Proper normalize attribute value normalization

Changelog.md

@@ -12,12 +12,25 @@
 
 ### New Features
 
+- [#379]: Improved compliance with the XML attribute value normalization process by
+  adding `Attribute::normalized_value()` and `Attribute::normalized_value_with()`,
+  which ought to be used in place of `Attribute::unescape_value()` and
+  `Attribute::unescape_value_with()`
+
 ### Bug Fixes
 
 ### Misc Changes
 
+- [#379]: Fixed error messages for some variants of `EscapeError`
+  which were incorrectly described as occurring during escaping (rather than unescaping).
 - [#650]: Change the type of `Event::PI` to a new dedicated `BytesPI` type.
 
+### New Tests
+
+- [#379]: Added tests for attribute value normalization
+
+
+[#379]: https://github.com/tafia/quick-xml/pull/379
 [#650]: https://github.com/tafia/quick-xml/issues/650
 
 
@@ -57,9 +70,6 @@ resolve predefined entities.
   - `quick_xml::escape::resolve_html5_entity`
 - [#753]: Added parser for processing instructions: `quick_xml::reader::PiParser`.
 - [#754]: Added parser for elements: `quick_xml::reader::ElementParser`.
-
-### Bug Fixes
-
 - [#622]: Fix wrong disregarding of not closed markup, such as lone `<`.
 - [#684]: Fix incorrect position reported for `Error::IllFormed(DoubleHyphenInComment)`.
 - [#684]: Fix incorrect position reported for `Error::IllFormed(MissingDoctypeName)`.
@@ -120,7 +130,6 @@ resolve predefined entities.
 [`PayloadEvent`]: https://docs.rs/quick-xml/latest/quick_xml/de/enum.PayloadEvent.html
 [`Text`]: https://docs.rs/quick-xml/latest/quick_xml/de/struct.Text.html
 
-
 ## 0.31.0 -- 2023-10-22
 
 MSRV bumped to 1.56! Crate now uses Rust 2021 edition.
@@ -290,7 +299,6 @@ serde >= 1.0.181
 - [#565]: Correctly set minimum required version of tokio dependency to 1.10
 - [#565]: Fix compilation error when build with serde <1.0.139
 
-
 [externally tagged]: https://serde.rs/enum-representations.html#externally-tagged
 [#490]: https://github.com/tafia/quick-xml/pull/490
 [#510]: https://github.com/tafia/quick-xml/issues/510

benches/macrobenches.rs

@@ -43,14 +43,13 @@ static INPUTS: &[(&str, &str)] = &[
     ("players.xml", PLAYERS),
 ];
 
-// TODO: use fully normalized attribute values
 fn parse_document_from_str(doc: &str) -> XmlResult<()> {
     let mut r = Reader::from_str(doc);
     loop {
         match criterion::black_box(r.read_event()?) {
             Event::Start(e) | Event::Empty(e) => {
                 for attr in e.attributes() {
-                    criterion::black_box(attr?.decode_and_unescape_value(&r)?);
+                    criterion::black_box(attr?.decode_and_normalize_value(&r)?);
                 }
             }
             Event::Text(e) => {
@@ -67,15 +66,14 @@ fn parse_document_from_str(doc: &str) -> XmlResult<()> {
     Ok(())
 }
 
-// TODO: use fully normalized attribute values
 fn parse_document_from_bytes(doc: &[u8]) -> XmlResult<()> {
     let mut r = Reader::from_reader(doc);
     let mut buf = Vec::new();
     loop {
         match criterion::black_box(r.read_event_into(&mut buf)?) {
             Event::Start(e) | Event::Empty(e) => {
                 for attr in e.attributes() {
-                    criterion::black_box(attr?.decode_and_unescape_value(&r)?);
+                    criterion::black_box(attr?.decode_and_normalize_value(&r)?);
                 }
             }
             Event::Text(e) => {
@@ -93,15 +91,14 @@ fn parse_document_from_bytes(doc: &[u8]) -> XmlResult<()> {
     Ok(())
 }
 
-// TODO: use fully normalized attribute values
 fn parse_document_from_str_with_namespaces(doc: &str) -> XmlResult<()> {
     let mut r = NsReader::from_str(doc);
     loop {
         match criterion::black_box(r.read_resolved_event()?) {
             (resolved_ns, Event::Start(e) | Event::Empty(e)) => {
                 criterion::black_box(resolved_ns);
                 for attr in e.attributes() {
-                    criterion::black_box(attr?.decode_and_unescape_value(&r)?);
+                    criterion::black_box(attr?.decode_and_normalize_value(&r)?);
                 }
             }
             (resolved_ns, Event::Text(e)) => {
@@ -120,7 +117,6 @@ fn parse_document_from_str_with_namespaces(doc: &str) -> XmlResult<()> {
     Ok(())
 }
 
-// TODO: use fully normalized attribute values
 fn parse_document_from_bytes_with_namespaces(doc: &[u8]) -> XmlResult<()> {
     let mut r = NsReader::from_reader(doc);
     let mut buf = Vec::new();
@@ -129,7 +125,7 @@ fn parse_document_from_bytes_with_namespaces(doc: &[u8]) -> XmlResult<()> {
             (resolved_ns, Event::Start(e) | Event::Empty(e)) => {
                 criterion::black_box(resolved_ns);
                 for attr in e.attributes() {
-                    criterion::black_box(attr?.decode_and_unescape_value(&r)?);
+                    criterion::black_box(attr?.decode_and_normalize_value(&r)?);
                 }
             }
             (resolved_ns, Event::Text(e)) => {

benches/microbenches.rs

@@ -1,6 +1,9 @@
+use std::borrow::Cow;
+
 use criterion::{self, criterion_group, criterion_main, Criterion};
 use pretty_assertions::assert_eq;
 use quick_xml::escape::{escape, unescape};
+use quick_xml::events::attributes::Attribute;
 use quick_xml::events::Event;
 use quick_xml::name::QName;
 use quick_xml::reader::{NsReader, Reader};
@@ -242,6 +245,74 @@ fn attributes(c: &mut Criterion) {
             assert_eq!(count, 150);
         })
     });
+
+    group.finish();
+}
+
+/// Benchmarks normalizing attribute values
+fn attribute_value_normalization(c: &mut Criterion) {
+    let mut group = c.benchmark_group("attribute_value_normalization");
+
+    group.bench_function("noop_short", |b| {
+        let attr = Attribute {
+            key: QName(b"foo"),
+            value: Cow::Borrowed(b"foobar"),
+        };
+        b.iter(|| {
+            criterion::black_box(attr.normalized_value()).unwrap();
+        })
+    });
+
+    group.bench_function("noop_long", |b| {
+        let attr = Attribute {
+            key: QName(b"foo"),
+            value: Cow::Borrowed(LOREM_IPSUM_TEXT.as_bytes()),
+        };
+        b.iter(|| {
+            criterion::black_box(attr.normalized_value()).unwrap();
+        })
+    });
+
+    group.bench_function("replacement_chars", |b| {
+        let attr = Attribute {
+            key: QName(b"foo"),
+            value: Cow::Borrowed(b"just a bit\n of text without\tany entities"),
+        };
+        b.iter(|| {
+            criterion::black_box(attr.normalized_value()).unwrap();
+        })
+    });
+
+    group.bench_function("char_reference", |b| {
+        let attr1 = Attribute {
+            key: QName(b"foo"),
+            value: Cow::Borrowed(b"prefix "some stuff","more stuff""),
+        };
+        let attr2 = Attribute {
+            key: QName(b"foo"),
+            value: Cow::Borrowed(b"&<"),
+        };
+        b.iter(|| {
+            criterion::black_box(attr1.normalized_value()).unwrap();
+            criterion::black_box(attr2.normalized_value()).unwrap();
+        })
+    });
+
+    group.bench_function("entity_reference", |b| {
+        let attr1 = Attribute {
+            key: QName(b"foo"),
+            value: Cow::Borrowed(b"age > 72 && age < 21"),
+        };
+        let attr2 = Attribute {
+            key: QName(b"foo"),
+            value: Cow::Borrowed(b""what's that?""),
+        };
+        b.iter(|| {
+            criterion::black_box(attr1.normalized_value()).unwrap();
+            criterion::black_box(attr2.normalized_value()).unwrap();
+        })
+    });
+
     group.finish();
 }
 
@@ -354,6 +425,7 @@ criterion_group!(
     read_resolved_event_into,
     one_event,
     attributes,
+    attribute_value_normalization,
     escaping,
     unescaping,
 );

src/errors.rs

@@ -254,6 +254,7 @@ impl From<EscapeError> for Error {
 }
 
 impl From<AttrError> for Error {
+    /// Creates a new `Error::InvalidAttr` from the given error
     #[inline]
     fn from(error: AttrError) -> Self {
         Error::InvalidAttr(error)