Make sure that the Docs clarify stream length is not always plaintext length

The S3EC attempts to be as "transparent" as possible when applying client-side encryption. For example, PutObject/GetObject "just works". A deeper example is ranged gets - in Java, a ranged get request will apply to the plaintext range, the customer does not need to manually adjust the range to account for encrypted blocks.

However, there is one aspect that is seemingly not possible, which is contentLength of the GetObject stream. It includes the auth tag. Customers need to read the entire thing, which is their original message length + auth tag (or padding in the case of CBC).

This needs to be well-documented.

S3EC GetObject streams と contentLength のドキュメントから始め、次に認証付き暗号化と CBC padding が顧客が読み取る必要のあるバイト数にどのような影響を与えるかを確認します。完了条件は、ドキュメントが stream の長さと元の平文の長さを明確に区別し、stream 全体を読み取る必要があることを説明していることです。

索引モデルが issue の本文から書いたものです。