From 02ad4bfab28c0cc82b7e4205a819e9fbd7a6786a Mon Sep 17 00:00:00 2001
From: ed <s@ocv.me>
Date: Sat, 10 Dec 2022 22:13:21 +0000
Subject: [PATCH] ensure consistency between db tables

---
 copyparty/__main__.py |  1 +
 copyparty/up2k.py     | 21 +++++++++++++++++++++
 2 files changed, 22 insertions(+)

diff --git a/copyparty/__main__.py b/copyparty/__main__.py
index 24f1ed8b..7bfaa7a6 100755
--- a/copyparty/__main__.py
+++ b/copyparty/__main__.py
@@ -834,6 +834,7 @@ def run_argparse(
     ap2.add_argument("--no-hash", metavar="PTN", type=u, help="regex: disable hashing of matching paths during e2ds folder scans (volflag=nohash)")
     ap2.add_argument("--no-idx", metavar="PTN", type=u, help="regex: disable indexing of matching paths during e2ds folder scans (volflag=noidx)")
     ap2.add_argument("--no-dhash", action="store_true", help="disable rescan acceleration; do full database integrity check -- makes the db ~5%% smaller and bootup/rescans 3~10x slower")
+    ap2.add_argument("--re-dhash", action="store_true", help="rebuild the cache if it gets out of sync (for example crash on startup during metadata scanning)")
     ap2.add_argument("--no-forget", action="store_true", help="never forget indexed files, even when deleted from disk -- makes it impossible to ever upload the same file twice (volflag=noforget)")
     ap2.add_argument("--dbd", metavar="PROFILE", default="wal", help="database durability profile; sets the tradeoff between robustness and speed, see --help-dbd (volflag=dbd)")
     ap2.add_argument("--xlink", action="store_true", help="on upload: check all volumes for dupes, not just the target volume (volflag=xlink)")
diff --git a/copyparty/up2k.py b/copyparty/up2k.py
index 55201d64..6869a82b 100644
--- a/copyparty/up2k.py
+++ b/copyparty/up2k.py
@@ -182,9 +182,14 @@ class Up2k(object):
         have_e2d = self.init_indexes(all_vols, [])
 
         if self.stop:
+            # up-mt consistency not guaranteed if init is interrupted;
+            # drop caches for a full scan on next boot
+            self._drop_caches()
+
             if self.pp:
                 self.pp.end = True
                 self.pp = None
+
             return
 
         if not self.pp and self.args.exit == "idx":
@@ -470,6 +475,10 @@ class Up2k(object):
         if next((zv for zv in vols if "e2ds" in zv.flags), None):
             self._block("indexing")
 
+        if self.args.re_dhash:
+            self.args.re_dhash = False
+            self._drop_caches()
+
         for vol in vols:
             if self.stop:
                 break
@@ -1248,6 +1257,18 @@ class Up2k(object):
 
         return ret
 
+    def _drop_caches(self) -> None:
+        self.log("dropping caches for a full filesystem scan")
+        for vol in self.asrv.vfs.all_vols.values():
+            reg = self.register_vpath(vol.realpath, vol.flags)
+            if not reg:
+                continue
+
+            cur, _ = reg
+            self._set_tagscan(cur, True)
+            cur.execute("delete from dh")
+            cur.connection.commit()
+
     def _set_tagscan(self, cur: "sqlite3.Cursor", need: bool) -> bool:
         if self.args.no_dhash:
             return False