cctally 1.79.1 → 1.80.0

package/CHANGELOG.md

@@ -5,6 +5,31 @@ based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 
 ## [Unreleased]
 
+## [1.80.0] - 2026-07-23
+
+### Added
+- Your usage history is now backed by an append-only journal (`~/.local/share/cctally/journal/`) that is the durable source of truth; `stats.db` becomes a disposable, rebuildable index, so a corrupted or deleted stats database now self-heals on the very next command — forensics bundle, quarantine of the damaged file, then an automatic rebuild from the journal — with zero loss of usage history and no manual repair step. (#336)
+- `cctally db rebuild --db stats` rebuilds the stats index from the journal on demand (the manual form of the automatic self-heal), reporting per-table row counts and timing. (#336)
+- `cctally doctor` gains a Journal section reporting the journal's presence and writability, torn/malformed-line counts, how far the stats index has fallen behind the journal, and the most recent self-heal incident. (#336)
+
+### Changed
+- Upgrading to this version migrates existing installs in place automatically on first run, with zero loss of irreplaceable usage history and no manual steps (a slow first run is expected while the one-time journal is written). (#336)
+- Every stats write now flows through a single-flight ingester, ending the `database is locked` pile-ups and wedged handles that concurrent multi-agent hook storms used to cause. (#336)
+- Opening any database is faster: schema work now runs only when the schema version actually changes, not on every open (the status line opens databases several times per render). (#336)
+- Codex quota observations are now preserved durably in the journal, so that history is no longer silently lost when the source rollout files are pruned. (#336)
+- Back up `~/.local/share/cctally/journal/` — hand-copying a `.db` file is now pointless rather than risky, because every database rebuilds from the journal plus surviving provider logs. (#336)
+
+### Fixed
+- The mobile dashboard hero now keeps a clear gap between its 7-day and 5-hour usage values instead of letting a wide 7-day percentage overflow into the adjacent metric.
+
+### Removed
+- `cctally db recover --db stats` is retired: a stats-index version mismatch now self-heals by rebuild, so use `cctally db rebuild --db stats` instead (`db recover --db cache` is unchanged). (#336)
+
+## [1.79.2] - 2026-07-23
+
+### Fixed
+- The dashboard's current-period five-hour milestone section now shows the complete block position and both navigation controls immediately, matching historic periods while keeping older block details lazy-loaded. Claude receives the fix now, and Codex uses the same behavior whenever native 300-minute blocks are retained.
+
 ## [1.79.1] - 2026-07-22
 
 ### Fixed

package/bin/_cctally_cache.py

@@ -331,9 +331,14 @@ def _iter_claude_jsonl_files():
                 yield jp
 
 _cctally_db_sib = _load_lib("_cctally_db")
+# Unified store opener (spec §6.1/§6.2). Owns the per-store PRAGMA policy and
+# the version gate; open_cache_db / open_conversations_db route their
+# connect+policy through it and skip the schema apply when schema_current().
+_cctally_store = _load_lib("_cctally_store")
 add_column_if_missing = _cctally_db_sib.add_column_if_missing
 _run_pending_migrations = _cctally_db_sib._run_pending_migrations
 _CACHE_MIGRATIONS = _cctally_db_sib._CACHE_MIGRATIONS
+_CONVERSATIONS_MIGRATIONS = _cctally_db_sib._CONVERSATIONS_MIGRATIONS
 # Storm-free conversation_messages + FTS full-clear (#138). Owns the trigger
 # drop/recreate dance so the per-row delete trigger never fires O(rows) under
 # the held lock on a rebuild / truncation escalation.
@@ -1191,6 +1196,61 @@ def _prune_inactive_codex_source_roots(
     )
 
 
+def _append_codex_quota_obs(quota_rows: list) -> None:
+    """Journal one Codex quota `obs` per newly-read quota observation (Task 7
+    Item 1, spec §4.2 / §5.3 / Appendix A).
+
+    The direct cache.db write in ``_write_codex_file_batch`` stays byte-identical;
+    this obs is the DURABLE truth for the observation. Codex rollout JSONL
+    evaporates over time (§1 latent data-loss hole), so ``quota_window_snapshots``
+    are silently irreplaceable today — journaling the raw capture closes that
+    hole, and the ingest cycle's ``QUOTA_APPLIER`` (spec §5.2 step 3)
+    re-materializes cache.db from these obs.
+
+    Called UNDER the ``cache.db.codex.lock`` provider flock (the journal append
+    lock is a LEAF — legal to take inside a provider flock, spec §4.3) and BEFORE
+    the cache write / offset advance, so a crash re-reads the same bytes and
+    re-appends (idempotent at the QUOTA_APPLIER's natural-key INSERT OR IGNORE)
+    rather than losing the observation. ``at`` is the detection clock
+    (``_command_as_of()``-aware, honoring the CCTALLY_AS_OF test hook); the raw
+    row (incl. its own ``captured_at_utc``) rides the payload (spec §4.2 raw
+    capture — no derivation baked into the stored line). $CODEX_HOME multi-root
+    scoping is preserved: each row carries its own ``source_root_key`` and roots
+    are never combined."""
+    if not quota_rows:
+        return
+    import _cctally_journal as _jr
+    import _lib_journal as _jl
+    at = (
+        _cctally_core._command_as_of()
+        .isoformat(timespec="seconds")
+        .replace("+00:00", "Z")
+    )
+    for row in quota_rows:
+        (source, source_root_key, source_path, line_offset, captured_at_utc,
+         observed_slot, logical_limit_key, limit_id, limit_name, window_minutes,
+         used_percent, resets_at_utc, plan_type, individual_limit_json,
+         reached_type, observed_model) = row
+        try:
+            _jr.append_record(_jl.make_obs(
+                at=at, src="codex-quota", provider="codex",
+                payload={
+                    "kind": "quota_window_snapshot",
+                    "source": source, "source_root_key": source_root_key,
+                    "source_path": source_path, "line_offset": line_offset,
+                    "captured_at_utc": captured_at_utc,
+                    "observed_slot": observed_slot,
+                    "logical_limit_key": logical_limit_key, "limit_id": limit_id,
+                    "limit_name": limit_name, "window_minutes": window_minutes,
+                    "used_percent": used_percent, "resets_at_utc": resets_at_utc,
+                    "plan_type": plan_type,
+                    "individual_limit_json": individual_limit_json,
+                    "reached_type": reached_type, "observed_model": observed_model,
+                }))
+        except Exception as exc:  # best-effort; a journal append must not break sync
+            eprint(f"[codex-cache] quota obs journal append failed: {exc}")
+
+
 def _write_codex_file_batch(
     conn: sqlite3.Connection,
     *,
@@ -4333,6 +4393,15 @@ def sync_codex_cache(
             # owns the authoritative transcript-local value.
             new_last_turn_id = prev_turn_id
 
+            # Task 7 Item 1: journal the Codex quota observations BEFORE the cache
+            # write (and before the offset advances), under the codex flock this
+            # function already holds. Durable-first: a crash after the append but
+            # before the commit re-reads the same bytes next sync and re-appends
+            # (idempotent at the QUOTA_APPLIER natural key) rather than losing the
+            # observation. Appended once here, not inside the retry loop, so a DB
+            # retry never double-journals.
+            _append_codex_quota_obs(quota_rows)
+
             # Every derived row above was buffered before the first DML. A
             # late database failure therefore rolls the whole file back and
             # retries that same in-memory batch exactly once.
@@ -4738,15 +4807,22 @@ def open_cache_db() -> sqlite3.Connection:
     except OSError as exc:
         eprint(f"[cache] could not chmod data dir 0700 ({exc}); continuing")
     try:
-        conn = sqlite3.connect(_cctally_core.CACHE_DB_PATH)
+        # Connect + install row factory / test trace hook via the unified
+        # opener (spec §6.1). PRAGMAs are applied AFTER the corruption probe so
+        # a mode-change never writes a WAL onto a DB we are about to unlink.
+        conn = _cctally_store.open_index("cache")
         conn.execute("SELECT 1").fetchone()
     except sqlite3.DatabaseError as exc:
         eprint(f"[cache] corrupt cache DB ({exc}); recreating")
+        try:
+            conn.close()
+        except Exception:
+            pass
         try:
             _cctally_core.CACHE_DB_PATH.unlink()
         except FileNotFoundError:
             pass
-        conn = sqlite3.connect(_cctally_core.CACHE_DB_PATH)
+        conn = _cctally_store.open_index("cache")
 
     # Best-effort 0600 on cache.db itself (the 0700 dir above backstops the
     # sidecars until the first write hardens them in sync_cache).
@@ -4755,56 +4831,46 @@ def open_cache_db() -> sqlite3.Connection:
     except OSError as exc:
         eprint(f"[cache] could not chmod cache.db 0600 ({exc}); continuing")
 
-    # #313 P3 reclaim: create a fresh cache.db in INCREMENTAL auto-vacuum mode so
-    # the transcript-retention prune can return freed pages to the OS via
-    # `PRAGMA incremental_vacuum` instead of only growing the freelist — the
-    # conversation-transcript store is what bloated cache.db to multiple GB and
-    # slowed every sync tick. This MUST run before the first page is written
-    # (before journal_mode=WAL / any CREATE TABLE) to take effect; it is a
-    # harmless no-op on an already-created DB, which keeps its existing mode
-    # until a full VACUUM (`cctally db vacuum`) or a `cache-sync --rebuild`.
-    conn.execute("PRAGMA auto_vacuum=INCREMENTAL")
-    conn.execute("PRAGMA journal_mode=WAL")
-    # #297: 15 s (was 5 s) lets a writer wait out a slow-but-normal sync (>5 s)
-    # instead of instantly erroring with "database is locked". Covers plain
-    # writer-vs-writer SQLITE_BUSY, not SQLITE_BUSY_SNAPSHOT (defended by
-    # write-lock-up-front at the write sites, #87).
-    conn.execute("PRAGMA busy_timeout=15000")
-    # #297: bound the persistent WAL file so a resetting checkpoint truncates it
-    # back to this cap instead of leaving it at its multi-GB high-water size.
-    conn.execute(f"PRAGMA journal_size_limit={CACHE_WAL_SIZE_LIMIT_BYTES}")
-    # Re-derivable DB under WAL: NORMAL risks at most the tail transaction on
-    # power loss, and cache.db can always be rebuilt (cache-sync --rebuild).
-    # Fewer ingest fsyncs than the default FULL. Matches stats.db
-    # (bin/_cctally_core.py open_db). #279 S1 F8.
-    conn.execute("PRAGMA synchronous=NORMAL")
-
-    # Apply the shared cache.db schema (cctally-dev#93, D4): Claude tables +
-    # indexes, the session_id / project_path column adds on session_files
-    # (A2 `session` metadata, populated lazily in sync_cache() /
-    # _ensure_session_files_row()), the Codex base tables + indexes, and the
-    # cache_meta sentinel table. This is the single cache.db schema source —
-    # the eager-apply path (_eagerly_apply_cache_migrations) uses the SAME
-    # helper, so the two can no longer drift. The Codex last_total_tokens
-    # ALTER + purge stays below (out of the shared helper — D4/P1#3).
-    _cctally_db_sib._apply_cache_schema(conn)
-
-    # Migration: add last_total_tokens to codex_session_files. When the column
-    # is newly added (i.e. this is the first run after upgrade), purge the
-    # Codex cache so the duplicate-counted rows produced by the previous
-    # iterator are reingested cleanly by sync_codex_cache(). The cache is
-    # fully re-derivable from ~/.codex/sessions/*.jsonl so this is safe.
-    if add_column_if_missing(conn, "codex_session_files", "last_total_tokens", "INTEGER"):
-        conn.execute("DELETE FROM codex_session_entries")
-        conn.execute("DELETE FROM codex_session_files")
-        conn.commit()
-        eprint("[cache] migrated codex cache — re-ingesting")
+    # §6.1 PRAGMA policy (auto_vacuum=INCREMENTAL emitted first — #313 P3
+    # reclaim needs it before the first page; journal_mode=WAL / busy_timeout /
+    # journal_size_limit #297; synchronous=NORMAL #279 S1 F8). Byte-equivalent
+    # to the former inline block, now owned by the shared policy table.
+    _cctally_store.apply_policy(conn, "cache")
+
+    # §6.2 version gate: run the full schema executescript + add_column_if_missing
+    # probes + the codex last_total_tokens ALTER/purge ONLY when the stamped
+    # user_version differs from the registry head. On a steady-state open of an
+    # up-to-date cache.db this whole block is skipped — the open becomes
+    # connect → PRAGMAs → one user_version read → the dispatcher fast-path.
+    if not _cctally_store.schema_current(conn, "cache"):
+        # Apply the shared cache.db schema (cctally-dev#93, D4): Claude tables +
+        # indexes, the session_id / project_path column adds on session_files
+        # (A2 `session` metadata, populated lazily in sync_cache() /
+        # _ensure_session_files_row()), the Codex base tables + indexes, and the
+        # cache_meta sentinel table. This is the single cache.db schema source —
+        # the eager-apply path (_eagerly_apply_cache_migrations) uses the SAME
+        # helper, so the two can no longer drift. The Codex last_total_tokens
+        # ALTER + purge stays below (out of the shared helper — D4/P1#3).
+        _cctally_db_sib._apply_cache_schema(conn)
+
+        # Migration: add last_total_tokens to codex_session_files. When the
+        # column is newly added (i.e. this is the first run after upgrade),
+        # purge the Codex cache so the duplicate-counted rows produced by the
+        # previous iterator are reingested cleanly by sync_codex_cache(). The
+        # cache is fully re-derivable from ~/.codex/sessions/*.jsonl so this is
+        # safe. Under the version gate now (spec §6.2): on an up-to-date DB the
+        # column already exists so add_column_if_missing was a no-op / no purge,
+        # making the gate byte-equivalent.
+        if add_column_if_missing(conn, "codex_session_files", "last_total_tokens", "INTEGER"):
+            conn.execute("DELETE FROM codex_session_entries")
+            conn.execute("DELETE FROM codex_session_files")
+            conn.commit()
+            eprint("[cache] migrated codex cache — re-ingesting")
 
-    # Migration framework dispatcher for cache.db. The registry is empty in
-    # v1 — this is preparatory wiring that activates when the next cache.db
-    # migration ships. With an empty registry the dispatcher hits the
-    # fast-path or fresh-install branch and returns immediately. See spec
-    # §2.5, §3.3 + the @cache_migration decorator further down in this file.
+    # Migration framework dispatcher for cache.db. Runs unconditionally (spec
+    # §6.2: only the schema apply moves under the gate; the dispatcher keeps its
+    # own fast-path that returns early when user_version == registry head). See
+    # the @cache_migration decorator further down in this file.
     _run_pending_migrations(
         conn, registry=_CACHE_MIGRATIONS, db_label="cache.db",
         recover_version_ahead=True,
@@ -4854,10 +4920,13 @@ def open_conversations_db(*, attach_cache: bool = True) -> sqlite3.Connection:
     path = _cctally_core.CONVERSATIONS_DB_PATH
     conn: sqlite3.Connection | None = None
     try:
-        # URI mode belongs to the connection, not only the later ATTACH value.
-        # Without it, some supported system-Python SQLite builds interpret the
-        # read-only ``file:...cache.db?mode=ro`` attachment as a literal path.
-        conn = sqlite3.connect(path, uri=True)
+        # Connect via the unified opener (spec §6.1). URI mode belongs to the
+        # connection, not only the later ATTACH value — without it, some
+        # supported system-Python SQLite builds interpret the read-only
+        # ``file:...cache.db?mode=ro`` attachment as a literal path; the
+        # conversations policy carries ``uri=True``. PRAGMAs are applied after
+        # the corruption probe.
+        conn = _cctally_store.open_index("conversations")
         conn.execute("SELECT 1").fetchone()
     except sqlite3.DatabaseError as exc:
         if conn is not None:
@@ -4878,13 +4947,31 @@ def open_conversations_db(*, attach_cache: bool = True) -> sqlite3.Connection:
             f"[conversations] could not chmod conversations.db 0600 ({exc}); continuing"
         )
 
-    conn.execute("PRAGMA auto_vacuum=INCREMENTAL")
-    conn.execute("PRAGMA journal_mode=WAL")
-    conn.execute("PRAGMA busy_timeout=15000")
-    conn.execute(f"PRAGMA journal_size_limit={CACHE_WAL_SIZE_LIMIT_BYTES}")
-    conn.execute("PRAGMA synchronous=NORMAL")
-    _cctally_db_sib._apply_conversations_schema(conn)
+    # §6.1 PRAGMA policy (auto_vacuum=INCREMENTAL first / WAL / busy_timeout /
+    # journal_size_limit / synchronous=NORMAL) via the shared table.
+    _cctally_store.apply_policy(conn, "conversations")
+    # §6.2 version gate. conversations.db is under the migration framework
+    # (spec §7.2): the schema apply runs only when the stamped user_version is
+    # behind the registry head; the dispatcher below then stamps user_version to
+    # the head so the next open gates the schema apply out. On an existing
+    # populated DB _apply_conversations_schema short-circuits on its own marker,
+    # so no transcript row is touched even when the gate re-runs it.
+    if not _cctally_store.schema_current(conn, "conversations"):
+        _cctally_db_sib._apply_conversations_schema(conn)
+    # Commit the schema apply BEFORE the dispatcher: _apply_conversations_schema
+    # can leave an open implicit transaction (its trailing marker INSERT), and
+    # the dispatcher's bootstrap-rename opens its own BEGIN — which would collide
+    # with a still-open transaction. Committing here leaves the connection in
+    # autocommit so the dispatcher starts clean.
     conn.commit()
+    # Migration framework dispatcher for conversations.db (spec §7.2). Runs
+    # unconditionally with its own fast-path (user_version == registry head).
+    # recover_version_ahead=True because conversations.db is re-derivable from
+    # provider JSONL, matching the cache.db posture.
+    _run_pending_migrations(
+        conn, registry=_CONVERSATIONS_MIGRATIONS, db_label="conversations.db",
+        recover_version_ahead=True,
+    )
 
     if attach_cache:
         # Ensure the compact schema exists before opening it read-only.  This

package/bin/_cctally_config.py

@@ -1050,8 +1050,11 @@ def _cmd_config_set(args: argparse.Namespace) -> int:
         except _BudgetConfigError as exc:
             eprint(f"cctally config: {exc}")
             return 2
-        # Deliberately outside the config lock: this helper opens and locks DBs.
-        c._reconcile_codex_budget_on_config_write({"codex": configured})
+        # Deliberately outside the config lock. 6f writer reroute: routed THROUGH
+        # the ingest cycle (opportunistic + exception-wrapped) so the latched
+        # Codex crossings are journaled and rebuild-replayable.
+        import _cctally_journal as _jr
+        _jr.reconcile_budget_config({"codex": configured}, axes={"codex_budget"})
         leaf = key.removeprefix(_CODEX_BUDGET_LEAF_PREFIX)
         value = configured[leaf]
         if getattr(args, "emit_json", False):
@@ -1765,21 +1768,32 @@ def _cmd_config_set(args: argparse.Namespace) -> int:
         # forward-only-from-set reconcile prevents (`budget set --period` already
         # reconciles via the same helper). Both run OUTSIDE config_writer_lock
         # (each helper has its own open_db lock).
+        # 6f writer reroute: the conditional per-leaf axis mapping is UNCHANGED,
+        # but the reconciles now run THROUGH the ingest cycle (opportunistic +
+        # exception-wrapped) so their latched crossings are journaled and
+        # rebuild-replayable. A leaf that feeds two axes (`alert_thresholds`
+        # feeds both budget + project_budget) reconciles both in ONE cycle via
+        # the axes union. `touched_projects=None` = reconcile every configured
+        # project (a `config set` write is wholesale, not project-scoped).
+        _reconcile_axes: set = set()
         if inner_key in (
             "weekly_usd", "alerts_enabled", "alert_thresholds", "period"
         ):
-            c._reconcile_budget_on_config_write(validated)
+            _reconcile_axes.add("budget")
         if inner_key in (
             "projects", "project_alerts_enabled", "alert_thresholds"
         ):
-            c._reconcile_project_budget_milestones_on_write(validated)
+            _reconcile_axes.add("project_budget")
         # Codex budget axis (spec §6): the nested budget.codex block is set
         # wholesale via `config set budget.codex '<json>'`, so the only key that
         # touches it is `codex` itself. Gated on the codex block carrying
         # alerts_enabled + thresholds (the helper re-checks); records nothing
         # otherwise.
         if inner_key == "codex":
-            c._reconcile_codex_budget_on_config_write(validated)
+            _reconcile_axes.add("codex_budget")
+        if _reconcile_axes:
+            import _cctally_journal as _jr
+            _jr.reconcile_budget_config(validated, axes=_reconcile_axes)
         out_val = validated[inner_key]
         if getattr(args, "emit_json", False):
             print(json.dumps({"budget": {inner_key: out_val}}, indent=2))
@@ -1843,8 +1857,11 @@ def _cmd_config_unset(args: argparse.Namespace) -> int:
             return 2
         # An optional leaf leaves a configured block, even when it already
         # held its default; the forward-only state must be reconciled once.
+        # 6f writer reroute: routed THROUGH the ingest cycle (opportunistic +
+        # exception-wrapped) so the latched Codex crossings are journaled.
         assert configured is not None
-        c._reconcile_codex_budget_on_config_write({"codex": configured})
+        import _cctally_journal as _jr
+        _jr.reconcile_budget_config({"codex": configured}, axes={"codex_budget"})
         return 0
     if key == "display.tz":
         with config_writer_lock():