Merge pull request #827 from Tpt/escape

Make `escape` and it variants take a `impl Into<Cow<str>>` argument and implement `From<(&'a str, Cow<'a, str>)>` on `Attribute`

Author: Mingun

Changelog.md

@@ -32,13 +32,16 @@
 - [#820]: Classify output of the `Serializer` by returning an enumeration with kind of written data
 - [#823]: Do not allow serialization of consequent primitives, for example `Vec<usize>` or
   `Vec<String>` in `$value` fields. They cannot be deserialized back with the same result
+- [#827]: Make `escape` and it variants take a `impl Into<Cow<str>>` argument and implement
+  `From<(&'a str, Cow<'a, str>)>` on `Attribute`
 
 [#227]: https://github.com/tafia/quick-xml/issues/227
 [#655]: https://github.com/tafia/quick-xml/issues/655
 [#810]: https://github.com/tafia/quick-xml/pull/810
 [#811]: https://github.com/tafia/quick-xml/pull/811
 [#820]: https://github.com/tafia/quick-xml/pull/820
 [#823]: https://github.com/tafia/quick-xml/pull/823
+[#827]: https://github.com/tafia/quick-xml/pull/827
 
 
 ## 0.36.2 -- 2024-09-20
@@ -979,7 +982,7 @@ serde >= 1.0.181
 
 ## 0.16.0
 - feat: (breaking change) set failure and encoding_rs crates as optional.
-You should now use respectively `use-failure` and `encoding` features to get the old behavior
+  You should now use respectively `use-failure` and `encoding` features to get the old behavior
 - perf: improve perf using memchr3 iterator. Reading is 18% better on benches
 
 ## 0.15.0

src/escape.rs

@@ -101,7 +101,7 @@ impl std::error::Error for EscapeError {
 /// | `&`       | `&`
 /// | `'`       | `'`
 /// | `"`       | `"`
-pub fn escape(raw: &str) -> Cow<str> {
+pub fn escape<'a>(raw: impl Into<Cow<'a, str>>) -> Cow<'a, str> {
     _escape(raw, |ch| matches!(ch, b'<' | b'>' | b'&' | b'\'' | b'\"'))
 }
 
@@ -126,7 +126,7 @@ pub fn escape(raw: &str) -> Cow<str> {
 /// | `<`       | `&lt;`
 /// | `>`       | `&gt;`
 /// | `&`       | `&amp;`
-pub fn partial_escape(raw: &str) -> Cow<str> {
+pub fn partial_escape<'a>(raw: impl Into<Cow<'a, str>>) -> Cow<'a, str> {
     _escape(raw, |ch| matches!(ch, b'<' | b'>' | b'&'))
 }
 
@@ -143,13 +143,17 @@ pub fn partial_escape(raw: &str) -> Cow<str> {
 /// | `&`       | `&amp;`
 ///
 /// [requires]: https://www.w3.org/TR/xml11/#syntax
-pub fn minimal_escape(raw: &str) -> Cow<str> {
+pub fn minimal_escape<'a>(raw: impl Into<Cow<'a, str>>) -> Cow<'a, str> {
     _escape(raw, |ch| matches!(ch, b'<' | b'&'))
 }
 
 /// Escapes an `&str` and replaces a subset of xml special characters (`<`, `>`,
 /// `&`, `'`, `"`) with their corresponding xml escaped value.
-pub(crate) fn _escape<F: Fn(u8) -> bool>(raw: &str, escape_chars: F) -> Cow<str> {
+pub(crate) fn _escape<'a, F: Fn(u8) -> bool>(
+    raw: impl Into<Cow<'a, str>>,
+    escape_chars: F,
+) -> Cow<'a, str> {
+    let raw = raw.into();
     let bytes = raw.as_bytes();
     let mut escaped = None;
     let mut iter = bytes.iter();
@@ -192,7 +196,7 @@ pub(crate) fn _escape<F: Fn(u8) -> bool>(raw: &str, escape_chars: F) -> Cow<str>
         // if unsafe code will be allowed
         Cow::Owned(String::from_utf8(escaped).unwrap())
     } else {
-        Cow::Borrowed(raw)
+        raw
     }
 }
 

src/events/attributes.rs

@@ -153,6 +153,31 @@ impl<'a> From<(&'a str, &'a str)> for Attribute<'a> {
     }
 }
 
+impl<'a> From<(&'a str, Cow<'a, str>)> for Attribute<'a> {
+    /// Creates new attribute from text representation.
+    /// Key is stored as-is, but the value will be escaped.
+    ///
+    /// # Examples
+    ///
+    /// ```
+    /// # use std::borrow::Cow;
+    /// use pretty_assertions::assert_eq;
+    /// use quick_xml::events::attributes::Attribute;
+    ///
+    /// let features = Attribute::from(("features", Cow::Borrowed("Bells & whistles")));
+    /// assert_eq!(features.value, "Bells &amp; whistles".as_bytes());
+    /// ```
+    fn from(val: (&'a str, Cow<'a, str>)) -> Attribute<'a> {
+        Attribute {
+            key: QName(val.0.as_bytes()),
+            value: match escape(val.1) {
+                Cow::Borrowed(s) => Cow::Borrowed(s.as_bytes()),
+                Cow::Owned(s) => Cow::Owned(s.into_bytes()),
+            },
+        }
+    }
+}
+
 impl<'a> From<Attr<&'a [u8]>> for Attribute<'a> {
     #[inline]
     fn from(attr: Attr<&'a [u8]>) -> Self {

src/events/mod.rs

@@ -746,9 +746,8 @@ impl<'a> BytesCData<'a> {
     pub fn escape(self) -> Result<BytesText<'a>, EncodingError> {
         let decoded = self.decode()?;
         Ok(BytesText::wrap(
-            match escape(&decoded) {
-                // Because result is borrowed, no replacements was done and we can use original content
-                Cow::Borrowed(_) => self.content,
+            match escape(decoded) {
+                Cow::Borrowed(escaped) => Cow::Borrowed(escaped.as_bytes()),
                 Cow::Owned(escaped) => Cow::Owned(escaped.into_bytes()),
             },
             Decoder::utf8(),
@@ -771,9 +770,8 @@ impl<'a> BytesCData<'a> {
     pub fn partial_escape(self) -> Result<BytesText<'a>, EncodingError> {
         let decoded = self.decode()?;
         Ok(BytesText::wrap(
-            match partial_escape(&decoded) {
-                // Because result is borrowed, no replacements was done and we can use original content
-                Cow::Borrowed(_) => self.content,
+            match partial_escape(decoded) {
+                Cow::Borrowed(escaped) => Cow::Borrowed(escaped.as_bytes()),
                 Cow::Owned(escaped) => Cow::Owned(escaped.into_bytes()),
             },
             Decoder::utf8(),
@@ -795,9 +793,8 @@ impl<'a> BytesCData<'a> {
     pub fn minimal_escape(self) -> Result<BytesText<'a>, EncodingError> {
         let decoded = self.decode()?;
         Ok(BytesText::wrap(
-            match minimal_escape(&decoded) {
-                // Because result is borrowed, no replacements was done and we can use original content
-                Cow::Borrowed(_) => self.content,
+            match minimal_escape(decoded) {
+                Cow::Borrowed(escaped) => Cow::Borrowed(escaped.as_bytes()),
                 Cow::Owned(escaped) => Cow::Owned(escaped.into_bytes()),
             },
             Decoder::utf8(),