Fix hc dna pack inconsistencies

[jdeepee]

Summary

Fixes #1475 by storing bundle resources as a BTreeMap vs Map. Before when writing out the bundle the resources were written in an inconsistent order, causing the hash of result .dna files to differ.

TODO:

• CHANGELOG(s) updated with appropriate info
• Just before pressing the merge button, ensure new entries to CHANGELOG(s) are still under the UNRELEASED heading

[pegaltier]

awesome!

[freesig]

Functionally this looks fine but the diff is a bit messed up. Maybe pull in develop and I can approve.

[jdeepee]

@freesig Thanks, develop pulled in

[lucksus]

AD4M is currently blocked with regards to updating to a recent Holochain version because of this. Templating AD4M Languages and the Holochain DNAs in them is quite a core aspect and used to verify Language source code when an agent installs a new Language. Having hc dna pack create different hashes in-deterministically breaks our templating checks.

Wondering what's holding up this one-line fix and just took a look at the failing CI workflow. Could it be failing just because the PR is coming in from an external repo?

[lucksus]

Yes, that must be it. This is a PR with the same changes but from an internal branch, and it passes all CI workflows: #1492