[melissalinkert]

Our usage of guava now appears to be very limited, so replacing likely makes more sense than continuing to address updates to this dependency.

[sbesson]

The repository tests have been passing with this and ome/bioformats#4499 included in the nightly builds - see https://bf-testing-results.s3.amazonaws.com/2026/2026-11-06/build-status.log.

As demonstrated by the diff, the usage of guava in this component was limited to a few classes and can be easily replaced by built-in APIs available since Java 8.

Defining guava at this level has been the source of downstream issues - see ome/openmicroscopy#6463 and also require a coordinated set of releases to bump the version e.g. in order to include a CVE fix.

Generally, getting rid of guava is consistent with approach taken in #108 to minimize the security surface area of this library and reduce the number of dependencies.