This commit adds support for unfurling static image URLs (not GIFs, not animated WebPs), such as https://placehold.co/600x400@2x.png. It also compresses images before returning them as data URIs to clients.
About compression: the compression strategy leverages the existing function images.CompressToFileLimits. A more comprehensive logic to consider the possibility of multiple image URLs being unfurled simultaneously is yet to be implemented.
Closes#3761
We should emit an error when the request to an image to be fetched
returns and HTTP error code. Otherwise, we'll run into other higher
level errors down the line, which are misleading
Example: I kept seeing "image content type not supported" errors,
although the content type *is* supported. The actual problem was that
the decode function operates on non existing image bytes.
