Suggestion: Add AccessKey / user information to S3 trait methods

[Sebastian Beyvers]

Hi, thanks for building this crate, very useful for building our own s3 compatible interface!

Unfortunately we had issues while experimenting with you S3_trait methods, because these methods lack any identity information about the user who issued the request. Would it be possible to add these information to the S3 trait ?

One suggestion would be to add the access_key to each method like this:

async fn abort_multipart_upload(&self, _input: AbortMultipartUploadInput, _access_key: String) -> S3Result<AbortMultipartUploadOutput> {
        Err(s3_error!(NotImplemented, "AbortMultipartUpload is not implemented yet"))
}

This would allow us to change or validate backend behavior based on identity information.

[Nugine]

Implemented in 4039538.
Implemented in 886a1c9.

You can try it with git dependencies.

s3s = { git = "https://github.com/Nugine/s3s", rev = "886a1c9b4798861a96014e01551a88bd6aa2ac9f" }

If everything is ok, I can release v0.4.0 for you.

[Sebastian Beyvers]

Very nice, thanks a lot for the addition. We will try to implement / integrate this now.

[Sebastian Beyvers]

Hi, just to give you a quick heads-up: In our preliminary tests everything worked as expected, especially in combination with a custom auth provider. From my side you are good to go for a v0.4.0 release. Thanks again for the addition, will make our work a lot easier !

[Nugine]

s3s v0.4.0 is now released!