From 8ab6c8a22d7b8d5dfbf37e84e315c849bc36cd2b Mon Sep 17 00:00:00 2001
From: Jake Moshenko <jake.moshenko@coreos.com>
Date: Mon, 11 Jan 2016 16:43:20 -0500
Subject: [PATCH] Fix torrent hash generation to work in mixed stacks

---
 data/database.py     | 4 ++--
 data/fields.py       | 7 +++++--
 endpoints/v2/blob.py | 8 +++++---
 3 files changed, 12 insertions(+), 7 deletions(-)

diff --git a/data/database.py b/data/database.py
index 7d87610d9..0573eadea 100644
--- a/data/database.py
+++ b/data/database.py
@@ -809,8 +809,8 @@ class BlobUpload(BaseModel):
   chunk_count = IntegerField(default=0)
   uncompressed_byte_count = IntegerField(null=True)
   created = DateTimeField(default=datetime.now, index=True)
-  piece_sha_state = ResumableSHA1Field(null=True, default=resumablehashlib.sha1)
-  piece_hashes = Base64BinaryField(null=True, default='')
+  piece_sha_state = ResumableSHA1Field(null=True)
+  piece_hashes = Base64BinaryField(null=True)
 
   class Meta:
     database = db
diff --git a/data/fields.py b/data/fields.py
index c069c2d6d..f73bc0cf1 100644
--- a/data/fields.py
+++ b/data/fields.py
@@ -10,6 +10,9 @@ class _ResumableSHAField(TextField):
     raise NotImplementedError
 
   def db_value(self, value):
+    if value is None:
+      return None
+
     sha_state = value.state()
 
     # One of the fields is a byte string, let's base64 encode it to make sure
@@ -19,14 +22,14 @@ class _ResumableSHAField(TextField):
     return json.dumps(sha_state)
 
   def python_value(self, value):
-    to_resume = self._create_sha()
     if value is None:
-      return to_resume
+      return None
 
     sha_state = json.loads(value)
 
     # We need to base64 decode the data bytestring.
     sha_state[3] = base64.b64decode(sha_state[3])
+    to_resume = self._create_sha()
     to_resume.set_state(sha_state)
     return to_resume
 
diff --git a/endpoints/v2/blob.py b/endpoints/v2/blob.py
index d8b1c9844..210e424f8 100644
--- a/endpoints/v2/blob.py
+++ b/endpoints/v2/blob.py
@@ -3,6 +3,8 @@ import re
 
 from flask import make_response, url_for, request, redirect, Response, abort as flask_abort
 
+import resumablehashlib
+
 from app import storage, app
 from auth.registry_jwt_auth import process_registry_jwt_auth
 from data import model, database
@@ -224,9 +226,9 @@ def _upload_chunk(namespace, repo_name, upload_uuid):
 
     piece_hasher = None
     # TODO remove this when all in-progress blob uploads reliably contain piece hashes
-    if start_offset == 0 or found.piece_sha_state is not None:
-      piece_hasher = PieceHasher(app.config['TORRENT_PIECE_SIZE'], start_offset, found.piece_hashes,
-                                 found.piece_sha_state)
+    if start_offset == 0:
+      piece_hasher = PieceHasher(app.config['TORRENT_PIECE_SIZE'], start_offset, '',
+                                 resumablehashlib.sha1())
       input_fp = wrap_with_handler(input_fp, piece_hasher.update)
 
     # If this is the first chunk and we're starting at the 0 offset, add a handler to gunzip the