work on deserialization

Author: kralverde

pumpkin-nbt/src/compound.rs

@@ -1,7 +1,7 @@
 use crate::deserializer::ReadAdaptor;
+use crate::serializer::WriteAdaptor;
 use crate::tag::NbtTag;
 use crate::{get_nbt_string, Error, Nbt, END_ID};
-use bytes::{BufMut, Bytes, BytesMut};
 use std::io::{ErrorKind, Read, Write};
 use std::vec::IntoIter;
 
@@ -56,19 +56,16 @@ impl NbtCompound {
         Ok(compound)
     }
 
-    pub fn serialize_content(&self) -> Bytes {
-        let mut bytes = BytesMut::new();
+    pub fn serialize_content<W>(&self, w: &mut WriteAdaptor<W>) -> Result<(), Error>
+    where
+        W: Write,
+    {
         for (name, tag) in &self.child_tags {
-            bytes.put_u8(tag.get_type_id());
-            bytes.put(NbtTag::String(name.clone()).serialize_data());
-            bytes.put(tag.serialize_data());
+            w.write_u8_be(tag.get_type_id())?;
+            NbtTag::String(name.clone()).serialize_data(w)?;
+            tag.serialize_data(w)?;
         }
-        bytes.put_u8(END_ID);
-        bytes.freeze()
-    }
-
-    pub fn serialize_content_to_writer<W: Write>(&self, mut writer: W) -> std::io::Result<()> {
-        writer.write_all(&self.serialize_content())?;
+        w.write_u8_be(END_ID)?;
         Ok(())
     }
 
@@ -152,19 +149,19 @@ impl NbtCompound {
         self.get(name).and_then(|tag| tag.extract_string())
     }
 
-    pub fn get_list(&self, name: &str) -> Option<&Vec<NbtTag>> {
+    pub fn get_list(&self, name: &str) -> Option<&[NbtTag]> {
         self.get(name).and_then(|tag| tag.extract_list())
     }
 
     pub fn get_compound(&self, name: &str) -> Option<&NbtCompound> {
         self.get(name).and_then(|tag| tag.extract_compound())
     }
 
-    pub fn get_int_array(&self, name: &str) -> Option<&Vec<i32>> {
+    pub fn get_int_array(&self, name: &str) -> Option<&[i32]> {
         self.get(name).and_then(|tag| tag.extract_int_array())
     }
 
-    pub fn get_long_array(&self, name: &str) -> Option<&Vec<i64>> {
+    pub fn get_long_array(&self, name: &str) -> Option<&[i64]> {
         self.get(name).and_then(|tag| tag.extract_long_array())
     }
 }

pumpkin-nbt/src/deserializer.rs

@@ -99,7 +99,7 @@ impl<R: Read> ReadAdaptor<R> {
         Ok(f64::from_be_bytes(buf))
     }
 
-    pub fn read_to_bytes(&mut self, count: usize) -> Result<Bytes> {
+    pub fn read_boxed_slice(&mut self, count: usize) -> Result<Box<[u8]>> {
         let mut buf = vec![0u8; count];
         self.reader
             .read_exact(&mut buf)
@@ -113,6 +113,7 @@ impl<R: Read> ReadAdaptor<R> {
 pub struct Deserializer<R: Read> {
     input: ReadAdaptor<R>,
     tag_to_deserialize: Option<u8>,
+    in_list: bool,
     is_named: bool,
 }
 
@@ -121,6 +122,7 @@ impl<R: Read> Deserializer<R> {
         Deserializer {
             input: ReadAdaptor { reader: input },
             tag_to_deserialize: None,
+            in_list: false,
             is_named,
         }
     }
@@ -148,7 +150,7 @@ impl<'de, R: Read> de::Deserializer<'de> for &mut Deserializer<R> {
     type Error = Error;
 
     forward_to_deserialize_any! {
-        u8 u16 u32 u64 i8 i16 i32 i64 f32 f64 char str string unit unit_struct seq tuple tuple_struct
+        i8 i16 i32 i64 f32 f64 char str string unit unit_struct seq tuple tuple_struct
         ignored_any bytes enum newtype_struct byte_buf option
     }
 
@@ -195,6 +197,51 @@ impl<'de, R: Read> de::Deserializer<'de> for &mut Deserializer<R> {
         result
     }
 
+    fn deserialize_u8<V>(self, visitor: V) -> Result<V::Value>
+    where
+        V: Visitor<'de>,
+    {
+        if self.in_list {
+            let value = self.input.get_u8_be()?;
+            visitor.visit_u8::<Error>(value)
+        } else {
+            panic!("{:?}", self.tag_to_deserialize);
+
+            /*
+            Err(Error::UnsupportedType(
+                "u8; NBT only supports signed values".to_string(),
+            ))
+            */
+        }
+    }
+
+    fn deserialize_u16<V>(self, _visitor: V) -> Result<V::Value>
+    where
+        V: Visitor<'de>,
+    {
+        Err(Error::UnsupportedType(
+            "u16; NBT only supports signed values".to_string(),
+        ))
+    }
+
+    fn deserialize_u32<V>(self, _visitor: V) -> Result<V::Value>
+    where
+        V: Visitor<'de>,
+    {
+        Err(Error::UnsupportedType(
+            "u32; NBT only supports signed values".to_string(),
+        ))
+    }
+
+    fn deserialize_u64<V>(self, _visitor: V) -> Result<V::Value>
+    where
+        V: Visitor<'de>,
+    {
+        Err(Error::UnsupportedType(
+            "u64; NBT only supports signed values".to_string(),
+        ))
+    }
+
     fn deserialize_bool<V>(self, visitor: V) -> Result<V::Value>
     where
         V: Visitor<'de>,
@@ -304,6 +351,10 @@ impl<'de, R: Read> SeqAccess<'de> for ListAccess<'_, R> {
 
         self.remaining_values -= 1;
         self.de.tag_to_deserialize = Some(self.list_type);
-        seed.deserialize(&mut *self.de).map(Some)
+        self.de.in_list = true;
+        let result = seed.deserialize(&mut *self.de).map(Some);
+        self.de.in_list = false;
+        result
     }
 }
+

pumpkin-nbt/src/lib.rs

@@ -4,11 +4,12 @@ use std::{
     ops::Deref,
 };
 
-use bytes::{BufMut, Bytes, BytesMut};
+use bytes::Bytes;
 use compound::NbtCompound;
 use deserializer::ReadAdaptor;
 use serde::{de, ser};
 use serde::{Deserialize, Deserializer};
+use serializer::WriteAdaptor;
 use tag::NbtTag;
 use thiserror::Error;
 
@@ -49,6 +50,8 @@ pub enum Error {
     Incomplete(io::Error),
     #[error("Negative list length {0}")]
     NegativeLength(i32),
+    #[error("Length too large {0}")]
+    LargeLength(usize),
 }
 
 impl ser::Error for Error {
@@ -111,11 +114,15 @@ impl Nbt {
     }
 
     pub fn write(&self) -> Bytes {
-        let mut bytes = BytesMut::new();
-        bytes.put_u8(COMPOUND_ID);
-        bytes.put(NbtTag::String(self.name.to_string()).serialize_data());
-        bytes.put(self.root_tag.serialize_content());
-        bytes.freeze()
+        let mut bytes = Vec::new();
+        let mut writer = WriteAdaptor::new(&mut bytes);
+        writer.write_u8_be(COMPOUND_ID).unwrap();
+        NbtTag::String(self.name.to_string())
+            .serialize_data(&mut writer)
+            .unwrap();
+        self.root_tag.serialize_content(&mut writer).unwrap();
+
+        bytes.into()
     }
 
     pub fn write_to_writer<W: Write>(&self, mut writer: W) -> Result<(), io::Error> {
@@ -125,10 +132,13 @@ impl Nbt {
 
     /// Writes NBT tag, without name of root compound.
     pub fn write_unnamed(&self) -> Bytes {
-        let mut bytes = BytesMut::new();
-        bytes.put_u8(COMPOUND_ID);
-        bytes.put(self.root_tag.serialize_content());
-        bytes.freeze()
+        let mut bytes = Vec::new();
+        let mut writer = WriteAdaptor::new(&mut bytes);
+
+        writer.write_u8_be(COMPOUND_ID).unwrap();
+        self.root_tag.serialize_content(&mut writer).unwrap();
+
+        bytes.into()
     }
 
     pub fn write_unnamed_to_writer<W: Write>(&self, mut writer: W) -> Result<(), io::Error> {
@@ -169,7 +179,7 @@ impl AsMut<NbtCompound> for Nbt {
 
 pub fn get_nbt_string<R: Read>(bytes: &mut ReadAdaptor<R>) -> Result<String, Error> {
     let len = bytes.get_u16_be()? as usize;
-    let string_bytes = bytes.read_to_bytes(len)?;
+    let string_bytes = bytes.read_boxed_slice(len)?;
     let string = cesu8::from_java_cesu8(&string_bytes).map_err(|_| Error::Cesu8DecodingError)?;
     Ok(string.to_string())
 }
@@ -204,6 +214,7 @@ impl_array!(BytesArray, "byte");
 
 #[cfg(test)]
 mod test {
+    use std::io::Read;
     use std::sync::LazyLock;
 
     use flate2::read::GzDecoder;
@@ -212,6 +223,7 @@ mod test {
     use pumpkin_world::world_info::{DataPacks, LevelData, WorldGenSettings, WorldVersion};
 
     use crate::deserializer::from_bytes;
+    use crate::serializer::to_bytes;
     use crate::BytesArray;
     use crate::IntArray;
     use crate::LongArray;
@@ -237,8 +249,10 @@ mod test {
             float: 1.00,
             string: "Hello test".to_string(),
         };
-        let bytes = to_bytes_unnamed(&test).unwrap();
-        let recreated_struct: Test = from_bytes_unnamed(&mut &bytes[..]).unwrap();
+
+        let mut bytes = Vec::new();
+        to_bytes_unnamed(&test, &mut bytes).unwrap();
+        let recreated_struct: Test = from_bytes_unnamed(&bytes[..]).unwrap();
 
         assert_eq!(test, recreated_struct);
     }
@@ -260,8 +274,10 @@ mod test {
             int_array: vec![13, 1321, 2],
             long_array: vec![1, 0, 200301, 1],
         };
-        let bytes = to_bytes_unnamed(&test).unwrap();
-        let recreated_struct: TestArray = from_bytes_unnamed(&mut &bytes[..]).unwrap();
+
+        let mut bytes = Vec::new();
+        to_bytes_unnamed(&test, &mut bytes).unwrap();
+        let recreated_struct: TestArray = from_bytes_unnamed(&bytes[..]).unwrap();
 
         assert_eq!(test, recreated_struct);
     }
@@ -316,13 +332,40 @@ mod test {
     });
 
     #[test]
-    fn test_serialize_deserialize_level_dat() {
+    fn test_deserialize_level_dat() {
         let raw_compressed_nbt = include_bytes!("../assets/level.dat");
         assert!(!raw_compressed_nbt.is_empty());
 
         let decoder = GzDecoder::new(&raw_compressed_nbt[..]);
-        let level_dat: LevelDat = from_bytes(decoder).unwrap();
+        let level_dat: LevelDat = from_bytes(decoder).expect("Failed to decode from file");
 
         assert_eq!(level_dat, *LEVEL_DAT);
     }
+
+    #[test]
+    fn test_serialize_level_dat() {
+        let raw_compressed_nbt = include_bytes!("../assets/level.dat");
+        assert!(!raw_compressed_nbt.is_empty());
+
+        let mut decoder = GzDecoder::new(&raw_compressed_nbt[..]);
+        let mut raw_bytes = Vec::new();
+        decoder.read_to_end(&mut raw_bytes).unwrap();
+
+        let mut serialized = Vec::new();
+        to_bytes(&*LEVEL_DAT, "".to_string(), &mut serialized).expect("Failed to encode to bytes");
+        raw_bytes
+            .iter()
+            .zip(serialized.iter())
+            .enumerate()
+            .for_each(|(index, (expected_byte, serialized_byte))| {
+                if expected_byte != serialized_byte {
+                    panic!("{} vs {} ({})", expected_byte, serialized_byte, index);
+                }
+            });
+
+        let level_dat_again: LevelDat =
+            from_bytes(&serialized[..]).expect("Failed to decode from bytes");
+
+        assert_eq!(level_dat_again, *LEVEL_DAT);
+    }
 }