Coder/deepseek-harness
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

docs: replace vague provenance prose with recorded facts

Browse Source
This commit is contained in:
Turtle
2026-08-09 15:35:02 +08:00
parent 8c124f84b6
commit 9704749b01
380 changed files with 946 additions and 874 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
docs/subsystems/compaction.md
View File

@@ -8,15 +8,15 @@ Source: [`packages/compact/compact/src/types.ts`](../../packages/compact/compact
## The `compact/*` session events
Compaction extends [`SessionEventMap`](session.md) with three event types via declaration merging. All three are **log-only** — they record the compaction lock and its provenance, and never join the surface. `SurfaceEventType` is deliberately NOT extended (only message-producing events reach the model), so the summary itself rides on a separate `user/message` with `surfaceOp: { op: 'replace', start, end }` — the only surface mutation performed by summary compaction. See the Agent Note for why reusing `user/message` is honest rather than a workaround.
Compaction extends [`SessionEventMap`](session.md) with three event types via declaration merging. All three are **log-only** — they record the lock, summary, selected range, shadowed event seqs, token count, and model call without joining the surface. `SurfaceEventType` is deliberately NOT extended (only message-producing events reach the model), so the summary itself rides on a separate `user/message` with `surfaceOp: { op: 'replace', start, end }` — the only surface mutation performed by summary compaction. See the Agent Note for why reusing `user/message` is honest rather than a workaround.
| Event | Payload | Role |
|---|---|---|
| `compact/start` | `{ turn }` | acquires the log-recorded lock; a number identifies the open automatic turn, while `null` identifies a standalone manual attempt |
| `compact/summary` | `{ summary, rawOutput?, llmStreamCall?, shadowedRange, shadowedSeqs, shadowedTokenCount, provider, model, maxTokens?, usage? }` | provenance: the safe summary projection, optional complete provider output and usage, an `llmStreamCall: true` marker when producing the result consumed exactly one call through this context's `ctx.llm.stream()` (which requires complete `rawOutput`), the shadowed surface-boundary pair (`start`/`end` seqs — a position span, not a numeric interval), the shadowed seqs in surface order, the estimated token count, and the summarize call's envelope (`provider`, `model`, plus its generation cap when one applied) — logged so the one-shot request is reconstructable from log + code (the reconstructability Agent Note); unmarked `rawOutput` does not identify the call path |
| `compact/summary` | `{ summary, rawOutput?, llmStreamCall?, shadowedRange, shadowedSeqs, shadowedTokenCount, provider, model, maxTokens?, usage? }` | the safe summary projection, optional complete provider output and usage, an `llmStreamCall: true` marker when producing the result consumed exactly one call through this context's `ctx.llm.stream()` (which requires complete `rawOutput`), the shadowed surface-boundary pair (`start`/`end` seqs — a position span, not a numeric interval), the shadowed seqs in surface order, the estimated token count, and the summarize call's envelope (`provider`, `model`, plus its generation cap when one applied) — logged so the one-shot request is reconstructable from log + code (the reconstructability Agent Note); unmarked `rawOutput` does not identify the call path |
| `compact/end` | `{ turn, error? }` | releases the lock with the same numeric-or-null owner (`error` records an unsuccessful attempt) |
The lock brackets the **whole** operation: `compact/start` is appended first, then summarization, the `compact/summary` provenance record, and the `user/message` replacement all land, and only then `compact/end`. Releasing the lock last turns a crash mid-operation into a detectable orphaned lock (a `compact/start` with no matching `compact/end`) rather than a `compact/end` that falsely claims compaction finished.
The lock brackets the **whole** operation: `compact/start` is appended first, then summarization, the `compact/summary` record, and the `user/message` replacement all land, and only then `compact/end`. Releasing the lock last turns a crash mid-operation into a detectable orphaned lock (a `compact/start` with no matching `compact/end`) rather than a `compact/end` that falsely claims compaction finished.
The markers are lock time points, not an exclusive container. An unrelated idle injection can appear between a standalone manual start and end while summarization is pending. The manual path revalidates only its selected positional span, so that injected context survives after the replacement checkpoint. A live unmatched start blocks every entry point; an unmatched start before a newer `session/end-seed` is stale evidence from a prior lifecycle and is ignored.
@@ -88,7 +88,7 @@ The seam exports `toolPairingBalancedBefore(session, seq)` and `toolPairingBalan
The optional tool-result pruning service reports each durable content replacement and the aggregate Unicode-code-point reduction. Its public result types live in [`compact-tool-result-prune/src/types.ts`](../../packages/compact/compact-tool-result-prune/src/types.ts).
```ts type-equiv
/** Provenance and size accounting for one landed surface replacement. */
/** Cited source event and size accounting for one landed surface replacement. */
interface PrunedEntry {
/** Full-fidelity tool-result event shadowed by the replacement. */
readonly originalSeq: number
@@ -212,7 +212,7 @@ pruneContent(blocks: readonly ContentBlock[]): ContentBlock[] | null
/**
* Prune every over-budget tool result from one stable current-surface snapshot.
* Each replacement preserves the complete event data except for `content`,
* points at the shadowed node for durable provenance and replay, and is
* cites the shadowed node so replay can recover the replacement input, and is
* immediately preceded by a `compact/prune` shadow-price event pricing the
* shadowed node through the injected token meter, so pure consumers can
* subtract it without per-node state.