When you post files to an API, you would normally pass the form to a FormData and it is 2 lines of Javascript, whatever the inputs. In the case of multiple files, each file uses the same identifier. This is problem with Phoenix since the request body is parsed into a map. You have to “append” directly each file with a unique name; you then need to do the same for the other inputs: evt.target[1], evt.target[2].... In addition, these other fields have no identifier in the params (or do something like formdata.append("width, JSON.stringify({width: 10})). It can be difficult to convince someone to produce this code to use your Phoenix API.
I was wondering if we could have an allow-upload for the conn? Or parse the request body into a keyword list?
