Merge pull request #72 from borman/bytestrings

Improved support for byte strings

serde_macros/src/de.rs

@@ -686,6 +686,16 @@ fn deserialize_field_visitor(
                         {
                             $body
                         }
+
+                        fn visit_bytes<E>(&mut self, value: &[u8]) -> ::std::result::Result<__Field, E>
+                            where E: ::serde::de::Error,
+                        {
+                            // TODO: would be better to generate a byte string literal match
+                            match ::std::str::from_utf8(value) {
+                                Ok(s) => self.visit_str(s),
+                                _ => Err(::serde::de::Error::syntax_error()),
+                            }
+                        }
                     }
 
                     deserializer.visit(

src/bytes.rs

@@ -1,6 +1,8 @@
 //! Helper module to enable serializing bytes more efficiently
 
 use std::ops;
+use std::fmt;
+use std::ascii;
 
 use ser;
 use de;
@@ -8,10 +10,17 @@ use de;
 ///////////////////////////////////////////////////////////////////////////////
 
 /// `Bytes` wraps a `&[u8]` in order to serialize into a byte array.
+#[derive(Clone, Copy, Eq, Hash, PartialEq, PartialOrd, Ord)]
 pub struct Bytes<'a> {
     bytes: &'a [u8],
 }
 
+impl<'a> fmt::Debug for Bytes<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(f, "b\"{}\"", escape_bytestring(self.bytes))
+    }
+}
+
 impl<'a> From<&'a [u8]> for Bytes<'a> {
     fn from(bytes: &'a [u8]) -> Self {
         Bytes {
@@ -28,6 +37,12 @@ impl<'a> From<&'a Vec<u8>> for Bytes<'a> {
     }
 }
 
+impl<'a> Into<&'a [u8]> for Bytes<'a> {
+    fn into(self) -> &'a [u8] {
+        self.bytes
+    }
+}
+
 impl<'a> ops::Deref for Bytes<'a> {
     type Target = [u8];
 
@@ -46,7 +61,7 @@ impl<'a> ser::Serialize for Bytes<'a> {
 ///////////////////////////////////////////////////////////////////////////////
 
 /// `ByteBuf` wraps a `Vec<u8>` in order to hook into serialize and from deserialize a byte array.
-#[derive(Clone, Debug, Eq, Hash, PartialEq, PartialOrd, Ord)]
+#[derive(Clone, Eq, Hash, PartialEq, PartialOrd, Ord)]
 pub struct ByteBuf {
     bytes: Vec<u8>,
 }
@@ -65,10 +80,22 @@ impl ByteBuf {
     }
 }
 
-impl<T> From<T> for ByteBuf where T: Into<Vec<u8>> {
-    fn from(bytes: T) -> Self {
+impl fmt::Debug for ByteBuf {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(f, "b\"{}\"", escape_bytestring(self.bytes.as_ref()))
+    }
+}
+
+impl Into<Vec<u8>> for ByteBuf {
+    fn into(self) -> Vec<u8> {
+        self.bytes
+    }
+}
+
+impl From<Vec<u8>> for ByteBuf {
+    fn from(bytes: Vec<u8>) -> Self {
         ByteBuf {
-            bytes: bytes.into(),
+            bytes: bytes,
         }
     }
 }
@@ -172,3 +199,15 @@ impl de::Deserialize for ByteBuf {
         deserializer.visit_bytes(ByteBufVisitor)
     }
 }
+
+///////////////////////////////////////////////////////////////////////////////
+
+fn escape_bytestring(bytes: &[u8]) -> String {
+    let mut result = String::new();
+    for &b in bytes {
+        for esc in ascii::escape_default(b) {
+            result.push(esc as char);
+        }
+    }
+    result
+}

src/de/impls.rs

@@ -3,6 +3,7 @@ use std::hash::Hash;
 use std::marker::PhantomData;
 use std::path;
 use std::rc::Rc;
+use std::str;
 use std::sync::Arc;
 use std::str;
 
@@ -216,6 +217,24 @@ impl Visitor for StringVisitor {
     {
         Ok(v)
     }
+
+    fn visit_bytes<E>(&mut self, v: &[u8]) -> Result<String, E>
+        where E: Error,
+    {
+        match str::from_utf8(v) {
+            Ok(s) => Ok(s.to_string()),
+            Err(_) => Err(Error::syntax_error()),
+        }
+    }
+
+    fn visit_byte_buf<'a, E>(&mut self, v: Vec<u8>) -> Result<String, E>
+        where E: Error,
+    {
+        match String::from_utf8(v) {
+            Ok(s) => Ok(s),
+            Err(_) => Err(Error::syntax_error()),
+        }
+    }
 }
 
 impl Deserialize for String {

src/de/mod.rs

@@ -272,10 +272,10 @@ pub trait Visitor {
         Err(Error::syntax_error())
     }
 
-    fn visit_byte_buf<E>(&mut self, _v: Vec<u8>) -> Result<Self::Value, E>
+    fn visit_byte_buf<E>(&mut self, v: Vec<u8>) -> Result<Self::Value, E>
         where E: Error,
     {
-        Err(Error::syntax_error())
+        self.visit_bytes(&v)
     }
 }
 

src/de/value.rs

@@ -12,6 +12,7 @@ use std::hash::Hash;
 use std::vec;
 
 use de;
+use bytes;
 
 ///////////////////////////////////////////////////////////////////////////////
 
@@ -409,3 +410,56 @@ impl<K, V> ValueDeserializer for HashMap<K, V>
         MapDeserializer::new(self.into_iter(), len)
     }
 }
+
+///////////////////////////////////////////////////////////////////////////////
+
+impl<'a> ValueDeserializer for bytes::Bytes<'a>
+{
+    type Deserializer = BytesDeserializer<'a>;
+
+    fn into_deserializer(self) -> BytesDeserializer<'a> {
+        BytesDeserializer(Some(self.into()))
+    }
+}
+
+pub struct BytesDeserializer<'a> (Option<&'a [u8]>);
+
+impl<'a> de::Deserializer for BytesDeserializer<'a> {
+    type Error = Error;
+
+    fn visit<V>(&mut self, mut visitor: V) -> Result<V::Value, Error>
+        where V: de::Visitor,
+    {
+        match self.0.take() {
+            Some(bytes) => visitor.visit_bytes(bytes),
+            None => Err(de::Error::end_of_stream_error()),
+        }
+    }
+}
+
+
+///////////////////////////////////////////////////////////////////////////////
+
+impl ValueDeserializer for bytes::ByteBuf
+{
+    type Deserializer = ByteBufDeserializer;
+
+    fn into_deserializer(self) -> Self::Deserializer {
+        ByteBufDeserializer(Some(self.into()))
+    }
+}
+
+pub struct ByteBufDeserializer(Option<Vec<u8>>);
+
+impl de::Deserializer for ByteBufDeserializer {
+    type Error = Error;
+
+    fn visit<V>(&mut self, mut visitor: V) -> Result<V::Value, Error>
+        where V: de::Visitor,
+    {
+        match self.0.take() {
+            Some(bytes) => visitor.visit_byte_buf(bytes),
+            None => Err(de::Error::end_of_stream_error()),
+        }
+    }
+}