)]}'
{"cinder/volume/drivers/sheepdog.py":[{"author":{"_account_id":4523,"name":"Eric Harney","email":"eharney@redhat.com","username":"eharney"},"change_message_id":"506cf305cc387a029b56ec15af9c484ae29d5bd5","unresolved":false,"context_lines":[{"line_number":188,"context_line":"        with image_utils.temporary_file() as tmp:"},{"line_number":189,"context_line":"            # image_utils.convert_image doesn\u0027t support \"sheepdog:\" source,"},{"line_number":190,"context_line":"            # so we use the qemu-img directly."},{"line_number":191,"context_line":"            cmd \u003d (\u0027qemu-img\u0027,"},{"line_number":192,"context_line":"                   \u0027convert\u0027,"},{"line_number":193,"context_line":"                   \u0027-t\u0027, \u0027none\u0027,"},{"line_number":194,"context_line":"                   \u0027-O\u0027, \u0027raw\u0027,"}],"source_content_type":"text/x-python","patch_set":5,"id":"9aa53dc9_dc483c8a","line":191,"updated":"2015-04-06 17:08:24.000000000","message":"I\u0027m not sure we should add another qemu-img convert call which is doing format guessing of the source volume.\n\nThis has been problematic for security in other areas and it may be better to wait until we have a standard, hardened way of doing this.  (Storing the volume format in Cinder and using that.)","commit_id":"abe6e2009109a778ca1fda89855ca81e5117a0de"},{"author":{"_account_id":4523,"name":"Eric Harney","email":"eharney@redhat.com","username":"eharney"},"change_message_id":"a47d3c7f815837fd62d47928fe2c65c274c978a8","unresolved":false,"context_lines":[{"line_number":188,"context_line":"        with image_utils.temporary_file() as tmp:"},{"line_number":189,"context_line":"            # image_utils.convert_image doesn\u0027t support \"sheepdog:\" source,"},{"line_number":190,"context_line":"            # so we use the qemu-img directly."},{"line_number":191,"context_line":"            cmd \u003d (\u0027qemu-img\u0027,"},{"line_number":192,"context_line":"                   \u0027convert\u0027,"},{"line_number":193,"context_line":"                   \u0027-t\u0027, \u0027none\u0027,"},{"line_number":194,"context_line":"                   \u0027-O\u0027, \u0027raw\u0027,"}],"source_content_type":"text/x-python","patch_set":5,"id":"9aa53dc9_60c17e2e","line":191,"in_reply_to":"9aa53dc9_1209ed95","updated":"2015-04-06 21:51:14.000000000","message":"This should work if it\u0027s known to always be raw-formatted.","commit_id":"abe6e2009109a778ca1fda89855ca81e5117a0de"},{"author":{"_account_id":124,"name":"Kazutaka Morita","email":"morita.kazutaka@gmail.com","username":"morita-kazutaka"},"change_message_id":"fff5517bb985e6745aca7eeef95cfadc47139df0","unresolved":false,"context_lines":[{"line_number":188,"context_line":"        with image_utils.temporary_file() as tmp:"},{"line_number":189,"context_line":"            # image_utils.convert_image doesn\u0027t support \"sheepdog:\" source,"},{"line_number":190,"context_line":"            # so we use the qemu-img directly."},{"line_number":191,"context_line":"            cmd \u003d (\u0027qemu-img\u0027,"},{"line_number":192,"context_line":"                   \u0027convert\u0027,"},{"line_number":193,"context_line":"                   \u0027-t\u0027, \u0027none\u0027,"},{"line_number":194,"context_line":"                   \u0027-O\u0027, \u0027raw\u0027,"}],"source_content_type":"text/x-python","patch_set":5,"id":"9aa53dc9_1209ed95","line":191,"in_reply_to":"9aa53dc9_dc483c8a","updated":"2015-04-06 17:37:48.000000000","message":"In this case, the source is a raw formatted volume stored in the Sheepdog storage system.\n\nHow about specifying \u0027-f raw\u0027 explicitly? Then, we can make sure that qemu-img convert never does format guessing at all.","commit_id":"abe6e2009109a778ca1fda89855ca81e5117a0de"}]}