Commit graph

16 commits

Author SHA1 Message Date
Christian Klotz
e92bf60a58
fix(agent): make SQLite branch caching scalable 2026-08-01 10:08:56 +03:00
Christian Klotz
c2adb4d1da
fix(agent): use SQLite trigger for rollback test 2026-08-01 02:52:21 +03:00
Christian Klotz
57583617ef
fix(agent): make SQLite session operations linear 2026-08-01 02:11:22 +03:00
Christian Klotz
53422a241a
feat(agent): add storage-owned session readers 2026-08-01 01:24:59 +03:00
Christian Klotz
9f4dfdddbf
fix(agent): remove obsolete SQLite session internals 2026-07-31 22:55:49 +03:00
Christian Klotz
2b765ef529
fix(agent): restore forward session cursor reads
Document complete session store and search composition for scanning and canonical SQLite search.
2026-07-31 17:17:41 +03:00
Christian Klotz
2132f8f75f
fix(agent): clarify session persistence ownership 2026-07-31 16:16:42 +03:00
Christian Klotz
a1bb7e4483
fix(agent): expose session stores through factories 2026-07-31 10:28:59 +03:00
Christian Klotz
2700511e45
fix(agent): refine session repository API 2026-07-31 09:48:24 +03:00
Cristina Poncela Cubeiro
aa784e5eed refactor, cleanup and conditionals for fts5 2026-07-31 05:51:24 +02:00
Cristina Poncela Cubeiro
fb0f0f7cd5 rename: vars 2026-07-31 05:19:23 +02:00
Cristina Poncela Cubeiro
5e48d41ecd refactor: repo facade 2026-07-30 14:07:58 +02:00
Cristina Poncela Cubeiro
6a9591ae78 refactor: repo owns session and storage separately 2026-07-29 16:33:20 +02:00
Cristina Poncela Cubeiro
3c2b71b570 refactor: search index swappable 2026-07-27 17:52:13 +02:00
Cristina Poncela Cubeiro
0f5a6e422b feat: search index sqlite 2026-07-27 09:47:22 +02:00
Cristina Poncela Cubeiro
9e7582aa03
feat: sqlite session storage (#6594)
This PR:

- Adds retainedTail to compaction entries in the new agent harness so we don't have to walk up the tree for the 2000 tokens before compaction,
- Changes getPathToRoot to getPathToRootOrCompaction to only load until last compaction, as unnecessary to access all nodes where it is called,
- Adds a SQLite storage backend, in a separate packages/session-backend-sqlite, with a migration system and schemas as per on-site discussions: sessions to match session header messages (except for metadata, which I couldn't understand what it's used for or where it gets written, so I omitted it), session_entries for shared entry types as columns plus payload as a json for what remains, session_sequences to represent the append-only, serialized nature of the jsonl files, branch_entries to attribute nodes to branches (relationship one-to-many), and session_materialized with the session info (see /session in TUI) to act as a "cache" or quick-access for costs, message count, token info, labels, session name, and model-thinking-level config (e.g. for fast resume).
- This is compatible with the new agent harness Session abstraction.
2026-07-21 11:36:31 +02:00