get_url - check for incomplete data read (#85164)

(cherry picked from commit aa8d58a174)

changelogs/fragments/84825-get-url-incomplete-read.yml

@@ -0,0 +1,2 @@
+bugfixes:
+  - get_url - add a check to recognize incomplete data transfers.

lib/ansible/modules/get_url.py

@@ -436,6 +436,23 @@ def url_get(module, url, dest, use_proxy, last_mod_time, force, timeout=10, head
         module.fail_json(msg="failed to create temporary content file: %s" % to_native(e), elapsed=elapsed)
     f.close()
     rsp.close()
+
+    # Since shutil.copyfileobj() will read from HTTPResponse in chunks, HTTPResponse.read() will not recognize
+    # if the entire content-length of data was not read. We need to do that validation here, unless a 'chunked'
+    # transfer-encoding was used, in which case we will not know content-length because it will not be returned.
+    # But in that case, HTTPResponse will behave correctly and recognize an IncompleteRead.
+
+    is_gzip = info.get('content-encoding') == 'gzip'
+
+    if not module.check_mode and 'content-length' in info:
+        # If data is decompressed, then content-length won't match the amount of data we've read, so skip.
+        if not is_gzip or (is_gzip and not decompress):
+            st = os.stat(tempname)
+            cl = int(info['content-length'])
+            if st.st_size != cl:
+                diff = cl - st.st_size
+                module.fail_json(msg=f'Incomplete read, ({rsp.length=}, {cl=}, {st.st_size=}) failed to read remaining {diff} bytes')
+
     return tempname, info
 
 

test/integration/targets/get_url/files/testserver.py

@@ -8,7 +8,15 @@ if __name__ == '__main__':
     PORT = int(sys.argv[1])
 
     class Handler(http.server.SimpleHTTPRequestHandler):
-        pass
+        def do_GET(self):
+            if self.path == '/incompleteRead':
+                self.send_response(200)
+                self.send_header("Content-type", "text/html")
+                self.send_header("Content-Length", "100")
+                self.end_headers()
+                self.wfile.write(b"ABCD")
+            else:
+                super().do_GET()
 
     Handler.extensions_map['.json'] = 'application/json'
     httpd = socketserver.TCPServer(("", PORT), Handler)

test/integration/targets/get_url/tasks/main.yml

@@ -761,3 +761,16 @@
 - assert:
     that:
       - get_dir_filename.dest == remote_tmp_dir ~ "/filename.json"
+
+- name: Test for incomplete data read (issue 85164)
+  get_url:
+    url: 'http://localhost:{{ http_port }}/incompleteRead'
+    dest: '{{ remote_tmp_dir }}/85164.txt'
+  ignore_errors: true
+  register: result
+
+- name: Assert we have an incomplete read failure
+  assert:
+    that:
+      - result is failed
+      - '"Incomplete read" in result.msg'