Retained state required to hydrate the same visible Projection.
One structural array and JavaScript-owned Footage.
Small snapshots use six-word insertion records after length-prefixed ACKs;
words[0] locates those records. Large snapshots begin with
[0, frontiersEnd, insertionCount] and omit predictable coordinate fields.
Both layouts encode the retained coordinates and acknowledgements.
Eligible masks, deleted Footage and compacted Session claims are omitted;
Actor membership remains. Compaction may translate retained coordinates.
Positive insertion diffs consume successive spans of footage.
String-encoded Footage preserves Insertion order, not visible text order.
It uses one UTF-16 code unit per Frame. Missing slots use
NUL (\u0000); hydration splits the string with split('').
[[], []] is an empty Sequence. Both numeric arrays and typed views are accepted.
Retained state required to hydrate the same visible Projection. One structural array and JavaScript-owned Footage. Small snapshots use six-word insertion records after length-prefixed ACKs;
words[0]locates those records. Large snapshots begin with[0, frontiersEnd, insertionCount]and omit predictable coordinate fields. Both layouts encode the retained coordinates and acknowledgements. Eligible masks, deleted Footage and compacted Session claims are omitted; Actor membership remains. Compaction may translate retained coordinates. Positive insertion diffs consume successive spans offootage. String-encoded Footage preserves Insertion order, not visible text order. It uses one UTF-16 code unit per Frame. Missing slots use NUL (\u0000); hydration splits the string withsplit('').[[], []]is an empty Sequence. Both numeric arrays and typed views are accepted.