Render explicit scenario lifetimes and preserve marker selection

This commit is contained in:
2026-09-11 05:25:04 -07:00
parent b853a1eca3
commit 52f6e176cd
15 changed files with 2231 additions and 97 deletions
+27
View File
@@ -319,3 +319,30 @@ Revisit if complete text/native tasks expose unacceptable hidden propagation,
wrong-context edits or inability to preserve reproducible independent studies.
The comparison establishes semantic/text maintenance evidence, not human task
timings or full quantity/library integration acceptance.
## Scenario lifetime rendering integration — 2026-09-11
Codex-ytr83r /root. Adopt the [scoped lifecycle proposal](release-scenario-lifecycle-design-20260910.md)
as derived presentation over existing source-7 semantics. Always retain a neutral
identity header/editing guide; depict definite live intervals separately, with
original-occurrence Create capsules and Delete crosses. Alternative, optional,
loop and parallel operands inherit their enclosing entry independently and leave
an unresolved continuation after a local lifetime change. Reused scenarios remain
opaque. This does not execute guards or add model/source/history records.
Full-height ordinary spines fail to communicate creation/deletion; one global
earliest-birth/latest-death range would falsely order exclusive operands. Both
alternatives lose to the scoped projection. The [independent review](release-scenario-lifecycle-independent-review-20260911.md)
requires true visible root bounds, a complete rendering lane map and local
activation clipping at Delete; insertion bounds/eligible lanes must keep their
separate roles. Receiver endpoints and selectable markers must agree, including
self messages. Those findings determine the integration contract in language
section 8.1 and the real-Scene/native acceptance fixtures.
Core/source semantics are unchanged. In section 4.3, replacing the old
guaranteed-only lifetime/duplicate-reply wording with possible shared-path wording
corrects documentation to `validation.rs` and the existing conditional/parallel
regressions; it does not broaden admission or add a new rule. Compilation and
native acceptance of the renderer are still pending at this decision. Revisit the
visual encoding if actual minimum-size engineering workflows make marker identity,
uncertain continuation or the distinction between guide and existence unclear.