implement element writing

Cargo.lock

@@ -17,6 +17,17 @@ version = "1.0.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f26201604c87b1e01bd3d98f8d5d9a8fcbb815e8cedb41ffccbeb4bf593a35fe"
 
+[[package]]
+name = "async-recursion"
+version = "1.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3b43422f69d8ff38f95f1b2bb76517c91589a924d1559a0e935d7c8ce0274c11"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn",
+]
+
 [[package]]
 name = "autocfg"
 version = "1.1.0"
@@ -272,6 +283,7 @@ dependencies = [
 name = "peanuts"
 version = "0.1.0"
 dependencies = [
+ "async-recursion",
  "circular",
  "futures",
  "nom",

Cargo.toml

@@ -6,6 +6,7 @@ edition = "2021"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
+async-recursion = "1.1.1"
 circular = { version = "0.3.0", path = "../circular" }
 futures = "0.3.30"
 nom = "7.1.3"

src/element.rs

@@ -1,10 +1,14 @@
 // elements resemble a final tree, including inherited namespace information
 
-use std::collections::{HashMap, HashSet};
+use std::{
+    collections::{HashMap, HashSet},
+    convert::Infallible,
+    str::FromStr,
+};
 
 use crate::{
     error::Error,
-    xml::{self, Attribute},
+    xml::{self, parsers_complete::Parser, Attribute},
 };
 
 // when are namespaces names chosen then if they are automatically calculated
@@ -18,7 +22,7 @@ pub struct NamespaceDeclaration {
 // names are qualified, they contain a reference to the namespace (held within the reader/writer)
 #[derive(PartialEq, Eq, Hash, Clone, Debug)]
 pub struct Name {
-    pub namespace: String,
+    pub namespace: Option<String>,
     pub local_name: String,
 }
 
@@ -50,6 +54,33 @@ pub struct Element {
     pub content: Vec<Content>,
 }
 
+pub fn escape_str(s: &str) -> String {
+    let mut string = String::new();
+    for str in s.split_inclusive(|c| c == '<' || c == '&' || c == '>') {
+        if let Some(str) = str.strip_suffix('<') {
+            if !str.is_empty() {
+                string.push_str(str)
+            }
+            string.push_str("&lt;");
+        } else if let Some(str) = str.strip_suffix('&') {
+            if !str.is_empty() {
+                string.push_str(str)
+            }
+            string.push_str("&amp;");
+        } else if let Some(str) = str.strip_suffix('>') {
+            if !str.is_empty() {
+                string.push_str(str)
+            }
+            string.push_str("&gt;");
+        } else {
+            if !str.is_empty() {
+                string.push_str(str)
+            }
+        }
+    }
+    string
+}
+
 // impl<'s> TryFrom<xml::Element<'s>> for Element<'s> {
 //     type Error = Error;
 

src/error.rs

@@ -16,6 +16,7 @@ pub enum Error {
     MismatchedEndTag(Name, Name),
     NotInElement(String),
     ExtraData(String),
+    UndeclaredNamespace(String),
 }
 
 impl From<std::io::Error> for Error {

src/reader.rs

@@ -48,7 +48,7 @@ where
         Ok(self.inner.read_buf(&mut self.buffer).await?)
     }
 
-    async fn read_prolog<'s>(&'s mut self) -> Result<()> {
+    pub async fn read_prolog<'s>(&'s mut self) -> Result<()> {
         loop {
             self.read_buf().await?;
             let input = str::from_utf8(self.buffer.data())?;
@@ -68,7 +68,7 @@ where
         }
     }
 
-    async fn read_start_tag<'s>(&'s mut self) -> Result<Element> {
+    pub async fn read_start_tag<'s>(&'s mut self) -> Result<Element> {
         loop {
             self.read_buf().await?;
             let input = str::from_utf8(self.buffer.data())?;
@@ -93,7 +93,7 @@ where
         }
     }
 
-    async fn read_end_tag<'s>(&'s mut self) -> Result<()> {
+    pub async fn read_end_tag<'s>(&'s mut self) -> Result<()> {
         loop {
             self.read_buf().await?;
             let input = str::from_utf8(self.buffer.data())?;
@@ -118,7 +118,7 @@ where
         }
     }
 
-    async fn read_element<'s>(&'s mut self) -> Result<Element> {
+    pub async fn read_element<'s>(&'s mut self) -> Result<Element> {
         loop {
             self.read_buf().await?;
             let input = str::from_utf8(self.buffer.data())?;
@@ -140,7 +140,7 @@ where
         }
     }
 
-    async fn read_content<'s>(&'s mut self) -> Result<Content> {
+    pub async fn read_content<'s>(&'s mut self) -> Result<Content> {
         let mut last_char = false;
         let mut text = String::new();
         loop {
@@ -312,7 +312,7 @@ impl<R> Reader<R> {
             .chain(element_namespace_declarations.iter())
             .collect();
 
-        // element name and default attribute namespace
+        // element name
 
         let element_namespace_declaration;
         let element_local_name = s_tag.name.local_part().to_string();
@@ -330,10 +330,8 @@ impl<R> Reader<R> {
             }
         }
 
-        let element_default_namespace = element_namespace_declaration
+        let element_default_namespace =
-            .ok_or_else(|| Error::UnqualifiedNamespace(s_tag.name.to_string()))?
+            element_namespace_declaration.map(|decl| decl.namespace.clone());
-            .namespace
-            .clone();
 
         let element_name = Name {
             namespace: element_default_namespace,
@@ -361,21 +359,25 @@ impl<R> Reader<R> {
                                 namespace_declaration.prefix.as_deref() == Some(prefix)
                             });
                 }
-                None => attribute_namespace_declaration = element_namespace_declaration,
+                None => attribute_namespace_declaration = None,
             }
+            let name;
             if let Some(namespace_declaration) = attribute_namespace_declaration {
-                let name = Name {
+                name = Name {
-                    namespace: namespace_declaration.namespace.clone(),
+                    namespace: Some(namespace_declaration.namespace.clone()),
                     local_name: attribute_local_name,
                 };
+            } else {
+                name = Name {
+                    namespace: None,
+                    local_name: attribute_local_name,
+                };
+            }
             let value = value.process()?;
             // check for duplicate attribute
             if let Some(_value) = attributes.insert(name, value) {
                 return Err(Error::DuplicateAttribute(q_name.to_string()));
             }
-            } else {
-                return Err(Error::UnqualifiedNamespace(q_name.to_string()));
-            }
         }
 
         depth.push(element_name.clone());
@@ -415,8 +417,7 @@ impl<R> Reader<R> {
             }
 
             let e_tag_namespace = e_tag_namespace_declaration
-                .ok_or_else(|| Error::UnqualifiedNamespace(xml_e_tag.name.to_string()))?
+                .map(|decl| decl.namespace.clone())
-                .namespace
                 .clone();
 
             let e_tag_name = Name {
@@ -512,10 +513,8 @@ impl<R> Reader<R> {
             }
         }
 
-        let element_default_namespace = element_namespace_declaration
+        let element_default_namespace =
-            .ok_or_else(|| Error::UnqualifiedNamespace(xml_name.to_string()))?
+            element_namespace_declaration.map(|decl| decl.namespace.clone());
-            .namespace
-            .clone();
 
         let element_name = Name {
             namespace: element_default_namespace,
@@ -541,10 +540,7 @@ impl<R> Reader<R> {
                 }
             }
 
-            let e_tag_namespace = e_tag_namespace_declaration
+            let e_tag_namespace = e_tag_namespace_declaration.map(|decl| decl.namespace.clone());
-                .ok_or_else(|| Error::UnqualifiedNamespace(xml_name.to_string()))?
-                .namespace
-                .clone();
 
             let e_tag_name = Name {
                 namespace: e_tag_namespace,
@@ -577,21 +573,25 @@ impl<R> Reader<R> {
                                 namespace_declaration.prefix.as_deref() == Some(prefix)
                             });
                 }
-                None => attribute_namespace_declaration = element_namespace_declaration,
+                None => attribute_namespace_declaration = None,
             }
+            let name;
             if let Some(namespace_declaration) = attribute_namespace_declaration {
-                let name = Name {
+                name = Name {
-                    namespace: namespace_declaration.namespace.clone(),
+                    namespace: Some(namespace_declaration.namespace.clone()),
                     local_name: attribute_local_name,
                 };
+            } else {
+                name = Name {
+                    namespace: None,
+                    local_name: attribute_local_name,
+                };
+            }
             let value = value.process()?;
             // check for duplicate attribute
             if let Some(_value) = attributes.insert(name, value) {
                 return Err(Error::DuplicateAttribute(q_name.to_string()));
             }
-            } else {
-                return Err(Error::UnqualifiedNamespace(q_name.to_string()));
-            }
         }
 
         let content;
@@ -674,19 +674,21 @@ impl<R: AsyncRead + Unpin> Stream for Reader<R> {
 }
 
 #[cfg(test)]
-mod test {
+pub(crate) mod test {
     use futures::{sink::Buffer, StreamExt};
     use tokio::io::AsyncRead;
 
+    use crate::element::Element;
+
     use super::Reader;
 
-    struct MockAsyncReader<'s> {
+    pub struct MockAsyncReader<'s> {
         put: bool,
         data: &'s str,
     }
 
     impl<'s> MockAsyncReader<'s> {
-        fn new(data: &'s str) -> Self {
+        pub fn new(data: &'s str) -> Self {
             Self { put: false, data }
         }
     }
@@ -705,7 +707,7 @@ mod test {
         }
     }
 
-    const TEST_DOC: &'static str = "<xs:schema
+    pub const TEST_DOC: &'static str = "<xs:schema
        xmlns:xs='http://www.w3.org/2001/XMLSchema'
        targetNamespace='http://etherx.jabber.org/streams'
        xmlns='http://etherx.jabber.org/streams'

src/writer.rs

@@ -1,53 +1,275 @@
-use std::collections::HashSet;
+use std::{collections::HashSet, str::FromStr};
 
+use async_recursion::async_recursion;
 use futures::Sink;
-use tokio::io::AsyncWrite;
+use tokio::io::{AsyncWrite, AsyncWriteExt};
 
 use crate::{
-    element::{Element, Name, NamespaceDeclaration},
+    element::{escape_str, Content, Element, Name, NamespaceDeclaration},
     error::Error,
     xml::{self, composers::Composer, parsers_complete::Parser, ETag},
+    Result,
 };
 
 // pub struct Writer<W, C = Composer> {
 pub struct Writer<W> {
     inner: W,
     depth: Vec<Name>,
-    namespaces: Vec<HashSet<NamespaceDeclaration>>,
+    namespace_declarations: Vec<HashSet<NamespaceDeclaration>>,
 }
 
-impl<W: AsyncWrite + Unpin> Writer<W> {
+impl<W> Writer<W> {
-    pub async fn write(&mut self, element: Element) -> Result<(), Error> {
+    pub fn new(writer: W) -> Self {
-        todo!()
+        Self {
+            inner: writer,
+            depth: Vec::new(),
+            namespace_declarations: Vec::new(),
+        }
+    }
 }
 
-    pub async fn write_start(&mut self, element: Element) -> Result<(), Error> {
+impl<W: AsyncWrite + Unpin + Send> Writer<W> {
-        todo!()
+    #[async_recursion]
+    pub async fn write_element(&mut self, element: &Element) -> Result<()> {
+        if element.content.is_empty() {
+            self.write_empty(element).await?;
+        } else {
+            self.write_start(element).await?;
+            for content in &element.content {
+                self.write_content(content).await?;
+            }
+            self.write_end().await?;
+        }
+        Ok(())
     }
 
-    pub async fn write_end(&mut self) -> Result<(), Error> {
+    pub async fn write_empty(&mut self, element: &Element) -> Result<()> {
-        todo!()
+        let namespace_declarations_stack: Vec<_> = self
-        // let e_tag;
+            .namespace_declarations
-        // if let Some(name) = self.depth.pop() {
+            .iter()
-        //     if let Some(prefix) = name.namespace.prefix {
+            .flatten()
-        //         e_tag = xml::ETag {
+            .chain(&element.namespace_declarations)
-        //             name: xml::QName::PrefixedName(xml::PrefixedName {
+            .collect();
-        //                 prefix: xml::Prefix::parse_full(&prefix)?,
+
-        //                 local_part: xml::LocalPart::parse_full(&name.name)?,
+        let prefix;
-        //             }),
+        if let Some(namespace) = &element.name.namespace {
-        //         };
+            let name_namespace_declaration = namespace_declarations_stack
-        //         e_tag.write(&mut self.inner).await?;
+                .iter()
-        //         Ok(())
+                .rfind(|namespace_declaration| namespace_declaration.namespace == *namespace)
-        //     } else {
+                .ok_or(Error::UndeclaredNamespace(namespace.clone()))?;
-        //         e_tag = xml::ETag {
+            prefix = name_namespace_declaration.prefix.as_ref();
-        //             name: xml::QName::UnprefixedName(xml::UnprefixedName::parse_full(&name.name)?),
+        } else {
-        //         };
+            prefix = None
-        //         e_tag.write(&mut self.inner).await?;
+        }
-        //         Ok(())
+
-        //     }
+        let name;
-        // } else {
+        if let Some(prefix) = &prefix {
-        //     return Err(Error::NotInElement("".to_string()));
+            name = xml::QName::PrefixedName(xml::PrefixedName {
-        // }
+                prefix: xml::Prefix::parse_full(prefix)?,
+                local_part: xml::LocalPart::parse_full(&element.name.local_name)?,
+            })
+        } else {
+            name = xml::QName::UnprefixedName(xml::UnprefixedName::parse_full(
+                &element.name.local_name,
+            )?)
+        }
+
+        let mut attributes = Vec::new();
+
+        for namespace_declaration in &element.namespace_declarations {
+            let ns_name = namespace_declaration
+                .prefix
+                .as_ref()
+                .map(|prefix| -> Result<_> {
+                    Ok(xml::NSAttName::PrefixedAttName(xml::PrefixedAttName(
+                        xml::NCName::parse_full(&prefix)?,
+                    )))
+                })
+                .unwrap_or(Ok(xml::NSAttName::DefaultAttName))?;
+            let value = xml::AttValue::from(namespace_declaration.namespace.as_str());
+            let xml_attribute = xml::Attribute::NamespaceDeclaration { ns_name, value };
+            attributes.push(xml_attribute);
+        }
+
+        for (name, value) in &element.attributes {
+            let prefix;
+            if let Some(namespace) = &name.namespace {
+                let name_namespace_declaration = namespace_declarations_stack
+                    .iter()
+                    .rfind(|namespace_declaration| namespace_declaration.namespace == *namespace)
+                    .ok_or(Error::UndeclaredNamespace(namespace.clone()))?;
+                prefix = name_namespace_declaration.prefix.as_ref();
+            } else {
+                prefix = None
+            }
+
+            let att_name;
+            if let Some(prefix) = &prefix {
+                att_name = xml::QName::PrefixedName(xml::PrefixedName {
+                    prefix: xml::Prefix::parse_full(prefix)?,
+                    local_part: xml::LocalPart::parse_full(&element.name.local_name)?,
+                })
+            } else {
+                att_name = xml::QName::UnprefixedName(xml::UnprefixedName::parse_full(
+                    &element.name.local_name,
+                )?)
+            }
+
+            let value = xml::AttValue::from(value.as_str());
+
+            let xml_attribute = xml::Attribute::Attribute {
+                name: att_name,
+                value,
+            };
+            attributes.push(xml_attribute);
+        }
+
+        let tag = xml::EmptyElemTag { name, attributes };
+
+        tag.write(&mut self.inner).await?;
+
+        Ok(())
+    }
+
+    pub async fn write_start(&mut self, element: &Element) -> Result<()> {
+        let namespace_declarations_stack: Vec<_> = self
+            .namespace_declarations
+            .iter()
+            .flatten()
+            .chain(&element.namespace_declarations)
+            .collect();
+
+        let prefix;
+        if let Some(namespace) = &element.name.namespace {
+            let name_namespace_declaration = namespace_declarations_stack
+                .iter()
+                .rfind(|namespace_declaration| namespace_declaration.namespace == *namespace)
+                .ok_or(Error::UndeclaredNamespace(namespace.clone()))?;
+            prefix = name_namespace_declaration.prefix.as_ref();
+        } else {
+            prefix = None
+        }
+
+        let name;
+        if let Some(prefix) = &prefix {
+            name = xml::QName::PrefixedName(xml::PrefixedName {
+                prefix: xml::Prefix::parse_full(prefix)?,
+                local_part: xml::LocalPart::parse_full(&element.name.local_name)?,
+            })
+        } else {
+            name = xml::QName::UnprefixedName(xml::UnprefixedName::parse_full(
+                &element.name.local_name,
+            )?)
+        }
+
+        let mut attributes = Vec::new();
+
+        for namespace_declaration in &element.namespace_declarations {
+            let ns_name = namespace_declaration
+                .prefix
+                .as_ref()
+                .map(|prefix| -> Result<_> {
+                    Ok(xml::NSAttName::PrefixedAttName(xml::PrefixedAttName(
+                        xml::NCName::parse_full(&prefix)?,
+                    )))
+                })
+                .unwrap_or(Ok(xml::NSAttName::DefaultAttName))?;
+            let value = xml::AttValue::from(namespace_declaration.namespace.as_str());
+            let xml_attribute = xml::Attribute::NamespaceDeclaration { ns_name, value };
+            attributes.push(xml_attribute);
+        }
+
+        for (name, value) in &element.attributes {
+            let prefix;
+            if let Some(namespace) = &name.namespace {
+                let name_namespace_declaration = namespace_declarations_stack
+                    .iter()
+                    .rfind(|namespace_declaration| namespace_declaration.namespace == *namespace)
+                    .ok_or(Error::UndeclaredNamespace(namespace.clone()))?;
+                prefix = name_namespace_declaration.prefix.as_ref();
+            } else {
+                prefix = None
+            }
+
+            let att_name;
+            if let Some(prefix) = &prefix {
+                att_name = xml::QName::PrefixedName(xml::PrefixedName {
+                    prefix: xml::Prefix::parse_full(prefix)?,
+                    local_part: xml::LocalPart::parse_full(&element.name.local_name)?,
+                })
+            } else {
+                att_name = xml::QName::UnprefixedName(xml::UnprefixedName::parse_full(
+                    &element.name.local_name,
+                )?)
+            }
+
+            let value = xml::AttValue::from(value.as_str());
+
+            let xml_attribute = xml::Attribute::Attribute {
+                name: att_name,
+                value,
+            };
+            attributes.push(xml_attribute);
+        }
+
+        let s_tag = xml::STag { name, attributes };
+
+        s_tag.write(&mut self.inner).await?;
+
+        self.depth.push(element.name.clone());
+        self.namespace_declarations
+            .push(element.namespace_declarations.clone());
+        Ok(())
+    }
+
+    pub async fn write_content(&mut self, content: &Content) -> Result<()> {
+        match content {
+            Content::Element(element) => self.write_element(element).await?,
+            Content::Text(text) => self.inner.write_all(escape_str(text).as_bytes()).await?,
+            // TODO: comments and PI
+            Content::PI => {}
+            Content::Comment(_) => {}
+        }
+        Ok(())
+    }
+
+    pub async fn write_end(&mut self) -> Result<()> {
+        if let Some(name) = &self.depth.pop() {
+            let e_tag;
+            let namespace_declarations_stack: Vec<_> =
+                self.namespace_declarations.iter().flatten().collect();
+
+            let prefix;
+            if let Some(namespace) = &name.namespace {
+                let name_namespace_declaration = namespace_declarations_stack
+                    .iter()
+                    .rfind(|namespace_declaration| namespace_declaration.namespace == *namespace)
+                    .ok_or(Error::UndeclaredNamespace(namespace.clone()))?;
+                prefix = name_namespace_declaration.prefix.as_ref();
+            } else {
+                prefix = None
+            }
+
+            if let Some(prefix) = &prefix {
+                e_tag = xml::ETag {
+                    name: xml::QName::PrefixedName(xml::PrefixedName {
+                        prefix: xml::Prefix::parse_full(prefix)?,
+                        local_part: xml::LocalPart::parse_full(&name.local_name)?,
+                    }),
+                };
+            } else {
+                e_tag = xml::ETag {
+                    name: xml::QName::UnprefixedName(xml::UnprefixedName::parse_full(
+                        &name.local_name,
+                    )?),
+                };
+            }
+            e_tag.write(&mut self.inner).await?;
+            self.namespace_declarations.pop();
+            Ok(())
+        } else {
+            return Err(Error::NotInElement("".to_string()));
+        }
     }
 }
 
@@ -57,25 +279,43 @@ impl<W: AsyncWrite, E: Into<Element>> Sink<E> for Writer<W> {
     fn poll_ready(
         self: std::pin::Pin<&mut Self>,
         cx: &mut std::task::Context<'_>,
-    ) -> std::task::Poll<Result<(), Self::Error>> {
+    ) -> std::task::Poll<Result<()>> {
         todo!()
     }
 
-    fn start_send(self: std::pin::Pin<&mut Self>, item: E) -> Result<(), Self::Error> {
+    fn start_send(self: std::pin::Pin<&mut Self>, item: E) -> Result<()> {
         todo!()
     }
 
     fn poll_flush(
         self: std::pin::Pin<&mut Self>,
         cx: &mut std::task::Context<'_>,
-    ) -> std::task::Poll<Result<(), Self::Error>> {
+    ) -> std::task::Poll<Result<()>> {
         todo!()
     }
 
     fn poll_close(
         self: std::pin::Pin<&mut Self>,
         cx: &mut std::task::Context<'_>,
-    ) -> std::task::Poll<Result<(), Self::Error>> {
+    ) -> std::task::Poll<Result<()>> {
         todo!()
     }
 }
+
+#[cfg(test)]
+mod test {
+    use crate::{
+        reader::{test::*, Reader},
+        writer::Writer,
+    };
+
+    #[tokio::test]
+    async fn test_element_write() {
+        let mock = MockAsyncReader::new(TEST_DOC);
+        let mut reader = Reader::new(mock);
+        let element = reader.read_element().await.unwrap();
+        let stdout = tokio::io::stdout();
+        let mut writer = Writer::new(stdout);
+        writer.write_element(&element).await.unwrap();
+    }
+}

src/xml/mod.rs

@@ -1,4 +1,6 @@
-use std::{char, ops::Deref};
+use std::{char, convert::Infallible, ops::Deref, str::FromStr};
+
+use parsers_complete::Parser;
 
 use crate::error::Error;
 
@@ -15,7 +17,7 @@ pub enum NSAttName<'s> {
 
 /// [2]   	PrefixedAttName	   ::=   	'xmlns:' NCName
 #[derive(Clone, Debug)]
-pub struct PrefixedAttName<'s>(NCName<'s>);
+pub struct PrefixedAttName<'s>(pub NCName<'s>);
 
 impl<'s> Deref for PrefixedAttName<'s> {
     type Target = NCName<'s>;
@@ -228,14 +230,14 @@ pub enum EntityValue<'s> {
     SingleQuoted(Vec<EntityValueData<'s>>),
 }
 
-#[derive(Clone, Debug)]
+#[derive(Clone, Debug, PartialEq, Eq)]
 pub enum AttValueData<'s> {
     String(&'s str),
     Reference(Reference<'s>),
 }
 /// [10]   	AttValue	   ::=   	'"' ([^<&"] | Reference)* '"'
 /// 			|  "'" ([^<&'] | Reference)* "'"
-#[derive(Clone, Debug)]
+#[derive(Clone, Debug, PartialEq, Eq)]
 pub enum AttValue<'s> {
     DoubleQuoted(Vec<AttValueData<'s>>),
     SingleQuoted(Vec<AttValueData<'s>>),
@@ -259,6 +261,34 @@ impl<'s> AttValue<'s> {
     }
 }
 
+impl<'s> From<&'s str> for AttValue<'s> {
+    fn from(s: &'s str) -> AttValue<'s> {
+        let mut data = Vec::new();
+        for str in s.split_inclusive(|c| c == '<' || c == '"') {
+            if let Some(str) = str.strip_suffix('<') {
+                if !str.is_empty() {
+                    data.push(AttValueData::String(str))
+                }
+                data.push(AttValueData::Reference(Reference::EntityRef(EntityRef(
+                    Name::parse_full("lt").unwrap(),
+                ))))
+            } else if let Some(str) = str.strip_suffix('"') {
+                if !str.is_empty() {
+                    data.push(AttValueData::String(str))
+                }
+                data.push(AttValueData::Reference(Reference::EntityRef(EntityRef(
+                    Name::parse_full("quot").unwrap(),
+                ))))
+            } else {
+                if !str.is_empty() {
+                    data.push(AttValueData::String(str))
+                }
+            }
+        }
+        AttValue::DoubleQuoted(data)
+    }
+}
+
 /// [11]   	SystemLiteral	   ::=   	('"' [^"]* '"') | ("'" [^']* "'")
 #[derive(Debug)]
 pub enum SystemLiteral<'s> {
@@ -673,7 +703,7 @@ pub struct IgnoreSectContents<'s> {
 pub struct Ignore<'s>(&'s str);
 
 /// [66]   	CharRef	   ::=   	'&#' [0-9]+ ';' | '&#x' [0-9a-fA-F]+ ';'
-#[derive(Clone, Debug)]
+#[derive(Clone, Debug, PartialEq, Eq)]
 pub enum CharRef<'s> {
     Decimal(&'s str),
     Hexadecimal(&'s str),
@@ -706,7 +736,7 @@ impl<'s> CharRef<'s> {
 }
 
 /// [67]   	Reference	   ::=   	EntityRef | CharRef
-#[derive(Clone, Debug)]
+#[derive(Clone, Debug, PartialEq, Eq)]
 pub enum Reference<'s> {
     EntityRef(EntityRef<'s>),
     CharRef(CharRef<'s>),
@@ -729,8 +759,8 @@ impl<'s> Reference<'s> {
 }
 
 /// [68]   	EntityRef	   ::=   	'&' Name ';'
-#[derive(Clone, Debug)]
+#[derive(Clone, Debug, PartialEq, Eq)]
-pub struct EntityRef<'s>(Name<'s>);
+pub struct EntityRef<'s>(pub(crate) Name<'s>);
 
 impl<'s> Deref for EntityRef<'s> {
     type Target = Name<'s>;
@@ -835,3 +865,60 @@ pub struct NotationDecl<'s> {
 /// [83]   	PublicID	   ::=   	'PUBLIC' S PubidLiteral
 #[derive(Debug)]
 pub struct PublicID<'s>(PubidLiteral<'s>);
+
+#[cfg(test)]
+mod test {
+    use super::{AttValue, AttValueData, EntityRef, Name, Reference};
+
+    #[test]
+    fn att_value_from_str() {
+        assert_eq!(
+            AttValue::from("hsdaflaskdf<laksdf<abdsf"),
+            AttValue::DoubleQuoted(vec![
+                AttValueData::String("hsdaflaskdf"),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("lt")))),
+                AttValueData::String("laksdf"),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("lt")))),
+                AttValueData::String("abdsf"),
+            ])
+        );
+        assert_eq!(
+            AttValue::from("hsdaflaskdf<laksdf\"abdsf"),
+            AttValue::DoubleQuoted(vec![
+                AttValueData::String("hsdaflaskdf"),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("lt")))),
+                AttValueData::String("laksdf"),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("quot")))),
+                AttValueData::String("abdsf"),
+            ])
+        );
+        assert_eq!(
+            AttValue::from("hsdaflaskdf<laksdf\""),
+            AttValue::DoubleQuoted(vec![
+                AttValueData::String("hsdaflaskdf"),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("lt")))),
+                AttValueData::String("laksdf"),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("quot")))),
+            ])
+        );
+        assert_eq!(
+            AttValue::from("hsdaflaskdf\"<<laksdf\""),
+            AttValue::DoubleQuoted(vec![
+                AttValueData::String("hsdaflaskdf"),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("quot")))),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("lt")))),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("lt")))),
+                AttValueData::String("laksdf"),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("quot")))),
+            ])
+        );
+        assert_eq!(
+            AttValue::from("<<\""),
+            AttValue::DoubleQuoted(vec![
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("lt")))),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("lt")))),
+                AttValueData::Reference(Reference::EntityRef(EntityRef(Name("quot")))),
+            ])
+        );
+    }
+}