Merge branch 'master' into no_std

Author: domenukk

.github/workflows/main.yml

@@ -53,7 +53,7 @@ jobs:
         if: |
           matrix.os == 'windows-latest' &&
           matrix.rust == 'nightly'
-        run: cargo test --test debugger_visualizer --features "url/serde,url/debugger_visualizer" -- --test-threads=1
+        run: cargo test --test debugger_visualizer --features "url/debugger_visualizer,url_debug_tests/debugger_visualizer" -- --test-threads=1
       - name: Test `no_std` support
         run: cargo test --no-default-features --features=alloc,no_std_net
 

Cargo.toml

@@ -1,2 +1,9 @@
 [workspace]
-members = ["url", "form_urlencoded", "idna", "percent_encoding", "data-url"]
+members = [
+  "url",
+  "form_urlencoded",
+  "idna",
+  "percent_encoding",
+  "data-url",
+  "url_debug_tests",
+]

data-url/src/forgiving_base64.rs

@@ -1,10 +1,26 @@
 //! <https://infra.spec.whatwg.org/#forgiving-base64-decode>
 
 use alloc::vec::Vec;
+use core::fmt;
 
 #[derive(Debug)]
 pub struct InvalidBase64(InvalidBase64Details);
 
+impl fmt::Display for InvalidBase64 {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self.0 {
+            InvalidBase64Details::UnexpectedSymbol(code_point) => {
+                write!(f, "symbol with codepoint {} not expected", code_point)
+            }
+            InvalidBase64Details::AlphabetSymbolAfterPadding => {
+                write!(f, "alphabet symbol present after padding")
+            }
+            InvalidBase64Details::LoneAlphabetSymbol => write!(f, "lone alphabet symbol present"),
+            InvalidBase64Details::Padding => write!(f, "incorrect padding"),
+        }
+    }
+}
+
 #[derive(Debug)]
 enum InvalidBase64Details {
     UnexpectedSymbol(u8),
@@ -19,6 +35,18 @@ pub enum DecodeError<E> {
     WriteError(E),
 }
 
+impl<E: fmt::Display> fmt::Display for DecodeError<E> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self {
+            Self::InvalidBase64(inner) => write!(f, "base64 not valid: {}", inner),
+            Self::WriteError(err) => write!(f, "write error: {}", err),
+        }
+    }
+}
+
+#[cfg(feature = "std")]
+impl<E: std::error::Error> std::error::Error for DecodeError<E> {}
+
 impl<E> From<InvalidBase64Details> for DecodeError<E> {
     fn from(e: InvalidBase64Details) -> Self {
         DecodeError::InvalidBase64(InvalidBase64(e))

data-url/src/lib.rs

@@ -18,7 +18,7 @@
 
 // For forwards compatibility
 #[cfg(feature = "std")]
-extern crate std as _;
+extern crate std;
 
 #[macro_use]
 extern crate alloc;
@@ -27,6 +27,7 @@ extern crate alloc;
 compile_error!("the `alloc` feature must be enabled");
 
 use alloc::{string::String, vec::Vec};
+use core::fmt;
 
 macro_rules! require {
     ($condition: expr) => {
@@ -51,6 +52,21 @@ pub enum DataUrlError {
     NoComma,
 }
 
+impl fmt::Display for DataUrlError {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self {
+            Self::NotADataUrl => write!(f, "not a valid data url"),
+            Self::NoComma => write!(
+                f,
+                "data url is missing comma delimiting attributes and body"
+            ),
+        }
+    }
+}
+
+#[cfg(feature = "std")]
+impl std::error::Error for DataUrlError {}
+
 impl<'a> DataUrl<'a> {
     /// <https://fetch.spec.whatwg.org/#data-url-processor>
     /// but starting from a string rather than a parsed `Url`, to avoid extra string copies.

data-url/src/mime.rs

@@ -26,6 +26,15 @@ impl Mime {
 #[derive(Debug)]
 pub struct MimeParsingError(());
 
+impl fmt::Display for MimeParsingError {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "invalid mime type")
+    }
+}
+
+#[cfg(feature = "std")]
+impl std::error::Error for MimeParsingError {}
+
 /// <https://mimesniff.spec.whatwg.org/#parsing-a-mime-type>
 impl FromStr for Mime {
     type Err = MimeParsingError;

url/Cargo.toml

@@ -16,16 +16,9 @@ include = ["src/**/*", "LICENSE-*", "README.md", "tests/**"]
 edition = "2018"
 rust-version = "1.56"
 
-[badges]
-travis-ci = { repository = "servo/rust-url" }
-appveyor = { repository = "Manishearth/rust-url" }
-
 [dev-dependencies]
 serde_json = "1.0"
 bencher = "0.1"
-# To test debugger visualizers defined for the url crate such as url.natvis
-debugger_test = "0.1"
-debugger_test_parser = "0.1"
 
 [dependencies]
 form_urlencoded = { version = "1.2.0", path = "../form_urlencoded", default-features = false, features = ["alloc"] }
@@ -53,12 +46,6 @@ name = "parse_url"
 path = "benches/parse_url.rs"
 harness = false
 
-[[test]]
-name = "debugger_visualizer"
-path = "tests/debugger_visualizer.rs"
-required-features = ["debugger_visualizer"]
-test = false
-
 [package.metadata.docs.rs]
 features = ["serde"]
 

url/src/lib.rs

@@ -1552,7 +1552,7 @@ impl Url {
         if let Some(input) = fragment {
             self.fragment_start = Some(to_u32(self.serialization.len()).unwrap());
             self.serialization.push('#');
-            self.mutate(|parser| parser.parse_fragment(parser::Input::no_trim(input)))
+            self.mutate(|parser| parser.parse_fragment(parser::Input::new_no_trim(input)))
         } else {
             self.fragment_start = None;
             self.strip_trailing_spaces_from_opaque_path();
@@ -1615,7 +1615,7 @@ impl Url {
                 parser.parse_query(
                     scheme_type,
                     scheme_end,
-                    parser::Input::trim_tab_and_newlines(input, vfn),
+                    parser::Input::new_trim_tab_and_newlines(input, vfn),
                 )
             });
         } else {
@@ -1736,10 +1736,14 @@ impl Url {
                     parser.serialization.push_str("%2F");
                     path = &path[1..];
                 }
-                parser.parse_cannot_be_a_base_path(parser::Input::new(path));
+                parser.parse_cannot_be_a_base_path(parser::Input::new_no_trim(path));
             } else {
                 let mut has_host = true; // FIXME
-                parser.parse_path_start(scheme_type, &mut has_host, parser::Input::new(path));
+                parser.parse_path_start(
+                    scheme_type,
+                    &mut has_host,
+                    parser::Input::new_no_trim(path),
+                );
             }
         });
         self.restore_after_path(old_after_path_pos, &after_path);
@@ -2435,7 +2439,7 @@ impl Url {
     #[allow(clippy::result_unit_err, clippy::suspicious_operation_groupings)]
     pub fn set_scheme(&mut self, scheme: &str) -> Result<(), ()> {
         let mut parser = Parser::for_setter(String::new());
-        let remaining = parser.parse_scheme(parser::Input::new(scheme))?;
+        let remaining = parser.parse_scheme(parser::Input::new_no_trim(scheme))?;
         let new_scheme_type = SchemeType::from(&parser.serialization);
         let old_scheme_type = SchemeType::from(self.scheme());
         // If url’s scheme is a special scheme and buffer is not a special scheme, then return.

url/src/parser.rs

@@ -188,17 +188,13 @@ pub struct Input<'i> {
 }
 
 impl<'i> Input<'i> {
-    pub fn new(input: &'i str) -> Self {
-        Input::with_log(input, None)
-    }
-
-    pub fn no_trim(input: &'i str) -> Self {
+    pub fn new_no_trim(input: &'i str) -> Self {
         Input {
             chars: input.chars(),
         }
     }
 
-    pub fn trim_tab_and_newlines(
+    pub fn new_trim_tab_and_newlines(
         original_input: &'i str,
         vfn: Option<&dyn Fn(SyntaxViolation)>,
     ) -> Self {
@@ -216,7 +212,10 @@ impl<'i> Input<'i> {
         }
     }
 
-    pub fn with_log(original_input: &'i str, vfn: Option<&dyn Fn(SyntaxViolation)>) -> Self {
+    pub fn new_trim_c0_control_and_space(
+        original_input: &'i str,
+        vfn: Option<&dyn Fn(SyntaxViolation)>,
+    ) -> Self {
         let input = original_input.trim_matches(c0_control_or_space);
         if let Some(vfn) = vfn {
             if input.len() < original_input.len() {
@@ -366,7 +365,7 @@ impl<'a> Parser<'a> {
 
     /// https://url.spec.whatwg.org/#concept-basic-url-parser
     pub fn parse_url(mut self, input: &str) -> ParseResult<Url> {
-        let input = Input::with_log(input, self.violation_fn);
+        let input = Input::new_trim_c0_control_and_space(input, self.violation_fn);
         if let Ok(remaining) = self.parse_scheme(input.clone()) {
             return self.parse_with_scheme(remaining);
         }

url/src/path_segments.rs

@@ -303,7 +303,7 @@ impl<'a> PathSegmentsMut<'a> {
                     scheme_type,
                     &mut has_host,
                     path_start,
-                    parser::Input::new(segment),
+                    parser::Input::new_no_trim(segment),
                 );
             }
         });

url/src/quirks.rs

@@ -154,7 +154,7 @@ pub fn set_host(url: &mut Url, new_host: &str) -> Result<(), ()> {
     }
     // Host parsing rules are strict,
     // We don't want to trim the input
-    let input = Input::no_trim(new_host);
+    let input = Input::new_no_trim(new_host);
     let host;
     let opt_port;
     {
@@ -205,7 +205,7 @@ pub fn set_hostname(url: &mut Url, new_hostname: &str) -> Result<(), ()> {
         return Err(());
     }
     // Host parsing rules are strict we don't want to trim the input
-    let input = Input::no_trim(new_hostname);
+    let input = Input::new_no_trim(new_hostname);
     let scheme_type = SchemeType::from(url.scheme());
     if scheme_type == SchemeType::File && new_hostname.is_empty() {
         url.set_host_internal(Host::Domain(String::new()), None);
@@ -251,7 +251,7 @@ pub fn set_port(url: &mut Url, new_port: &str) -> Result<(), ()> {
             return Err(());
         }
         result = Parser::parse_port(
-            Input::new(new_port),
+            Input::new_no_trim(new_port),
             || default_port(scheme),
             Context::Setter,
         )