Improved unicode support in mutator, flattener, and more

[bohendo]

solc, foundry, hardhat, everything reports source_map offsets denominated in bytes (solidity#14733 is a false positive). And (almost) all slither detectors/properly properly index the source code byte-wise. Great.

But some tools, notably the mutator and flattener, use per-byte offsets to index strings per-character. Once this PR is merged, they will not.

Summary of changes:

• fixed source_mapping.content to index source code correctly and used this property instead of manual indexing in flat/mutate tools
• fixed src_mapping usage in the documentation tool too
• small bugfix to Makefile
• manual review of all other src_mapping, source_code, and utf8 encodings to ensure we aren't applying byte-offsets to strings anywhere else
  • resolved subtle unicode bugs in the unused_import, upgradability and codex detectors.
  • standardize encoding strings from a mixture of "utf-8" and "utf8" to just the latter
• try/catch mutant generation, so one bugged mutator won't crash the entire campaign (and it'll print more helpful logs now)
• more tweaks to mutation logging (eg uncaught mutants use white text so the above errors stand out better)
• bug fixes to a few mutators that eg crashed on assembly instructions

Note that the last 3 of these were merged into this branch from PR#2648 bc most of the changes in that PR would have needed to be duplicated in this one.