Polish

[izeye]

This PR polishes a bit.

[wilkinsona]

I think "rely" is correct here.

[izeye]

@wilkinsona Thanks for the feedback!

I reverted it in 32c9bbe.

[wilkinsona]

OpenTelemetryConfiguration isn't an @AutoConfiguration so this is fine as it was.

[izeye]

@wilkinsona I don't know exactly what it does behind the scenes, but I saw the following code in the same class, so I thought that a part of the auto-configuration also uses AutoConfigurations.of():

spring-boot/module/spring-boot-tracing/src/test/java/org/springframework/boot/tracing/autoconfigure/zipkin/ZipkinConfigurationsOpenTelemetryConfigurationTests.java

Lines 42 to 43 in 4babe2e

	private final ApplicationContextRunner contextRunner = new ApplicationContextRunner()
	.withConfiguration(AutoConfigurations.of(DefaultEncodingConfiguration.class, OpenTelemetryConfiguration.class));

There are more occurrences like this, for example, as follows:

spring-boot/module/spring-boot-pulsar/src/test/java/org/springframework/boot/pulsar/autoconfigure/PulsarConfigurationTests.java

Lines 73 to 75 in 4babe2e

	private final ApplicationContextRunner contextRunner = new ApplicationContextRunner()
	.withConfiguration(AutoConfigurations.of(PulsarConfiguration.class))
	.withBean(PulsarClient.class, () -> mock(PulsarClient.class));

I also thought that "User" in the withUserConfiguration() doesn't sound right for a part of the auto-configuration.

Are you suggesting that the mixed uses are okay, or that they need to be used with the withUserConfiguration() if it's not an @AutoConfiguration?

[wilkinsona]

AutoConfigurations.of does two things:

1. It sorts all of the passed-in configurations based on their before and after attributes
2. It orders the configurations so that they're processed before any user configurations

In this case, the configuration isn't an auto-configuration so there are no before and after attributes to consider and there's only of them so there's nothing to sort. There's also only one configuration class in the test so the order in which they're processed doesn't matter either.

We're pretty inconsistent about this across our tests. What was there before is fine and what you've proposed is also fine. Given this, I'd prefer to leave things as they were.

[izeye]

@wilkinsona Thanks for the explanation!

I reverted it in 00e2d1b.