scoped variables. exc is only scoped inside the brackets. Thus, you can reuse the same name in another bracket, without having to fear copy paste errors, because you may have used exc1, exc2 etc. (did happen to me more than once ;-) )

Further, it separates the cases visually.

It is something I came across a few months ago and found suitable use case for this with certain test setups. I think tests are a perfect fit for this if you don't want to use a separate method for each and every check.
i.e. The alternatives are: To use a separate method for each check (which is totally valid).
Or use @ParameterizedTest.
Or to be careful with the var-naming, so you don't mix up vars by accident.

Unfortunately, it does not have any impact on the test execution.
For the use case you mentioned, assertAll()is a suitable method. It executes all given assertions, even if some fail.

[Vankog]

Alright, changed it to assertAll.
for example, running the tests against the old equals code, produces these test failures:

[jqno]

Thanks for the PR and the tests, I'll gladly accept them! If I knew you wanted to contribute the fix yourself, I would have waited 😄

Regarding Prettier, it should just work out of the box. But I admit I've only ever tested it on Linux and macOS, so it's interesting to know that it might not work on Windows. Can you give me the full error message that you get?
Prettier is, unfortunately, a Javascript tool, so it will install node behind the scenes. I'm planning to move away from Prettier soon-ish for this reason, but haven't had the time yet to actually do so.

[jqno]

I see now that I never updated the PR template when I switched to another Maven plugin for Prettier. The command to use (for now) is mvn prettier:write. Hopefully that works for you!

[Vankog]

Thanks for the PR and the tests, I'll gladly accept them! If I knew you wanted to contribute the fix yourself, I would have waited 😄

Don't worry, I was just tinkering. After all it was just a 1-liner. ;-)

The prettier error is:

PS D:\ext_code\equalsverifier> mvn prettier:write
[INFO] Scanning for projects...
[INFO] ------------------------------------------------------------------------
[INFO] Reactor Build Order:
[INFO] 
[INFO] EqualsVerifier | parent                                            [pom]
[INFO] EqualsVerifier | core                                              [jar]
Downloading from central: https://repo.maven.apache.org/maven2/org/codehaus/mojo/maven-metadata.xml
Downloading from central: https://repo.maven.apache.org/maven2/org/apache/maven/plugins/maven-metadata.xml
Downloaded from central: https://repo.maven.apache.org/maven2/org/apache/maven/plugins/maven-metadata.xml (14 kB at 25 kB/s)
Downloaded from central: https://repo.maven.apache.org/maven2/org/codehaus/mojo/maven-metadata.xml (21 kB at 36 kB/s)
[INFO] ------------------------------------------------------------------------
[INFO] Reactor Summary for EqualsVerifier | parent 3.17.3-SNAPSHOT:
[INFO] 
[INFO] EqualsVerifier | parent ............................ SKIPPED
[INFO] EqualsVerifier | core .............................. SKIPPED
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time:  1.360 s
[INFO] Finished at: 2024-11-05T16:25:08+01:00
[INFO] ------------------------------------------------------------------------
[ERROR] No plugin found for prefix 'prettier' in the current project and in the plugin groups [org.apache.maven.plugins, org.codehaus.mojo] available from the repositories [local (C:\Users\dmoench\.m2\repository), central (https://repo.maven.apache.org/maven2)] -> [Help 1]
[ERROR]
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR]
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/NoPluginFoundForPrefixException

I have Node 22.9.0
and npm Prettier also.

I also tried updating the POM with the latest prettier-java.maven version 2.6.5 and updated the <nodeVersion> to match mine. But to no avail.

[Vankog]

OK, I used the npm version prettier-plugin-java directly via npx prettier --write "**/*.java".

However, it changed quite some other files, too. So I am not very confident that the maven version gives the same result. Maybe that's caused by the version difference of Prettier and the plugin.

[Vankog]

He....
I found the issue.
After telling maven to use my JDK21, instead of my JAVA_HOME as runtime, it worked.
JDK 8 is seemingly not supported by some dependency anymore.

The good message is: The file output was the same. So it should be conforming now.
The not-so-good:
I had to adapt the endOfLine config as it conflicted with my git config (checkout CRLF, commit as LF).
But running prettier converts all my line endings to LF again, marking them as touched.
I'll create another PR with my suggestion on how to fix this.

EDIT: #1015

[jqno]

Sorry, I was unable to respond sooner, or I'd have told you not to bother with Prettier. It's such a hassle to set up, with node and the specific version you have to install in order to use it (the most recent version makes a lot of unnecessary changes, so I prefer sticking with the older version for now). But good on you that you figured it out anyway!

Turns out that the build doesn't fail on Prettier, but on Checkstyle, unfortunately it doesn't like the nested scope in the test. Thanks for explaining the reasoning behind that. But I think I agree with Checkstyle on this one. Whenever I do more than one assert per method, I prefer to use assertAll() because that way, a failing test can't hide behind another failing test. I've been bitten too often by fixing an assertion, only to find out that the next assertion in the test also fails but for a different reason.

[Vankog]

Sure, I'll update it tomorrow.

[jqno]

Thanks for these tests! I might tweak them a little bit further to fit in more with the other tests, but they're certainly helpful.