- Hardlinking duplicate files before compacting,
- & then tracking inodes during compacting would be beneficial.
dupemerge does № 1 well, & № 2 could be handled via code here.
Homepage: https://www.schinagl.priv.at/nt/dupemerge/dupemerge.html
Source repo:
https://gitlab.com/schinagl/link-shell-extension