README.md's "Split rounding and dust" section correctly describes the current algorithm: "compute_split now uses a largest-remainder allocation... Earlier versions assigned all accumulated dust to the final recipient." This matches the actual code (sort_remainders_desc / largest-remainder logic in both contracts/escrow/src/lib.rs and contracts/milestones/src/lib.rs, landed via the already-closed #25/#55).
But further down, under "1. contracts/escrow — mergefi-escrow", the release: bullet (README.md:181) still says: "Deducts fee_bps off the top to the treasury, splits the rest pro-rata, with the last recipient absorbing integer-division remainder so no dust is stranded in the contract." That's the old, already-replaced behavior, directly contradicting the "Split rounding and dust" section a few dozen lines earlier in the same file. Update that bullet to describe largest-remainder rounding, matching the milestones section's equivalent bullet (which already correctly says "largest-remainder rounding").
README.md's "Split rounding and dust" section correctly describes the current algorithm: "
compute_splitnow uses a largest-remainder allocation... Earlier versions assigned all accumulated dust to the final recipient." This matches the actual code (sort_remainders_desc/ largest-remainder logic in bothcontracts/escrow/src/lib.rsandcontracts/milestones/src/lib.rs, landed via the already-closed #25/#55).But further down, under "1.
contracts/escrow—mergefi-escrow", therelease:bullet (README.md:181) still says: "Deductsfee_bpsoff the top to the treasury, splits the rest pro-rata, with the last recipient absorbing integer-division remainder so no dust is stranded in the contract." That's the old, already-replaced behavior, directly contradicting the "Split rounding and dust" section a few dozen lines earlier in the same file. Update that bullet to describe largest-remainder rounding, matching the milestones section's equivalent bullet (which already correctly says "largest-remainder rounding").