[bitreq] Check utf-8 while parsing JSON responses

While deserializing, `serde_json::from_slice` validates utf-8 as
needed. So instead of making two passes on the response body, one
to validate utf-8, and another to deserialize the object, we let
`serde_json::from_slice` check utf-8 as needed during deserialization.

Making a single pass over large response bodies reduces the number of
cache misses, and hence decreases the cycles taken to fully deserialize
such responses.

`Response::json` now returns `Error::SerdeJsonError` instead of
`Error::InvalidUtf8InBody` if invalid utf-8 is found during
deserialization.

For this error case, the `Error::SerdeJsonError` inner type
`serde_json::error::Error` is of category
`serde_json::error::Category::Syntax`. Other JSON syntax errors are
also assigned to this category.

We accept this loss of information given the performance gain described
above.

Author: tankyleo

bitreq/src/response.rs

@@ -248,7 +248,7 @@ impl Response {
     where
         T: serde::de::Deserialize<'a>,
     {
-        match serde_json::from_str(self.as_str()?) {
+        match serde_json::from_slice(self.as_bytes()) {
             Ok(json) => Ok(json),
             Err(err) => Err(Error::SerdeJsonError(err)),
         }