I have written my own software to access "Google Drive" and I am seeing that even providing a MIME framing, Google will happily write a partial file if the HTTP connection is severed.
I am USING the API as a POST to "https://www.googleapis.com/upload/drive/v3/files" indicating 'uploadType': 'multipart', a "content-type": "multipart/related" headers and a correctly formatted MIME multipart file. So, it is trivial for Google to detect that the file is incomplete but, apparently, it is not doing it. It is trivial because a MIME multipart upload has a particular trailer that will be missing if the upload is incomplete.
Is there anything I can do so Google doesn't store partial files?. A MD5 header, maybe? A content-length header?.
If not, is there any way to detect it (searchable metadata, maybe) beside my program "knowing" that a particular file upload was partially done?. Maybe the "CHANGES" API?.
Thanks for your time in advance!.
PS: I could do a "save to another filename" + "rename to the correct filename when you are sure that Google has the complete file" but then I would upload half speed because latency (yes, I could do pipelining) and, WORSE, waste half of my scarce API rate limits. If there is no other choice...