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

feat(tui): add safe session resume flow

Browse Source
This commit is contained in:
NI0317
2026-07-24 12:31:26 +08:00
committed by ZiyaZhang
parent 65d29da8a1
commit 2ae9f4fdf3
57 changed files with 2312 additions and 192 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
packages/session-persistence/session-persistence-jsonl/README.md
View File

@@ -6,6 +6,9 @@ The JSONL durable session-persistence backend — a concrete `SessionPersistence
```
<root>/
.live/
<encoded-id>.lock # PID + nonce cross-process live lease
<encoded-id>.lock.reclaim # ephemeral stale-owner takeover guard
cwd-<sha256(cwd)[:12]>/ # per-project bucket (or _no-cwd/ when no cwd)
<encoded-id>.jsonl.zstd # default: checksummed header frame + append frames
<encoded-id>.jsonl # only with compression: 'none'
@@ -43,7 +46,7 @@ A root belongs to one encoding. Startup discovery and targeted lookup reject the
## Write path
The plugin copies frozen session events into one controller per live session and starts an eager drain. Concurrent events share the current write; events admitted during it form a follow-up batch, while `session/flush` waits until both current and pending batches are durable. A per-session cursor prevents resumed sessions from re-appending stored events, and live sessions are seeded when the plugin loads. The owning backend instance serializes operations for one session; disposal drains every retained controller before teardown.
The plugin copies frozen session events into one controller per live session and starts an eager drain. Before a session can flush or resume, the coordinator claims an exclusive `.live/<encoded-id>.lock` containing the process PID and an exec-stable nonce; another live process is rejected, while a dead owner is reclaimed under the separate `.reclaim` guard. Concurrent events share the current write; events admitted during it form a follow-up batch, while `session/flush` waits until both current and pending batches are durable. A per-session cursor prevents resumed sessions from re-appending stored events, and live sessions are seeded when the plugin loads. Disposal drains every retained controller before releasing its lease.
## Model Experience
@@ -66,5 +69,6 @@ JSONL storage does not mutate live request prefixes. A resumed loop can reuse pr
- **Only the configured encoding and current `SESSION_FORMAT_VERSION` (v0) load** — changing compression requires a separate/fresh root or selecting the legacy raw mode; the pre-release format has no migration.
- **Compressed files are not directly line-readable** — use the backend to load them, or select `compression: 'none'` before writing a fresh root when text fixtures or external line readers are required.
- **Nothing deletes session files** — logs accumulate under `root` until removed externally (the seam has no deletion surface).
- **One live writer per session** — append and repair are coordinated only inside the owning backend instance. Another backend instance or process must not write the same session until that owner reaches quiescent disposal; initial same-id publication remains collision-safe through the POSIX no-overwrite hard link or Windows write-through rename without replacement.
- **Lease scope is local-host advisory ownership** — PID liveness prevents two ordinary local Harness processes from resuming the same id, but it is not a distributed lease for shared network filesystems or hostile principals.
- **A crash during stale-lease takeover fails closed** — if the reclaiming process itself crashes while holding the short-lived `.reclaim` guard, an operator must remove that guard after confirming no recovery is active.
- **POSIX materialization requires hard-link support** — first append uses `link()` so same-id races fail instead of overwriting a committed log; Windows uses write-through rename without replacement.