feat(url): add quirks::internal_components (#788)

Co-authored-by: Aaron O'Mullan <[email protected]>

Author: littledivy

.github/workflows/main.yml

@@ -49,7 +49,7 @@ jobs:
         uses: actions-rs/cargo@v1
         with:
           command: test
-          args: --features "url/serde"
+          args: --features "url/serde,url/expose_internals"
       # The #[debugger_visualizer] attribute is currently gated behind an unstable feature flag.
       # In order to test the visualizers for the url crate, they have to be tested on a nightly build.
       - name: Run debugger_visualizer tests

url/Cargo.toml

@@ -38,6 +38,8 @@ default = ["idna"]
 # UNSTABLE FEATURES (requires Rust nightly)
 # Enable to use the #[debugger_visualizer] attribute.
 debugger_visualizer = []
+# Expose internal offsets of the URL.
+expose_internals = []
 
 [[bench]]
 name = "parse_url"

url/src/quirks.rs

@@ -14,6 +14,49 @@
 use crate::parser::{default_port, Context, Input, Parser, SchemeType};
 use crate::{Host, ParseError, Position, Url};
 
+/// Internal components / offsets of a URL.
+///
+/// https://user@pass:example.com:1234/foo/bar?baz#quux
+///      |      |    |          | ^^^^|       |   |
+///      |      |    |          | |   |       |   `----- fragment_start
+///      |      |    |          | |   |       `--------- query_start
+///      |      |    |          | |   `----------------- path_start
+///      |      |    |          | `--------------------- port
+///      |      |    |          `----------------------- host_end
+///      |      |    `---------------------------------- host_start
+///      |      `--------------------------------------- username_end
+///      `---------------------------------------------- scheme_end
+#[derive(Copy, Clone)]
+#[cfg(feature = "expose_internals")]
+pub struct InternalComponents {
+    pub scheme_end: u32,
+    pub username_end: u32,
+    pub host_start: u32,
+    pub host_end: u32,
+    pub port: Option<u16>,
+    pub path_start: u32,
+    pub query_start: Option<u32>,
+    pub fragment_start: Option<u32>,
+}
+
+/// Internal component / parsed offsets of the URL.
+///
+/// This can be useful for implementing efficient serialization
+/// for the URL.
+#[cfg(feature = "expose_internals")]
+pub fn internal_components(url: &Url) -> InternalComponents {
+    InternalComponents {
+        scheme_end: url.scheme_end,
+        username_end: url.username_end,
+        host_start: url.host_start,
+        host_end: url.host_end,
+        port: url.port,
+        path_start: url.path_start,
+        query_start: url.query_start,
+        fragment_start: url.fragment_start,
+    }
+}
+
 /// https://url.spec.whatwg.org/#dom-url-domaintoascii
 pub fn domain_to_ascii(domain: &str) -> String {
     match Host::parse(domain) {

url/tests/unit.rs

@@ -731,6 +731,34 @@ fn test_domain_encoding_quirks() {
     }
 }
 
+#[cfg(feature = "expose_internals")]
+#[test]
+fn test_expose_internals() {
+    use url::quirks::internal_components;
+    use url::quirks::InternalComponents;
+
+    let url = Url::parse("https://example.com/path/file.ext?key=val&key2=val2#fragment").unwrap();
+    let InternalComponents {
+        scheme_end,
+        username_end,
+        host_start,
+        host_end,
+        port,
+        path_start,
+        query_start,
+        fragment_start,
+    } = internal_components(&url);
+
+    assert_eq!(scheme_end, 5);
+    assert_eq!(username_end, 8);
+    assert_eq!(host_start, 8);
+    assert_eq!(host_end, 19);
+    assert_eq!(port, None);
+    assert_eq!(path_start, 19);
+    assert_eq!(query_start, Some(33));
+    assert_eq!(fragment_start, Some(51));
+}
+
 #[test]
 fn test_windows_unc_path() {
     if !cfg!(windows) {