This will not be retroactive but it is fixed for the next release.
Rather than relying on the file name coming from the response headers when we fetch the image, we will effectively fall back to whatever file name we can parse out of the URL.
Thanks!