feat: add encoder/decoder for withdrawal id compression (for storage in OP_RETURN)

[cylewitruk]

Description

Part of: #1478 and supersedes #1486 (due to the size of the changes and the fact that most comments are now invalidated).

This introduces a new module idpack to the sbtc crate which contains an encapsulated integer sequence compression encoding using a multi-bitmap segmentation algorithm.

Compared to #1486, this PR is substantially slimmed down and drops support for the eventual handling of multiple encodings within the same encoding format, i.e. adding or changing the encodings will require a new implementation.

Changes

• Removed EncodingStrategy trait and moved the full encode/decode implementations to encoder.rs and decoder.rs respectively.
• Removed a bunch of other stuff as well.
• The LEB128 implementation is moved to feat: add leb128 variable length integer implementation to sbtc crate #1523 and this PR targets that.
• The encode-path is now infallible.
• The packaging path remains fallible:
  • The only errors that can be produced (in practice) are due to input validation.
  • The same applies to size-estimations from Segmenter.
  • Reason being that these work on slices instead of requiring new allocs to sortable types (we expect our input to be sorted, anyway - see this comment for more context).
• Added a bunch of tests.
• The Segments type now upholds the invariant that its inner segments must be sequential without overlaps.

Checklist:

• I have performed a self-review of my code
• My changes generate no new warnings
• New and existing unit tests pass locally with my changes
• Any dependent changes have been merged and published in downstream modules

[djordon]

Largely looks good, just a couple of comments.

The base branch was changed.

[djordon]

Looks good