using ratchetFrom in a branch fails to detect the previous index in origin/main

Hi,

I’m using maven (wrapper) 3.8.4 with spotless-maven-plugin 2.27.1 with the configuration defined as follows

<configuration>
	<ratchetFrom>origin/main</ratchetFrom>
	<java>
		<palantirJavaFormat>
			<version>xxx</version>
		</palantirJavaFormat>
		<removedUnusedImports/>
	</java>
	<upToDateChecking>
		<enabled>true</enabled>
		<indexFile>${project.basedir/.spotless-index</indexFile>
	</upToDateChecking>
</configuration>

So I have been working on main for a while, and it already has a .spotless-index file with the checksum and the listing of all the files. So far, no issue with that. But recently, I created a branch off my main called feature-a, and after the first commit in feature-a branch, I ran spotless:apply, then the .spotless-index file now has a new checksum and only listing files that was changed in feature-a branch.

Imagine in main branch

ajoicjaicjoiajscoijasjc098-
// listing all my source files (say 50 lines)

Then in feature-a branch (after a commit, and ran spotless:apply)

osdjvoidjsoyfboudf0bd= // new checksum
// listing 3 files (those that was changed in this new branch commit)

There was a log indicating

[INFO] Up-to-date checking enabled
[INFO] Fingerprint mismatch in the index file. Fallback to an empty index

I can’t have the full logs pasted as this occurs in an air-gapped repo, so if there is further information required, please let me know.

However, based on one of my previous feature branch, it doesn’t seem to suffer the same issue. So I don’t quite understand if there was something that I did wrong when creating the branch, or some mis-configuration on my end.

Thanks!

[lutovich]

Thanks for reporting this problem and sorry for the delay! I will investigate this week.

[bwgjoseph]

I understand that Fingerprint mismatch ... is to be expected when I change spotless configuration, and I generally agree with the approach as you have explained previously.

But I’m more concerned on how to use spotless when working with others, and what to expect. Given that, this index file can change quite frequently (with the current setup) and whatever tracked changes may frequently get reset, and thus, making the index file pretty much “not as useful” as one might have thought.

I think as you have explained before, fingerprint mismatch should only happen when spotless plugin configuration changes and not any other configuration within pom.xml which should greatly reduce the chance for the index file to keep resetting.

And I still have concern when using with large codebase, if the index file resets, how much impact would it make (as it needs to rebuild the index again each time spotless version gets upgraded - non-configuration change)? And how would it work in a monorepo (maven multi-module) project setup

Thanks for investigating, and answering to my queries. Appreciate it.