do_bloom(): remove unused "count" variable

[bup.git] / cmd / bloom-cmd.py

diff --git a/cmd/bloom-cmd.py b/cmd/bloom-cmd.py
index 47b4294ba7eec5fac62110b7a54a766c32886f9b..d74cce2704567597fc1fce934032dfd025041e55 100755 (executable)
--- a/cmd/bloom-cmd.py
+++ b/cmd/bloom-cmd.py
@@ -6,6 +6,7 @@ from bup.helpers import *
 optspec = """
 bup bloom [options...]
 --
+ruin       ruin the specified bloom file (clearing the bitfield)
 f,force    ignore existing bloom file and regenerate it from scratch
 o,output=  output bloom filename (default: auto)
 d,dir=     input directory to look for idx files (default: auto)
@@ -13,6 +14,17 @@ k,hashes=  number of hash functions to use (4 or 5) (default: auto)
 c,check=   check the given .idx file against the bloom filter
 """
 
+
+def ruin_bloom(bloomfilename):
+    rbloomfilename = git.repo_rel(bloomfilename)
+    if not os.path.exists(bloomfilename):
+        log("%s\n" % bloomfilename)
+        add_error("bloom: %s not found to ruin\n" % rbloomfilename)
+        return
+    b = bloom.ShaBloom(bloomfilename, readwrite=True, expected=1)
+    b.map[16:16+2**b.bits] = '\0' * 2**b.bits
+
+
 def check_bloom(path, bloomfilename, idx):
     rbloomfilename = git.repo_rel(bloomfilename)
     ridx = git.repo_rel(idx)
@@ -69,14 +81,14 @@ def do_bloom(path, outfilename):
 
     if b:
         if len(b) != rest_count:
-            log("bloom: size %d != idx total %d, regenerating\n"
-                    % (len(b), rest_count))
+            debug1("bloom: size %d != idx total %d, regenerating\n"
+                   % (len(b), rest_count))
             b = None
         elif (b.bits < bloom.MAX_BLOOM_BITS and
               b.pfalse_positive(add_count) > bloom.MAX_PFALSE_POSITIVE):
-            log("bloom: regenerating: adding %d entries gives "
-                "%.2f%% false positives.\n"
-                    % (add_count, b.pfalse_positive(add_count)))
+            debug1("bloom: regenerating: adding %d entries gives "
+                   "%.2f%% false positives.\n"
+                   % (add_count, b.pfalse_positive(add_count)))
             b = None
         else:
             b = bloom.ShaBloom(outfilename, readwrite=True, expected=add_count)
@@ -89,7 +101,7 @@ def do_bloom(path, outfilename):
     msg = b is None and 'creating from' or 'adding'
     if not _first: _first = path
     dirprefix = (_first != path) and git.repo_rel(path)+': ' or ''
-    log('bloom: %s%s %d file%s (%d object%s).\n'
+    progress('bloom: %s%s %d file%s (%d object%s).\n'
         % (dirprefix, msg,
            len(add), len(add)!=1 and 's' or '',
            add_count, add_count!=1 and 's' or ''))
@@ -97,18 +109,20 @@ def do_bloom(path, outfilename):
     tfname = None
     if b is None:
         tfname = os.path.join(path, 'bup.tmp.bloom')
-        tf = open(tfname, 'w+')
-        b = bloom.create(tfname, f=tf, expected=add_count, k=opt.k)
-    count = 0
+        b = bloom.create(tfname, expected=add_count, k=opt.k)
+
     icount = 0
     for name in add:
         ix = git.open_idx(name)
         qprogress('bloom: writing %.2f%% (%d/%d objects)\r' 
                   % (icount*100.0/add_count, icount, add_count))
         b.add_idx(ix)
-        count += 1
         icount += len(ix)
 
+    # Currently, there's an open file object for tfname inside b.
+    # Make sure it's closed before rename.
+    b.close()
+
     if tfname:
         os.rename(tfname, outfilename)
 
@@ -132,6 +146,8 @@ for path in paths:
     outfilename = opt.output or os.path.join(path, 'bup.bloom')
     if opt.check:
         check_bloom(path, outfilename, opt.check)
+    elif opt.ruin:
+        ruin_bloom(outfilename)
     else:
         do_bloom(path, outfilename)
 
@@ -139,4 +155,4 @@ if saved_errors:
     log('WARNING: %d errors encountered during bloom.\n' % len(saved_errors))
     sys.exit(1)
 elif opt.check:
-    log('all tests passed.\n')
+    log('All tests passed.\n')