Best headless browser for AI agents. Lite, Fast, High-Compatibility. Built in Rust
Rust ai-agentsai-toolsbrowserbrowser-automationcloud-browserheadlessheadless-browserkitesurfplaywrightpuppeteerrustservoweb-crawlerweb-crawlingweb-scraperweb-scraping
docs(content-type): clarify parser selection
ldm0e684648346d5
diff --git a/moli-content-type/src/lib.rs b/moli-content-type/src/lib.rsindex bf53e96d7..019731cee 100644--- a/moli-content-type/src/lib.rs+++ b/moli-content-type/src/lib.rs@@ -1,20 +1,21 @@//! MIME parsing for two distinct browser contexts.//!-//! Choose the entry point from the source of the value:+//! Choose the entry point from the browser algorithm being implemented://!//! - [`parse_mime_type`] applies the WHATWG MIME parsing and serialization-//! rules. Use it for standards-facing MIME operations such as Blob/File type-//! handling, MIME classification, and interpreting an already selected MIME-//! value.+//! rules. Use it when a web-platform algorithm says to parse a MIME type,+//! including Blob/File type handling, MIME classification and sniffing,+//! Fetch response MIME handling, and media capability checks.//! - [`parse_response_content_type`] matches Chromium's network-layer parsing-//! of a raw HTTP response `Content-Type` field. Use it when transport-//! metadata must retain Chromium behavior for values such as `charset` and-//! multipart `boundary`.+//! of an HTTP response `Content-Type` field. Use it only when reproducing+//! Chromium transport-metadata behavior, such as selecting a response+//! `charset` or multipart `boundary`.//!//! These parsers intentionally have different validity and recovery rules. A-//! value accepted by one is not necessarily accepted by the other, so callers-//! must select the parser from the value's origin rather than treating the two-//! result types as interchangeable.+//! response header can legitimately reach either parser depending on the+//! operation: MIME classification uses [`parse_mime_type`], while transport+//! charset selection uses [`parse_response_content_type`]. The result types+//! are therefore not interchangeable.mod response;mod whatwg;diff --git a/moli-content-type/src/response.rs b/moli-content-type/src/response.rsindex 718a650ea..1fc52be40 100644--- a/moli-content-type/src/response.rs+++ b/moli-content-type/src/response.rs@@ -40,9 +40,10 @@ impl ResponseContentType {////// This intentionally accepts malformed values seen on the network, including/// an unterminated quoted parameter and an empty subtype. Use this only for-/// response transport metadata, such as extracting `charset` or multipart-/// `boundary`; standards-facing MIME operations must use-/// [`crate::parse_mime_type`].+/// Chromium transport-metadata behavior, such as selecting a response+/// `charset` or multipart `boundary`. A value does not belong here merely+/// because it came from a response header: web-platform MIME classification+/// and sniffing use [`crate::parse_mime_type`].pub fn parse_response_content_type(input: &str) -> Option<ResponseContentType> {// Chromium treats an exact bare wildcard as meaningless, while retaining// wildcard values that have parameters or even trailing whitespace.diff --git a/moli-content-type/src/whatwg.rs b/moli-content-type/src/whatwg.rsindex 860bb1693..50dbbcf07 100644--- a/moli-content-type/src/whatwg.rs+++ b/moli-content-type/src/whatwg.rs@@ -47,10 +47,10 @@ impl fmt::Display for MimeType {/// Parses a standards-facing MIME value using WHATWG validation,/// normalization, and parameter recovery rules.///-/// Use this for Web API MIME operations and for interpreting an already-/// selected MIME value. Raw HTTP response `Content-Type` metadata must instead-/// use [`crate::parse_response_content_type`] so its Chromium network behavior-/// is preserved.+/// Use this whenever a web-platform algorithm says to parse a MIME type. That+/// includes MIME classification and sniffing even when the input originated+/// in an HTTP response header. Use [`crate::parse_response_content_type`] only+/// for Chromium's separate transport-metadata interpretation.pub fn parse_mime_type(input: &str) -> Option<MimeType> {input.parse().ok().map(|inner| MimeType { inner })}