From ebe4bb725d434277b0f0b639cc554de822f145bb Mon Sep 17 00:00:00 2001
From: Victor Moene <victor.moene@northern.tech>
Date: Fri, 13 Dec 2024 11:45:33 +0100
Subject: [PATCH] Improved handling of already downloaded packages

Ticket: ENT-12526
Signed-off-by: Victor Moene <victor.moene@northern.tech>
---
 cf_remote/utils.py |  8 ++++++++
 cf_remote/web.py   | 19 +++++++++++--------
 2 files changed, 19 insertions(+), 8 deletions(-)

diff --git a/cf_remote/utils.py b/cf_remote/utils.py
index d0522a1..235fbe5 100644
--- a/cf_remote/utils.py
+++ b/cf_remote/utils.py
@@ -1,3 +1,4 @@
+import hashlib
 import os
 import sys
 import re
@@ -202,3 +203,10 @@ def whoami():
 def print_progress_dot(*args):
     print(".", end="")
     sys.stdout.flush()  # STDOUT is line-buffered
+
+
+def is_different_checksum(checksum, content) :
+    assert type(content) == bytes
+
+    digest = hashlib.sha256(content).digest().hex()
+    return checksum != digest
diff --git a/cf_remote/web.py b/cf_remote/web.py
index 28d95c5..dc9ef09 100644
--- a/cf_remote/web.py
+++ b/cf_remote/web.py
@@ -5,7 +5,7 @@
 import urllib.request
 import json
 from collections import OrderedDict
-from cf_remote.utils import user_error, write_json, mkdir, parse_json
+from cf_remote.utils import is_different_checksum, user_error, write_json, mkdir, parse_json
 from cf_remote import log
 from cf_remote.paths import cf_remote_dir, cf_remote_packages_dir
 
@@ -39,22 +39,25 @@ def download_package(url, path=None, checksum=None):
 
     # Use "ab" to prevent truncation of the file in case it is already being
     # downloaded by a different thread.
-    with open(path, "ab") as f:
+    with open(path, "ab+") as f:
         # Get an exclusive lock. If the file size is != 0 then it's already
         # downloaded, otherwise we download.
         fcntl.flock(f.fileno(), fcntl.LOCK_EX)
         st = os.stat(path)
         if st.st_size != 0:
-            log.debug("Package '{}' already downloaded".format(path))
+            print("Package '{}' already downloaded".format(path))
+
+            f.seek(0)
+            content = f.read()
+            if checksum and is_different_checksum(checksum, content):
+                user_error("Downloaded file '{}' does not match expected checksum '{}'. Please delete the file.".format(filename, checksum))
+
         else:
             print("Downloading package: '{}'".format(path))
 
             answer = urllib.request.urlopen(url).read()
-
-            if checksum:
-                digest = hashlib.sha256(answer).digest().hex()
-                if checksum != digest:
-                    user_error("Downloaded file '{}' does not match expected checksum '{}'".format(filename, checksum))
+            if checksum and is_different_checksum(checksum, answer):
+                user_error("Downloaded file '{}' does not match expected checksum '{}'".format(filename, checksum))
 
             f.write(answer)
             f.flush()