feat: add path operators for Uri<String>

some part copy from url lib

Author: Decodetalkers

Cargo.toml

@@ -3,7 +3,7 @@ name = "fluent-uri"
 version = "0.3.2"
 authors = ["Scallop Ye <[email protected]>"]
 edition = "2021"
-rust-version = "1.65"
+rust-version = "1.80"
 description = "A generic URI/IRI handling library compliant with RFC 3986/3987."
 documentation = "https://docs.rs/fluent-uri"
 repository = "https://github.com/yescallop/fluent-uri-rs"
@@ -18,7 +18,9 @@ net = []
 
 [dependencies]
 borrow-or-share = "0.2"
+percent-encoding = "2.3.1"
 ref-cast = "1.0"
+thiserror = "2.0.12"
 
 [dependencies.serde]
 version = "1.0"

src/ri.rs

@@ -5,7 +5,7 @@ use crate::{
     },
     component::{Authority, IAuthority, Scheme},
     encoding::{encode_byte, encoder::*, EStr, Encoder},
-    error::{ParseError, ResolveError},
+    error::{BuildError, ParseError, ResolveError},
     internal::{Criteria, HostMeta, Meta, Parse, RiRef, Value},
     normalizer, parser, resolver,
 };
@@ -963,6 +963,285 @@ ri_maybe_ref! {
     FragmentEncoderType = Fragment,
 }
 
+/// Error
+#[derive(Debug, thiserror::Error)]
+pub enum UriError {
+    #[error("No Segments")]
+    NoSegments,
+    #[error("HostError")]
+    HostError,
+    #[error("build error")]
+    BuildError(#[from] BuildError),
+    #[error("not absolute path")]
+    NotAbsolutePath,
+    #[error("Path illegal")]
+    IllegalPath,
+}
+const SCHEME_FILE: &Scheme = Scheme::new_or_panic("file");
+impl<'i, 'o, T: BorrowOrShare<'i, 'o, str>> Uri<T> {
+    /// Assuming the URL is in the `file` scheme or similar,
+    /// convert its path to an absolute `std::path::Path`.
+    ///
+    /// **Note:** This does not actually check the URL’s `scheme`,
+    /// and may give nonsensical results for other schemes.
+    /// It is the user’s responsibility to check the URL’s scheme before calling this.
+    ///
+    /// ```
+    /// # use fluent_uri::Uri;
+    /// # let url = Uri::parse("file:///etc/passwd").unwrap();
+    /// let path = url.to_file_path();
+    /// ```
+    ///
+    /// # Errors
+    /// Returns `Err` if the host is neither empty nor `"localhost"` (except on Windows, where
+    /// `file:` URLs may have a non-local host),
+    /// or if `Path::new_opt()` returns `None`.
+    /// (That is, if the percent-decoded path contains a NUL byte or,
+    /// for a Windows path, is not UTF-8.)
+    ///
+    /// This method is only available if the `std` Cargo feature is enabled.
+    ///
+    pub fn to_file_path(&'i self) -> Result<std::path::PathBuf, UriError> {
+        let segments = self.path();
+        let segments = segments
+            .segments_if_absolute()
+            .ok_or(UriError::NoSegments)?;
+        let host: Option<&str> = match self.authority().map(|authority| authority.host()) {
+            None | Some("localhost") => None,
+            Some(host_data) if self.scheme().as_str() == "file" => Some(host_data),
+            Some(_) => return Err(UriError::NoSegments),
+        };
+        file_url_segments_to_pathbuf(host, segments)
+    }
+}
+
+mod control_chars {
+    use percent_encoding::AsciiSet;
+    /// https://url.spec.whatwg.org/#fragment-percent-encode-set
+    const FRAGMENT: &AsciiSet = &percent_encoding::CONTROLS
+        .add(b' ')
+        .add(b'"')
+        .add(b'<')
+        .add(b'>')
+        .add(b'`');
+
+    /// https://url.spec.whatwg.org/#path-percent-encode-set
+    const PATH: &AsciiSet = &FRAGMENT.add(b'#').add(b'?').add(b'{').add(b'}');
+    pub(crate) const PATH_SEGMENT: &AsciiSet = &PATH.add(b'/').add(b'%');
+
+    pub(crate) const SPECIAL_PATH_SEGMENT: &AsciiSet = &PATH_SEGMENT.add(b'\\');
+}
+
+impl Uri<String> {
+    /// Convert a file name as `std::path::Path` into an URL in the `file` scheme.
+    ///
+    /// This returns `Err` if the given path is not absolute or,
+    /// on Windows, if the prefix is not a disk prefix (e.g. `C:`) or a UNC prefix (`\\`).
+    ///
+    /// # Examples
+    ///
+    /// On Unix-like platforms:
+    ///
+    /// ```
+    /// # if cfg!(unix) {
+    /// # use fluent_uri::Uri;
+    ///
+    /// let uri = Uri::from_file_path("/tmp/foo.txt").unwrap();
+    /// assert_eq!(uri.as_str(), "file:///tmp/foo.txt");
+    ///
+    /// let uri = Uri::from_file_path("../foo.txt");
+    /// assert!(uri.is_err());
+    ///
+    /// let uri = Uri::from_file_path("https://google.com/");
+    /// assert!(uri.is_err());
+    /// # }
+    /// ```
+    ///
+    /// # Errors
+    ///
+    /// Will return error when the path is illegal
+    pub fn from_file_path<P: AsRef<std::path::Path>>(path: P) -> Result<Self, UriError> {
+        use control_chars::*;
+        use percent_encoding::percent_encode;
+        use std::os::unix::ffi::OsStrExt;
+        let path = path.as_ref();
+        if !path.is_absolute() {
+            return Err(UriError::NotAbsolutePath);
+        }
+        let mut serialization = "".to_owned();
+        let mut empty = true;
+        for component in path.components().skip(1) {
+            empty = false;
+            serialization.push('/');
+            #[cfg(not(target_os = "wasi"))]
+            serialization.extend(percent_encode(
+                component.as_os_str().as_bytes(),
+                SPECIAL_PATH_SEGMENT,
+            ));
+
+            #[cfg(target_os = "wasi")]
+            serialization.extend(percent_encode(
+                component.as_os_str().to_string_lossy().as_bytes(),
+                SPECIAL_PATH_SEGMENT,
+            ));
+        }
+        if empty {
+            serialization.push('/');
+        }
+        let path = EStr::new(&serialization).ok_or(UriError::IllegalPath)?;
+        Ok(Self::builder()
+            .scheme(SCHEME_FILE)
+            .authority(Authority::EMPTY)
+            .path(path)
+            .build()?)
+    }
+}
+
+use crate::encoding::Split;
+#[cfg(all(
+    feature = "std",
+    any(unix, target_os = "redox", target_os = "wasi", target_os = "hermit")
+))]
+fn file_url_segments_to_pathbuf(
+    host: Option<&str>,
+
+    segments: Split<'_, Path>,
+) -> Result<std::path::PathBuf, UriError> {
+    use alloc::vec::Vec;
+
+    use percent_encoding::percent_decode;
+
+    #[cfg(not(target_os = "wasi"))]
+    use std::ffi::OsStr;
+
+    #[cfg(target_os = "hermit")]
+    use std::os::hermit::ffi::OsStrExt;
+
+    #[cfg(any(unix, target_os = "redox"))]
+    use std::os::unix::prelude::OsStrExt;
+
+    use std::path::PathBuf;
+
+    if host.is_some_and(|host| !host.is_empty()) {
+        return Err(UriError::HostError);
+    }
+
+    let mut bytes = if cfg!(target_os = "redox") {
+        b"file:".to_vec()
+    } else {
+        Vec::new()
+    };
+
+    for segment in segments {
+        bytes.push(b'/');
+
+        bytes.extend(percent_decode(segment.as_str().as_bytes()));
+    }
+
+    // A windows drive letter must end with a slash.
+
+    if bytes.len() > 2
+        && bytes[bytes.len() - 2].is_ascii_alphabetic()
+        && matches!(bytes[bytes.len() - 1], b':' | b'|')
+    {
+        bytes.push(b'/');
+    }
+
+    #[cfg(not(target_os = "wasi"))]
+    let path = PathBuf::from(OsStr::from_bytes(&bytes));
+
+    #[cfg(target_os = "wasi")]
+    let path = String::from_utf8(bytes)
+        .map(|path| PathBuf::from(path))
+        .map_err(|_| ())?;
+
+    debug_assert!(
+        path.is_absolute(),
+        "to_file_path() failed to produce an absolute Path"
+    );
+
+    Ok(path)
+}
+
+#[cfg(all(feature = "std", windows))]
+fn file_url_segments_to_pathbuf(
+    host: Option<&str>,
+    segments: Split<'_, Path>,
+) -> Result<std::path::PathBuf, UriError> {
+    file_url_segments_to_pathbuf_windows(host, segments)
+}
+
+/// https://url.spec.whatwg.org/#ascii-alpha
+#[allow(unused)]
+#[inline]
+fn ascii_alpha(ch: char) -> bool {
+    ch.is_ascii_alphabetic()
+}
+
+// Build this unconditionally to alleviate https://github.com/servo/rust-url/issues/102
+#[cfg(feature = "std")]
+#[cfg_attr(not(windows), allow(dead_code))]
+fn file_url_segments_to_pathbuf_windows(
+    host: Option<&str>,
+    mut segments: Split<'_, Path>,
+) -> Result<std::path::PathBuf, UriError> {
+    use percent_encoding::percent_decode;
+    use std::path::PathBuf;
+
+    let mut string = if let Some(host) = host {
+        r"\\".to_owned() + host
+    } else {
+        let first = segments.next().ok_or(UriError::HostError)?.as_str();
+
+        match first.len() {
+            2 => {
+                if !first.starts_with(ascii_alpha) || first.as_bytes()[1] != b':' {
+                    return Err(UriError::HostError);
+                }
+
+                first.to_owned()
+            }
+
+            4 => {
+                if !first.starts_with(ascii_alpha) {
+                    return Err(UriError::HostError);
+                }
+
+                let bytes = first.as_bytes();
+
+                if bytes[1] != b'%' || bytes[2] != b'3' || (bytes[3] != b'a' && bytes[3] != b'A') {
+                    return Err(UriError::HostError);
+                }
+
+                first[0..1].to_owned() + ":"
+            }
+
+            _ => return Err(UriError::HostError),
+        }
+    };
+
+    for segment in segments.map(|seg| seg.as_str()) {
+        string.push('\\');
+
+        // Currently non-unicode windows paths cannot be represented
+
+        match String::from_utf8(percent_decode(segment.as_bytes()).collect()) {
+            Ok(s) => string.push_str(&s),
+
+            Err(..) => return Err(UriError::HostError),
+        }
+    }
+
+    let path = PathBuf::from(string);
+
+    debug_assert!(
+        path.is_absolute(),
+        "to_file_path() failed to produce an absolute Path"
+    );
+
+    Ok(path)
+}
+
 ri_maybe_ref! {
     Type = UriRef,
     type_name = "UriRef",

tests/to_path_buf.rs

@@ -0,0 +1,14 @@
+use fluent_uri::Uri;
+use std::path::Path;
+#[test]
+fn test_path_roundtrip_conversion() {
+    let src = std::fs::canonicalize(Path::new(".")).unwrap();
+    let conv: Uri<String> = Uri::from_file_path(&src).unwrap();
+
+    let roundtrip = conv.to_file_path().unwrap();
+    assert_eq!(src, roundtrip, "conv={conv:?}",);
+
+    let url = Uri::from_file_path("/tmp/foo.txt").unwrap();
+
+    assert_eq!(url.as_str(), "file:///tmp/foo.txt");
+}