FA-75005 / CRDT convergence / Member archive

Delta causal context join: compaction jumps over gaps · case 05

The vector claims dots that were never observed, so unseen additions are later discarded.

Member previewVariant 5 · 3 implementations · 10 checks per implementation

Case contract

A state is {"dots": [[element, replica, n]], "vv": {replica: n}, "cloud": [[replica, n]]}. A context covers (r, n) when n <= vv[r] or [r, n] is in the cloud. Join keeps dots present in both states plus dots of one state not covered by the other's context; contexts join by vector maximum and cloud union, then the cloud is compacted: an entry equal to vv[r]+1 advances vv[r], entries at or below vv[r] are pruned. Return elements, dots, vector and cloud, all sorted.

Why this case matters

Delta-state add-wins sets rely on causal contexts to tell removals from unseen additions.

One recorded failure

Sample boundary fixture

This sample comes from the broken implementation of a controlled reproducer.

Boundary fixture	Actual	Expected	Outcome
gaps stay in the cloud	`{"cloud": [], "dots": [["q", "a", 9]], "elements": ["q"], "vv": [["a", 9]]}`	`{"cloud": [["a", 3], ["a", 9]], "dots": [["q", "a", 9]], "elements": ["q"], "vv": [["a", 1]]}`	Failed

MEMBER ARCHIVE

The complete case is available to members.

This record includes three runnable implementations, regression fixtures, execution results, and source hashes.

Member access is invitation-based. Sign in with your invited account to inspect the sources.