)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"1ac67100fea21693412ab25d55abc81ac8d85fc7","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"84f2e47d_08ad93a1","updated":"2025-04-21 20:37:33.000000000","message":"Thanks for getting started on this.\n\nI know this is WIP, so maybe you\u0027re planning to fix the data pipeline as well. However, I wouldn\u0027t even bother checking the filename against a list of disallowed files as I don\u0027t think it really gains us anything. I can also `mknod c 1 5 foo` and then use that instead of `/dev/zero` and this check wouldn\u0027t catch it. The real fix has to be a chunked read pipeline, IMHO.","commit_id":"8aec0ecdd6b5905a11ac9f3649c3dd665d52f7bc"},{"author":{"_account_id":9303,"name":"Abhishek Kekane","email":"akekane@redhat.com","username":"abhishekkekane"},"change_message_id":"b5c95d62b61ee0aaf61520a734393cdc79c76c3f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"85ca7583_2c7ebc1e","in_reply_to":"84f2e47d_08ad93a1","updated":"2025-04-22 14:25:55.000000000","message":"Ack, I think first thing is to make this transfer in chunks, will work on that.","commit_id":"8aec0ecdd6b5905a11ac9f3649c3dd665d52f7bc"}],"openstackclient/image/v2/image.py":[{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"1ac67100fea21693412ab25d55abc81ac8d85fc7","unresolved":true,"context_lines":[{"line_number":55,"context_line":"    \"ploop\","},{"line_number":56,"context_line":"]"},{"line_number":57,"context_line":"MEMBER_STATUS_CHOICES \u003d [\"accepted\", \"pending\", \"rejected\", \"all\"]"},{"line_number":58,"context_line":"SENSITIVE_FILES \u003d [\u0027/dev/zero\u0027, \u0027/dev/null\u0027, \u0027/dev/random\u0027, \u0027/dev/urandom\u0027]"},{"line_number":59,"context_line":""},{"line_number":60,"context_line":"LOG \u003d logging.getLogger(__name__)"},{"line_number":61,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"91567475_ad39276e","line":58,"updated":"2025-04-21 20:37:33.000000000","message":"This is not a very good check, because I can fool it with:\n```\ncat /dev/zero | openstack image create\n```\nAlso, the actual problem is that we\u0027re reading the whole thing in, so if I create a 1TB image and try to upload it, I\u0027m going to OOM all the same.\n\nThis code needs to be able to read and send the file (or stdin stream) in chunks. There\u0027s no way around it.","commit_id":"8aec0ecdd6b5905a11ac9f3649c3dd665d52f7bc"}]}