The /camli/enumerate-blobs endpoint enumerates all blobs that the
server knows about.

They're returned in sorted order, sorted by (digest_type,
digest_value).  That is, md5-acbd18db4cc2f85cedef654fccc4a4d8 sorts
before sha1-0beec7b5ea3f0fdbc95d0dd47f3c5bc275da8a33 because "m" sorts
before "s", even though "0" sorts before "a".

GET /camli/enumerate-blobs?after=&limit= HTTP/1.1
Host: example.com

URL GET parameters:

     after     optional    If provided, only blobs GREATER THAN this
                           value are returned.

     limit     optional    Limit the number of returned blobrefs.  The
                           server may have its own lower limit, however,
                           so be sure to pay attention to the presence
                           of a "continueAfter" key in the JSON response.

Response:

HTTP/1.1 200 OK
Content-Type: text/javascript

{
  "blobs": [
    {"blobRef": "md5-acbd18db4cc2f85cedef654fccc4a4d8",
     "size": 3},
    {"blobRef": "sha1-0beec7b5ea3f0fdbc95d0dd47f3c5bc275da8a33",
     "size": 3},
  ],
  "after": "sha1-0beec7b5ea3f0fdbc95d0dd47f3c5bc275da8a33",
}

Response keys:

   blobs          required   Array of {"blobRef": BLOBREF, "size": INT_bytes}
                             will be an empty list if no blobs are present.

   after          optional   If present, the result is truncated and there are
                             are more blobs after the provided blobref, which
                             should be passed to the next request's "after"
                             request parameter.