Defering disk writes is still probably a good idea, but unfortunately
there are some tradeoffs with rust's async story that make it non-ideal.
Ideally we would defer writes, but have a Drop impl on DiskStorage that
waited for all the deferred writes to complete. While it's trival to
create a future that waits for all deferred writes to finish it's not
currently possible to write a Drop impl that waits on a future.
It would be possible to write an inherent async fn that takes self by
value and waits for the writes, but that method would need to be
propogated up all the way to users of the library and they would need to
remember to invoke it before dropping the Authenticator.
This Removes RefreshError and PollError. Both those types can be fully
represented within Error and there seems little value in distinguishing
that they were resulting from device polling or refreshes. In either
case the user will need to handle the response from token() calls
similarly. This also removes the AuthenticatorDelegate since it only
served to notify users when refreshes failed, which can already be done
by looking at the return code from token. DeviceFlow no longer has the
ability to set a wait_timeout. This is trivial to do by wrapping the
token() call in a tokio::Timeout future so there's little benefit for
users specifying this value. The DeviceFlowDelegate also no longer has
the ability to specify when to abort, or alter the interval polling
happens on, but it does gain understanding of the 'slow_down' response
as documented in the oauth rfc. It seemed very unlikely the delegate was
going to do anything other that timeout after a given time and that's
already possible using tokio::Timeout so it needlessly complicated the
implementation.
BadServerResponse didn't seem to be adequately different from
NegativeServerResponse so it's been removed. NegativeServerResponse is
now a struct variant with field names 'error' and 'error_description' to
self document what it contains.
InvalidClient and InvalidScope were only ever created based on string
parsing of the returned error message from the server. This seemed
overly specific to a particular implementation and didn't provide much
value to callers. Printing a NegativeServerResponse would provide users
the same information.
The caching layer never returns errors anymore so remove that variant.
No more need to macro_use serde. Order the imports consistently (albeit
somewhat arbitrary), starting with items from this crate, followed by
std, followed by external crates.
RequestError is the error value that encompasses all errors from the
authenticators. Their is an established convention of using Error as the
name for those types.
Restructure the modules and imports to increase the signal to noise
ration on the cargo doc landing page. This includes exposing some
modules as public so that they can contain things that need to be public
but that users will rarely need to interact with. Most items from
types.rs were moved into an error.rs module that is now exposed
publicly.
