trigger the db progress handler after more operations

We originally were triggering on 100 opcodes, because at the time we were
doing write-heavy alterations to the DB for inactive tags, and a higher level
of opcodes would pause the interface for a long time. The query structure is
different now, so we can afford to save the overhead of more frequent calls.

With the change, a .reset() triggers the handler 3 times; fixIntegrity()
triggers it 30 times over a period of 4.5 seconds.

anki/db.py

@@ -71,3 +71,6 @@ class DB(object):
 
     def close(self):
         self._db.close()
+
+    def set_progress_handler(self, *args):
+        self._db.set_progress_handler(*args)

anki/deck.py

@@ -535,12 +535,6 @@ where c.fid == f.id and f.mid == m.id and c.gid = g.id
     def tagList(self):
         return self.db.list("select name from tags order by name")
 
-    def cardsWithNoTags(self):
-        return self.db.list("""
-select cards.id from cards, facts where
-facts.tags = ""
-and cards.fid = facts.id""")
-
     def cardHasTag(self, card, tag):
         tags = self.db.scalar("select tags from fact where id = :fid",
                               fid=card.fid)
@@ -630,16 +624,14 @@ update facts set tags = :t, mod = :n where id = :id""", [fix(row) for row in res
             return
         self.progressHandlerCalled = time.time()
         if self.progressHandlerEnabled:
+            # things which hook on this should be very careful not to touch
+            # the db as they run
             runHook("dbProgress")
 
     def setupProgressHandler(self):
         self.progressHandlerCalled = 0
         self.progressHandlerEnabled = False
-        try:
+        self.db.set_progress_handler(self.progressHandler, 100000)
-            self.engine.raw_connection().set_progress_handler(
-                deck.progressHandler, 100)
-        except:
-            pass
 
     def enableProgressHandler(self):
         self.progressHandlerEnabled = True