TL;DR

Evolving Data::Resolver

While trying to move on with PDF::Collage, I figured that the next step would be to fully adopt Data::Resolver’s new OOP interface, get rid of all the overloading over one single sub call as in the previous iteration, and finally embrace the new asset-oriented interface.

Such good intentions made a big loud noise while crashing against reality.

It turns out that the directory-based interface was pretty much to the point, while the TAR-based had some restrictions that surely were attractive at coding time, but proved to make the interface cumbersome while actually using it.

What? Well, asking for a key that has a hierarchy in it (like foo/bar.txt, which has a directory/container part, as well as a file/content part) is processed fine in the directory case, but it was stopped in the TAR alternative.

This is no more the case in the newest trial release, though!

Just as an added bonus, I also decided to force a Unix syntax to all hierarchical keys in the directory case. It might seem arbitrary, but I think that it’s probably something that helps with portability as keys are meant to go beyond the system. Plus I like Unix much better and the TAR interface sticks to it anyway.

So there we go, Data-Resolver-0.005-TRIAL is out, and I’m crossing fingers seeking feedback from CPAN Testers!

Stay safe!