fsck: row indices are too big for slice.indices, so calculate manually

Normally, indexes for an array are expected to fit in a platform's
native long (32 or 64-bit).  In nilmdb, tables aren't real arrays and
we need to handle unbounded indices.

nilmdb/fsck/fsck.py

@@ -425,11 +425,15 @@ class Fsck(object):
         for intv in ints:
             last_ts = None
             (stime, etime, spos, epos) = intv
-            if spos == epos:
-                continue
-            for start in xrange(*slice(spos, epos, maxrows).indices(epos)):
+
+            # Break interval into maxrows-sized chunks
+            next_start = spos
+            while next_start < epos:
+                start = next_start
                 stop = min(start + maxrows, epos)
                 count = stop - start
+                next_start = stop
+
                 # Get raw data, convert to NumPy arary
                 try:
                     raw = tab.get_data(start, stop, binary = True)