add beet.contrib.mod_export

[jacobsjo]

This adds a plugin to bundle a data pack and resource pack into a single .jar file, including mod identifier files for fabric, neoforge, mcforge, and quilt.

example config:

pipeline:
- beet.contrib.mod_export
meta:
  mod_export:
    fabric: true
    quilt: true
    neoforge: true
    mcforge: true
    name: '{{project_id}}_{{project_version}}' # filename of resulting .jar file without extension
    group: dev.mcbeet # only used by quilt, but then required

    # these are 2 fields since every loader has a different format to specify format ranges
    # if unspecified uses version specified in beet config
    min_minecraft_version: 1.21.4
    max_minecraft_version: 1.21.5

    license: MIT
    contact:
      # any field is allowed here, but only homepage and issues is used for [mc|neo]forge
      homepage: https://mcbeet.dev/
      issues: https://github.com/mcbeet/beet/issues
    authors: # overwriting author from beet config to allow multiple authors listed separately
    - A
    - B
    contributors:
    - C

Things left to do:

• Deal with files that are in both resource and data pack (i.e. pack.mcmeta and pack.png, but also other extra files such as a README)
  • Currently this prints warnings, and is actually resulting in duplicate files in the jar archive; probably leading to hard to debug issues with different tools using different files for the same filename
  • I think this would require changes to core, so I would need some help here.
• Deal with (resource pack) overlays
  • Since there is only one pack.mcmeta, it is unclear how this can work
  • I think fabric uses data-pack version numbers to select resource pack overlays, might be difficult to handle properly
  • I don't know how other loaders handle overlays, if at all

---

I've written this first for my own datapack at https://github.com/jacobsjo/lavaflow/blob/main/src/modded.py ; for the avoidance of doubt: I own full copyright of this file and am re-licensing this file under MIT