Misc dependency rake improvements

[ZacSweers]

See individual commits!

[github-actions]

1. In the Usage section, it would be helpful to explain what the command does and what the expected output is. This will provide more context for users who are not familiar with the tool.

2. In the Implementation section, it would be helpful to provide a brief overview of the DependencyRake.kt file and its purpose in the project.

3. In the Usage section, it would be helpful to provide an example of how to run the aggregateMissingIdentifiers task and explain what it does. This will help users understand how to use this task and what its purpose is.

4. In the DependencyRake.kt file, the missingIdentifiersFile property should be annotated with @OutputFile instead of @InputFile, since it is an output file.

5. In the DependencyRake.kt file, the missingIdentifiers parameter in the rakeProject function should be passed by reference (MutableSet<String>) instead of by value (Set<String>). This will allow the function to modify the set and add missing identifiers to it.

6. In the DependencyRake.kt file, the mapIdentifier function should return a nullable Coordinates object instead of a non-null Coordinates object. This will allow the function to return null when an unknown identifier is encountered, instead of throwing an error.

7. In the DependencyRake.kt file, the toDependencyNotation function should take an additional parameter missingIdentifiers: MutableSet<String>. This parameter should be passed to the mapIdentifier function to add missing identifiers to the set.

8. In the DependencyRake.kt file, the toDependencyString function should take an additional parameter missingIdentifiers: MutableSet<String>. This parameter should be passed to the toDependencyNotation function to add missing identifiers to the set.

9. In the DependencyRake.kt file, the aggregate function of the MissingIdentifiersAggregatorTask should use sortedSetOf() instead of toSortedSet() to ensure consistent ordering of the aggregated missing identifiers.

10. In the DependencyRake.kt file, the MissingIdentifiersAggregatorTask should be registered only if the slackProperties.enableAnalysisPlugin is true and the project is not the platform project. This will prevent the task from being registered multiple times and avoid unnecessary configuration.

11. In the StandardProjectConfigurations.kt file, the rakeDependencies task should be registered only if the slackProperties.enableAnalysisPlugin is true and the project is not the platform project. This will prevent the task from being registered multiple times and avoid unnecessary configuration.

12. In the StandardProjectConfigurations.kt file, the aggregator task should be configured to use the rakeDependencies task's missingIdentifiersFile as input. This will ensure that the aggregator task runs after the rakeDependencies task and collects the missing identifiers from all subprojects.

[autonomousapps]

Should I make a comment about the lack of tests? :D

[autonomousapps]

--no-configuration-cache so sad, I should fix that.

[autonomousapps]

I'm curious about this, can you elaborate?

[ZacSweers]

Basically we have a lot of transitive deps that DAGP offers as "add" advice but don't exist in our version catalogs, so this is intended to help make setting those up easier

[autonomousapps]

Also just curious about these magic strings.

[ZacSweers]

Just debug logging!

[autonomousapps]

missing identifiers just means missing from your version catalog, right?

[ZacSweers]

yup!