dvgaba / easy-random Goto Github PK

I'm the original author of easy-random-protobuf and I just discovered that my code was integrated in this project.

However my work - published under ASL2 - has been re-licensed here to MIT,

https://github.com/murdos/easy-random-protobuf/blob/8e9d20f173b15f312d5ab55d802667489d4b51fa/src/main/java/io/github/murdos/easyrandom/protobuf/ByteStringRandomizer.java#L4C16-L4C16

And I'm not even credited as author of the code: Mahmoud Ben Hassine is credited as the author of the code.

https://github.com/murdos/easy-random-protobuf/blob/8e9d20f173b15f312d5ab55d802667489d4b51fa/src/main/java/io/github/murdos/easyrandom/protobuf/ByteStringRandomizer.java#L2

datafaker, mockito, classgraph - have new versions.

Could you make a release with their latest versions 🙏

When trying to generate random record, it seems that EasyRandomParameters are not used for generation.
Example:

import org.jeasy.random.EasyRandom;
import org.jeasy.random.EasyRandomParameters;
import org.jeasy.random.randomizers.text.StringRandomizer;
import org.junit.jupiter.api.Test;

import static org.jeasy.random.FieldPredicates.named;
import static org.junit.jupiter.api.Assertions.assertEquals;

public class TestParameters {

    private final EasyRandom easyRandom = new EasyRandom(new EasyRandomParameters()
        .randomize(named("name"), new StringRandomizer(1))
    );

    @Test
    void test() {
        var result = easyRandom.nextObject(TestRecord.class);
        assertEquals(1, result.name().length());
    }

    record TestRecord(
        String name,
        String phone
    ){}
}

Result:

Expected :1
Actual   :23

Hi,

Just found a bug regarding records. When you randomize a list of records, it's always returning the same value, so all the objects inside the stream are the same (see attached test). I opened the issue just in case someone is facing the same and can provide a workaround or guidelines to fix it.

This is the test I made:

    @Test
    void typeRecordListShouldBePopulatedWithDifferentValues() {
        List<TestRecord> recordList = easyRandom.objects(TestRecord.class, 2).toList();

        TestRecord rnd1 = recordList.get(0);
        TestRecord rnd2 = recordList.get(1);

        assertThat(rnd1).isNotEqualTo(rnd2);

    }

Found it using version 6.1.7

Given a record like:

public record MyRecord(String fieldOne, AnotherRecord fieldTwo) {}

where AnotherRecord is a record

When I do:

easyRandomParameters.randomize(
            field -> field.equals(fieldTwoField), () -> null
        );

I would expect the value for fieldTwo to be null. This works with most primitives & other classes that are not records. However, with records, fieldTwo is not null and still subject to value generation.

I've sort of worked around this by also adding:

easyRandomParameters.randomize(AnotherRecord.class, new SkipRandomizer());

however if in my object I have more than one field that is AnotherRecord, ie:

public record SomeOtherRecord(AnotherRecord fieldOne, AnotherRecord fieldTwo) {}

naturally both fieldOne & fieldTwo would be skipped, so it doesn't work in those scenarios. The intention is just to null out that single field (fieldTwo).

jitpack runs your build with Java 8

Easy fix

jitpack.yml

jdk:
  - openjdk17
install:
  - mvn install -Dmaven.javadoc.skip=true -DskipTests

Thank you for taking time to open this issue! Please check the known limitations section before opening an issue.

Please provide as much context as possible to help us fixing the issue (at least easy-random version you are using).

• If you are reporting a bug, the best way is to provide a failing test.
• If you are requesting a feature, don't hesitate to explain in detail your suggestion with code examples.
• If you have a question, please first check if there is no (closed) issue about it. You may ask your question on the Gitter channel of the project.

Many thanks upfront!

hi,
I just noticed that protobuf field predicates are not working when specifying custom random values providers.

For example, the following test fails

    void shouldUseCustomProviderRegistryToFillFields(){
        EasyRandomParameters parameters = new EasyRandomParameters().randomize(ProtobufPredicates.named("int32Field"), () -> 42);
        EasyRandom easyRandom = new EasyRandom(parameters);

        Proto3Message protoInstance = easyRandom.nextObject(Proto3Message.class);

        assertThat(protoInstance.getInt32Field()).isEqualTo(42);
    }

I've found a solution and pushed it to my fork, will create PR now because it involves exposing EasyRandomParameters's custom randomizer provider so that we are able to access it. Any suggestion is welcome.

There is a test that is supposed to verify the scenario, i.e. shouldLimitTheNestingLevel_whenInRecursiveStructures, but it was changed from:

        assertThat(actual.children().get(0))
            .as("On the 2nd level, the field should be initialized with empty list, i.e. end of nesting.")
            .isEqualTo(new NestedRecordThroughCollection(List.of()));

to:

        assertThat(actual.children().get(0))
            .as("On the 2nd level, the field should be initialized with empty list, i.e. end of nesting.")
            .isEqualTo(new NestedRecordThroughCollection(null));

When reverting the test, it fails as the list 'in children' is not initialized.