Finding (Medium). Text parts are decoded twice — once via get_body_raw() and again via get_body() (mail_parser.rs:53–65). Storing the raw bytes once and deriving the string removes ~10–20% of body-decode overhead.
Acceptance
Audit ref: M2-T3 (do after the classifier rewrite).
Finding (Medium). Text parts are decoded twice — once via
get_body_raw()and again viaget_body()(mail_parser.rs:53–65). Storing the raw bytes once and deriving the string removes ~10–20% of body-decode overhead.Acceptance
get_body_raw()is called once per partAudit ref: M2-T3 (do after the classifier rewrite).