fix(backup): skip DB backup when the index is unchanged

backup-db.sh

@@ -6,7 +6,7 @@ DB_DIR="$(dirname "$(realpath "$0")")"
 DB_PATH="${CODE_SEARCH_DB:-$DB_DIR/code_index.db}"
 BACKUP_DIR="${CODE_SEARCH_BACKUP_DIR:-$DB_DIR/backups}"
 TIMESTAMP=$(date +%Y-%m-%d_%H%M)
-MAX_BACKUPS="${CODE_SEARCH_MAX_BACKUPS:-14}"
+MAX_BACKUPS="${CODE_SEARCH_MAX_BACKUPS:-4}"
 
 mkdir -p "$BACKUP_DIR"
 
@@ -15,16 +15,49 @@ if [ ! -f "$DB_PATH" ]; then
     exit 1
 fi
 
-# Use Python sqlite3 backup for consistency (sqlite3 CLI not installed)
+# Skip when nothing changed since the last good backup. Each copy is a full
+# 15G file. The stamp records the source state (size, ns mtime, inode) and the
+# backup file it produced; skip only if that file still exists, the source
+# state matches, and there are no pending WAL frames.
+STAMP_FILE="$BACKUP_DIR/.last-source-stamp"
+source_state() { stat -c '%s %.9Y %i' "$DB_PATH"; }
+wal_size() { stat -c %s "$DB_PATH-wal" 2>/dev/null || echo 0; }
+
+BEFORE="$(source_state)"
+WAL_BEFORE="$(wal_size)"
+if [ "$WAL_BEFORE" -eq 0 ] && [ -f "$STAMP_FILE" ]; then
+    IFS=$'\t' read -r LAST_STATE LAST_BACKUP < "$STAMP_FILE" || true
+    if [ "${LAST_STATE:-}" = "$BEFORE" ] && [ -n "${LAST_BACKUP:-}" ] \
+        && [ -f "$BACKUP_DIR/$LAST_BACKUP" ]; then
+        echo "Unchanged since $LAST_BACKUP; skipping backup"
+        exit 0
+    fi
+fi
+
+# Use Python sqlite3 backup for consistency (sqlite3 CLI not installed).
+# Write to a temp name and rename on success so a killed run never leaves a
+# truncated file that looks like a valid backup.
+BACKUP_NAME="code_index_$TIMESTAMP.db"
+TMP_BACKUP="$BACKUP_DIR/$BACKUP_NAME.tmp"
+trap 'rm -f "$TMP_BACKUP"' EXIT
+rm -f "$TMP_BACKUP"
 python3 -c "
 import sqlite3, sys
 src = sqlite3.connect('$DB_PATH')
-dst = sqlite3.connect('$BACKUP_DIR/code_index_$TIMESTAMP.db')
+dst = sqlite3.connect('$TMP_BACKUP')
 src.backup(dst)
 dst.close()
 src.close()
 "
-echo "Backed up to $BACKUP_DIR/code_index_$TIMESTAMP.db ($(du -sh "$BACKUP_DIR/code_index_$TIMESTAMP.db" | cut -f1))"
+mv -f "$TMP_BACKUP" "$BACKUP_DIR/$BACKUP_NAME"
+echo "Backed up to $BACKUP_DIR/$BACKUP_NAME ($(du -sh "$BACKUP_DIR/$BACKUP_NAME" | cut -f1))"
+
+# Record the stamp only if the source did not move during the copy.
+if [ "$WAL_BEFORE" -eq 0 ] && [ "$(wal_size)" -eq 0 ] && [ "$(source_state)" = "$BEFORE" ]; then
+    printf '%s\t%s\n' "$BEFORE" "$BACKUP_NAME" > "$STAMP_FILE"
+else
+    rm -f "$STAMP_FILE"
+fi
 
 # Rotate: keep only the newest MAX_BACKUPS
 cd "$BACKUP_DIR"

tests/test_backup_db.py

@@ -0,0 +1,89 @@
+"""Regression tests for backup-db.sh skip-if-unchanged behavior.
+
+Each backup is a full copy of the index, so an unchanged DB must not produce a
+new copy on every scheduled run. These tests use a throwaway SQLite file and a
+temp backup dir; they never touch the real index.
+"""
+
+from __future__ import annotations
+
+import os
+import sqlite3
+import subprocess
+from pathlib import Path
+
+SCRIPT = Path(__file__).resolve().parent.parent / "backup-db.sh"
+
+
+def run_backup(db: Path, backups: Path) -> str:
+    env = {**os.environ, "CODE_SEARCH_DB": str(db), "CODE_SEARCH_BACKUP_DIR": str(backups)}
+    result = subprocess.run(["bash", str(SCRIPT)], env=env, capture_output=True, text=True, check=True)
+    return result.stdout
+
+
+def make_db(path: Path) -> None:
+    """Create a WAL-mode DB, like the real index."""
+    conn = sqlite3.connect(path)
+    conn.execute("pragma journal_mode=wal")
+    conn.execute("create table t (a)")
+    conn.commit()
+    conn.close()
+
+
+def test_unchanged_db_is_skipped(tmp_path: Path) -> None:
+    db, backups = tmp_path / "index.db", tmp_path / "backups"
+    make_db(db)
+
+    assert "Backed up" in run_backup(db, backups)
+    assert "skipping backup" in run_backup(db, backups)
+    assert len(list(backups.glob("code_index_*.db"))) == 1
+
+
+def test_changed_db_is_backed_up_again(tmp_path: Path) -> None:
+    db, backups = tmp_path / "index.db", tmp_path / "backups"
+    make_db(db)
+    run_backup(db, backups)
+
+    # No mtime bump: a same-second write must still be detected.
+    conn = sqlite3.connect(db)
+    conn.execute("insert into t values (1)")
+    conn.commit()
+    conn.close()
+
+    assert "Backed up" in run_backup(db, backups)
+
+
+def test_missing_stamp_forces_backup(tmp_path: Path) -> None:
+    db, backups = tmp_path / "index.db", tmp_path / "backups"
+    make_db(db)
+    run_backup(db, backups)
+    (backups / ".last-source-stamp").unlink()
+
+    assert "Backed up" in run_backup(db, backups)
+
+
+def test_pending_wal_forces_backup(tmp_path: Path) -> None:
+    db, backups = tmp_path / "index.db", tmp_path / "backups"
+    make_db(db)
+    run_backup(db, backups)
+
+    writer = sqlite3.connect(db)
+    writer.execute("pragma wal_autocheckpoint=0")
+    writer.execute("insert into t values (2)")
+    writer.commit()
+    try:
+        assert (tmp_path / "index.db-wal").stat().st_size > 0
+        assert "Backed up" in run_backup(db, backups)
+    finally:
+        writer.close()
+
+
+def test_deleted_newest_backup_forces_backup(tmp_path: Path) -> None:
+    db, backups = tmp_path / "index.db", tmp_path / "backups"
+    make_db(db)
+    run_backup(db, backups)
+    for backup in backups.glob("code_index_*.db"):
+        backup.unlink()
+
+    assert "Backed up" in run_backup(db, backups)
+    assert not list(backups.glob("*.tmp"))