Fix handling of v2 hashes

[bernt-matthias]

Added tests first for the mulled container resolvers here https://github.com/galaxyproject/galaxy/pull/18522/commits/5b3c4bddd38d88bd195176be9f20f3e81d5336df

Then fixed (and more tests) here: https://github.com/galaxyproject/galaxy/pull/18522/commits/e7bca837f08345980f240f633921dbcce174345e

See details in https://github.com/galaxyproject/galaxy/pull/18522/commits/e7bca837f08345980f240f633921dbcce174345e

I'm tempted to backport this to 24.1 in order to make this effective in the multi-package-container repo.

How to test the changes?

(Select all options that apply)

• [x] I've included appropriate automated tests.
• [ ] This is a refactoring of components with existing test coverage.
• [ ] Instructions for manual testing are as follows:
  1. [add testing steps and prerequisites here if you didn't write automated tests covering all your changes]

License

• [x] I agree to license these and all my past contributions to the core galaxy codebase under the MIT license.

[bernt-matthias]

I only remove it from the function that parses targets from a string (comma separated list of tools and versions) which is used in build-hash and the other two tools.

The remaining code (e.g. the mulled container resolver etc) keeps the build as part of the version. I think that the tests which I added in my first commit show this (the version part of the hash differs when builds are added, i.e. the build is considered as part of the version when computing the hash).

Wondering if this got a bit confusing when we merged Target and CondaTarget https://github.com/galaxyproject/galaxy/commit/d65721c78de7adf9ad749a0e07d0dc7ad8684c2a (guess for conda we parse the build).

[bernt-matthias]

Anything else to do here?

[mvdbeek]

If github let's me I will merge this 😆

[github-actions[bot]]

This PR was merged without a "kind/" label, please correct.