Project typed failure origins and join correction evidence by identity

This commit is contained in:
2026-11-16 15:00:05 +00:00
parent b98d463958
commit 5167154a1c
16 changed files with 473 additions and 363 deletions

View File

@@ -51,10 +51,15 @@ count, and a `failures` sequence. Each failure identifies the operation
(`proposal`, `validation`, `validation_input`, or `correction_repair`), optional validator and
zero-based batch index, safe reason/message, sorted proposal indices when
known, and execution counts when available. Multiple failed batches in one
work unit form one error record. `scope: run` instead records phase, safe reason
and message, optional origin, and output path where relevant. It can describe
a fatal error or secondary output-write failure without inventing a failed
model work unit. Raw provider errors and transcript text are absent.
work unit form one error record. A record with scope `run` records a fatal
processing failure or a separate output delivery failure. It identifies the
phase, safe reason and message, and available origin. A processing origin can
include the operation, validator, batch, proposal index, and repair attempt.
A delivery failure includes the output path when known. A record with scope
`interrupted` preserves an interrupted peer's cause and location when
available. These records do not add to the recoverable work-unit or run-error
counts. The work-unit matrix still marks the interrupted work. Raw provider
errors and transcript text are absent.
For `validation_decisions_invalid`, proposal indices identify only candidates
whose required decisions were unavailable. Other candidates in the same batch
may still be approved or rejected normally.