diff options
| -rwxr-xr-x | arabesque.py | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/arabesque.py b/arabesque.py index 5b8e209..1c5663b 100755 --- a/arabesque.py +++ b/arabesque.py @@ -416,7 +416,11 @@ def backward(log_file, map_db, output_db, hit_mimetypes=FULLTEXT_MIMETYPES): continue if line.mimetype == "application/octet-stream" and int(line.size_bytes) < 1000: - counts['skip-tiny-octetstream-'] += 1 + counts['skip-tiny-octetstream'] += 1 + continue + + if int(line.size_bytes) == 0 or line.sha1 == "3I42H3S6NNFQ2MSVX7XZKYAYSCX5QBYJ": + counts['skip-empty-file'] += 1 continue #print(time.time()) |