add LEB128 types

androscalpel_serializer/src/core/leb.rs

@@ -0,0 +1,318 @@
+//! The implementation of serializable for LEB128 types.
+
+use std::io::Write;
+
+use crate::{Error, ReadSeek, Result, Serializable};
+
+/// Signed LEB128, variable-length
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct Sleb128(i32);
+
+/// Unsigned LEB128, variable-length
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct Uleb128(u32);
+
+/// Unsigned LEB128 plus 1, variable-length
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct Uleb128p1(u32);
+
+impl Sleb128 {
+    fn get_serialized_bytes(&self) -> Vec<u8> {
+        let Self(mut val) = self;
+        let mut bytes = vec![];
+        loop {
+            let byte = val as u8 & 0b0111_1111;
+            val /= 64;
+            if val == 0 || val == -1 {
+                bytes.push(byte);
+                return bytes;
+            } else {
+                val /= 2;
+            }
+            bytes.push(byte | 0b1000_0000);
+        }
+    }
+}
+
+impl Serializable for Sleb128 {
+    fn serialize(&self, output: &mut dyn Write) -> Result<()> {
+        output
+            .write_all(&self.get_serialized_bytes())
+            .map_err(|err| {
+                Error::SerializationError(format!(
+                    "Failed to write serialize Sleb128 to output: {err}"
+                ))
+            })?;
+        Ok(())
+    }
+
+    fn deserialize(input: &mut dyn ReadSeek) -> Result<Self> {
+        let mut shift = 0;
+        let mut uvalue: u64 = 0;
+        let mut buffer = [0u8; 1];
+        for _ in 0..5 {
+            input.read_exact(&mut buffer).map_err(|_| {
+                Error::InputTooSmall("Failed to read all bytes for Sleb128 from the input".into())
+            })?;
+            let byte = buffer[0];
+            let is_last_byte = byte & 0b1000_0000 == 0;
+
+            if is_last_byte {
+                uvalue += ((byte & 0b0011_1111) as u64) << shift;
+                let sign = ((byte & 0b0100_0000) as u64) << shift;
+                let value = uvalue as i64 - sign as i64;
+                let value: i32 = value.try_into().map_err(|err| Error::DeserializationError(
+                        format!("Failed to parse the Sleb128 value, the encoded value is supposed to fit in 32 bits: {err}")
+                ))?;
+                return Ok(Self(value));
+            } else {
+                uvalue |= ((byte & 0b0111_1111) as u64) << shift;
+                shift += 7;
+            }
+        }
+        Err(Error::DeserializationError(
+            "Malfored LEB128 encoding: the value cannot be longer than 5 bytes".into(),
+        ))
+    }
+
+    fn size(&self) -> usize {
+        self.get_serialized_bytes().len()
+    }
+}
+
+impl Uleb128 {
+    fn get_serialized_bytes(&self) -> Vec<u8> {
+        let Self(mut val) = self;
+        let mut bytes = vec![];
+        loop {
+            let byte = val as u8 & 0b0111_1111;
+            val /= 128;
+            if val == 0 {
+                bytes.push(byte);
+                return bytes;
+            }
+            bytes.push(byte | 0b1000_0000);
+        }
+    }
+}
+
+impl Serializable for Uleb128 {
+    fn serialize(&self, output: &mut dyn Write) -> Result<()> {
+        output
+            .write_all(&self.get_serialized_bytes())
+            .map_err(|err| {
+                Error::SerializationError(format!(
+                    "Failed to write serialize Uleb128 to output: {err}"
+                ))
+            })?;
+        Ok(())
+    }
+
+    fn deserialize(input: &mut dyn ReadSeek) -> Result<Self> {
+        let mut shift = 0;
+        let mut value: u32 = 0;
+        let mut buffer = [0u8; 1];
+        for _ in 0..5 {
+            input.read_exact(&mut buffer).map_err(|_| {
+                Error::InputTooSmall("Failed to read all bytes for Uleb128 from the input".into())
+            })?;
+            let byte = buffer[0];
+            let is_last_byte = byte & 0b1000_0000 == 0;
+
+            value += ((byte & 0b0111_1111) as u32) << shift;
+            if is_last_byte {
+                return Ok(Self(value));
+            }
+            shift += 7;
+        }
+        Err(Error::DeserializationError(
+            "Malfored LEB128 encoding: the value cannot be longer than 5 bytes".into(),
+        ))
+    }
+
+    fn size(&self) -> usize {
+        self.get_serialized_bytes().len()
+    }
+}
+
+impl Uleb128p1 {
+    pub const MINUS_ONE: Uleb128p1 = Uleb128p1(0xFFFFFFFF);
+}
+
+impl Serializable for Uleb128p1 {
+    fn serialize(&self, output: &mut dyn Write) -> Result<()> {
+        let Self(val) = self;
+        Uleb128(val.overflowing_add(1).0).serialize(output)
+    }
+
+    fn deserialize(input: &mut dyn ReadSeek) -> Result<Self> {
+        let Uleb128(val) = Uleb128::deserialize(input)?;
+        Ok(Self(val.overflowing_sub(1).0))
+    }
+
+    fn size(&self) -> usize {
+        let Self(val) = self;
+        Uleb128(val.overflowing_add(1).0).size()
+    }
+}
+
+impl From<i32> for Sleb128 {
+    fn from(item: i32) -> Self {
+        Self(item)
+    }
+}
+
+impl From<u32> for Uleb128 {
+    fn from(item: u32) -> Self {
+        Self(item)
+    }
+}
+
+impl From<u32> for Uleb128p1 {
+    fn from(item: u32) -> Self {
+        Self(item)
+    }
+}
+
+impl From<Sleb128> for i32 {
+    fn from(item: Sleb128) -> Self {
+        item.0
+    }
+}
+
+impl From<Uleb128> for u32 {
+    fn from(item: Uleb128) -> Self {
+        item.0
+    }
+}
+
+impl From<Uleb128p1> for u32 {
+    fn from(item: Uleb128p1) -> Self {
+        item.0
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use super::*;
+
+    #[test]
+    fn serialize_sleb128() {
+        assert_eq!(Sleb128(0).serialize_to_vec().unwrap(), vec![0x00u8]);
+        assert_eq!(Sleb128(1).serialize_to_vec().unwrap(), vec![0x01u8]);
+        assert_eq!(Sleb128(-1).serialize_to_vec().unwrap(), vec![0x7Fu8]);
+        assert_eq!(
+            Sleb128(-128).serialize_to_vec().unwrap(),
+            vec![0x80u8, 0x7F]
+        );
+    }
+
+    #[test]
+    fn deserialize_sleb128() {
+        assert_eq!(
+            Sleb128::deserialize_from_slice(&[0x00u8]).unwrap(),
+            Sleb128(0)
+        );
+        assert_eq!(
+            Sleb128::deserialize_from_slice(&[0x01u8]).unwrap(),
+            Sleb128(1)
+        );
+        assert_eq!(
+            Sleb128::deserialize_from_slice(&[0x7Fu8]).unwrap(),
+            Sleb128(-1)
+        );
+        assert_eq!(
+            Sleb128::deserialize_from_slice(&[0x80u8, 0x7Fu8]).unwrap(),
+            Sleb128(-128)
+        );
+    }
+
+    #[test]
+    fn size_sleb128() {
+        assert_eq!(Sleb128(0).size(), 1);
+        assert_eq!(Sleb128(1).size(), 1);
+        assert_eq!(Sleb128(-1).size(), 1);
+        assert_eq!(Sleb128(-128).size(), 2);
+    }
+
+    #[test]
+    fn serialize_uleb128() {
+        assert_eq!(Uleb128(0).serialize_to_vec().unwrap(), vec![0x00u8]);
+        assert_eq!(Uleb128(1).serialize_to_vec().unwrap(), vec![0x01u8]);
+        assert_eq!(Uleb128(127).serialize_to_vec().unwrap(), vec![0x7Fu8]);
+        assert_eq!(
+            Uleb128(16256).serialize_to_vec().unwrap(),
+            vec![0x80u8, 0x7F]
+        );
+    }
+
+    #[test]
+    fn deserialize_uleb128() {
+        assert_eq!(
+            Uleb128::deserialize_from_slice(&[0x00u8]).unwrap(),
+            Uleb128(0)
+        );
+        assert_eq!(
+            Uleb128::deserialize_from_slice(&[0x01u8]).unwrap(),
+            Uleb128(1)
+        );
+        assert_eq!(
+            Uleb128::deserialize_from_slice(&[0x7Fu8]).unwrap(),
+            Uleb128(127)
+        );
+        assert_eq!(
+            Uleb128::deserialize_from_slice(&[0x80u8, 0x7Fu8]).unwrap(),
+            Uleb128(16256)
+        );
+    }
+
+    #[test]
+    fn size_uleb128() {
+        assert_eq!(Uleb128(0).size(), 1);
+        assert_eq!(Uleb128(1).size(), 1);
+        assert_eq!(Uleb128(127).size(), 1);
+        assert_eq!(Uleb128(16256).size(), 2);
+    }
+
+    #[test]
+    fn serialize_uleb128p1() {
+        assert_eq!(
+            Uleb128p1::MINUS_ONE.serialize_to_vec().unwrap(),
+            vec![0x00u8]
+        );
+        assert_eq!(Uleb128p1(0).serialize_to_vec().unwrap(), vec![0x01u8]);
+        assert_eq!(Uleb128p1(126).serialize_to_vec().unwrap(), vec![0x7Fu8]);
+        assert_eq!(
+            Uleb128p1(16255).serialize_to_vec().unwrap(),
+            vec![0x80u8, 0x7F]
+        );
+    }
+
+    #[test]
+    fn deserialize_uleb128p1() {
+        assert_eq!(
+            Uleb128p1::deserialize_from_slice(&[0x00u8]).unwrap(),
+            Uleb128p1::MINUS_ONE
+        );
+        assert_eq!(
+            Uleb128p1::deserialize_from_slice(&[0x01u8]).unwrap(),
+            Uleb128p1(0)
+        );
+        assert_eq!(
+            Uleb128p1::deserialize_from_slice(&[0x7Fu8]).unwrap(),
+            Uleb128p1(126)
+        );
+        assert_eq!(
+            Uleb128p1::deserialize_from_slice(&[0x80u8, 0x7Fu8]).unwrap(),
+            Uleb128p1(16255)
+        );
+    }
+
+    #[test]
+    fn size_uleb128p1() {
+        assert_eq!(Uleb128p1::MINUS_ONE.size(), 1);
+        assert_eq!(Uleb128p1(0).size(), 1);
+        assert_eq!(Uleb128p1(126).size(), 1);
+        assert_eq!(Uleb128p1(16255).size(), 2);
+    }
+}

androscalpel_serializer/src/core/mod.rs

@@ -5,10 +5,14 @@ use std::io::{Cursor, Read, Seek, SeekFrom, Write};
 
 pub use androscalpel_serializer_derive::*;
 
+mod leb;
+pub use leb::*;
+
 #[derive(Debug, PartialEq, Eq)]
 pub enum Error {
     InputTooSmall(String), // TODO: find a better name
-    SerializeationError(String),
+    SerializationError(String),
+    DeserializationError(String),
 }
 
 pub type Result<T> = core::result::Result<T, Error>;
@@ -20,7 +24,8 @@ impl std::fmt::Display for Error {
     fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
         match self {
             Self::InputTooSmall(msg) => write!(f, "Error: {}", msg),
-            Self::SerializeationError(msg) => write!(f, "Error: {}", msg),
+            Self::SerializationError(msg) => write!(f, "Error: {}", msg),
+            Self::DeserializationError(msg) => write!(f, "Error: {}", msg),
         }
     }
 }
@@ -85,15 +90,13 @@ impl<D: Serializable, U: Serializable + Eq> SerializableUntil<D, U> for Vec<D> {
         let mut data = Self::new();
         loop {
             let pos = input.stream_position().map_err(|err| {
-                Error::SerializeationError(format!("Failled to get position in steam: {err}"))
+                Error::SerializationError(format!("Failled to get position in steam: {err}"))
             })?;
             match U::deserialize(input) {
                 Ok(val) if val == end_val => break Ok(data),
                 Err(err) => break Err(err),
                 Ok(_) => input.seek(SeekFrom::Start(pos)).map_err(|err| {
-                    Error::SerializeationError(format!(
+                    Error::SerializationError(format!("Failled to get to position in steam: {err}"))
-                        "Failled to get to position in steam: {err}"
-                    ))
                 })?,
             };
             data.push(D::deserialize(input)?);
@@ -108,7 +111,7 @@ impl<D: Serializable, U: Serializable + Eq> SerializableUntil<D, U> for Vec<D> {
 impl Serializable for u8 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize u8 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize u8 to output: {err}"))
         })?;
         Ok(())
     }
@@ -129,7 +132,7 @@ impl Serializable for u8 {
 impl Serializable for i8 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize i8 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize i8 to output: {err}"))
         })?;
         Ok(())
     }
@@ -150,7 +153,7 @@ impl Serializable for i8 {
 impl Serializable for u16 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize u16 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize u16 to output: {err}"))
         })?;
         Ok(())
     }
@@ -171,7 +174,7 @@ impl Serializable for u16 {
 impl Serializable for i16 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize i16 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize i16 to output: {err}"))
         })?;
         Ok(())
     }
@@ -192,7 +195,7 @@ impl Serializable for i16 {
 impl Serializable for u32 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize u32 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize u32 to output: {err}"))
         })?;
         Ok(())
     }
@@ -213,7 +216,7 @@ impl Serializable for u32 {
 impl Serializable for i32 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize i32 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize i32 to output: {err}"))
         })?;
         Ok(())
     }
@@ -234,7 +237,7 @@ impl Serializable for i32 {
 impl Serializable for u64 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize u64 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize u64 to output: {err}"))
         })?;
         Ok(())
     }
@@ -255,7 +258,7 @@ impl Serializable for u64 {
 impl Serializable for i64 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize i64 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize i64 to output: {err}"))
         })?;
         Ok(())
     }
@@ -276,7 +279,7 @@ impl Serializable for i64 {
 impl Serializable for u128 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize u128 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize u128 to output: {err}"))
         })?;
         Ok(())
     }
@@ -297,7 +300,7 @@ impl Serializable for u128 {
 impl Serializable for i128 {
     fn serialize(&self, output: &mut dyn Write) -> Result<()> {
         output.write_all(&self.to_be_bytes()).map_err(|err| {
-            Error::SerializeationError(format!("Failed to write serialize i128 to output: {err}"))
+            Error::SerializationError(format!("Failed to write serialize i128 to output: {err}"))
         })?;
         Ok(())
     }

androscalpel_serializer_derive/src/lib.rs

@@ -508,7 +508,7 @@ fn get_implem_deserialize(data: &Data, params: &ParamsStruct) -> TokenStream {
                 let prefix = <#prefix_ty as androscalpel_serializer::Serializable>::deserialize(input)?;
                 #(#recurse)*
 
-                Err(androscalpel_serializer::Error::SerializeationError(format!(
+                Err(androscalpel_serializer::Error::SerializationError(format!(
                     "Found prefix {:?} that did not match any variant.", prefix
                 )))
             }

flake.lock

@@ -0,0 +1,61 @@
+{
+  "nodes": {
+    "flake-utils": {
+      "inputs": {
+        "systems": "systems"
+      },
+      "locked": {
+        "lastModified": 1689068808,
+        "narHash": "sha256-6ixXo3wt24N/melDWjq70UuHQLxGV8jZvooRanIHXw0=",
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "rev": "919d646de7be200f3bf08cb76ae1f09402b6f9b4",
+        "type": "github"
+      },
+      "original": {
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "type": "github"
+      }
+    },
+    "nixpkgs": {
+      "locked": {
+        "lastModified": 1692447944,
+        "narHash": "sha256-fkJGNjEmTPvqBs215EQU4r9ivecV5Qge5cF/QDLVn3U=",
+        "owner": "NixOS",
+        "repo": "nixpkgs",
+        "rev": "d680ded26da5cf104dd2735a51e88d2d8f487b4d",
+        "type": "github"
+      },
+      "original": {
+        "owner": "NixOS",
+        "ref": "nixos-unstable",
+        "repo": "nixpkgs",
+        "type": "github"
+      }
+    },
+    "root": {
+      "inputs": {
+        "flake-utils": "flake-utils",
+        "nixpkgs": "nixpkgs"
+      }
+    },
+    "systems": {
+      "locked": {
+        "lastModified": 1681028828,
+        "narHash": "sha256-Vy1rq5AaRuLzOxct8nz4T6wlgyUR7zLU309k9mBC768=",
+        "owner": "nix-systems",
+        "repo": "default",
+        "rev": "da67096a3b9bf56a91d16901293e51ba5b49a27e",
+        "type": "github"
+      },
+      "original": {
+        "owner": "nix-systems",
+        "repo": "default",
+        "type": "github"
+      }
+    }
+  },
+  "root": "root",
+  "version": 7
+}

flake.nix

@@ -0,0 +1,31 @@
+{
+  description = "An android bytecode manipulation library";
+
+  inputs.flake-utils.url = "github:numtide/flake-utils";
+  inputs.nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
+
+  outputs = { self, nixpkgs, flake-utils }:
+    flake-utils.lib.eachDefaultSystem (system:
+      let
+        pkgs = nixpkgs.legacyPackages.${system};
+      in
+      {
+#        packages = {
+#          androscalpel = { };
+#          docker = pkgs.dockerTools.buildImage {
+#            name = "androscalpel";
+#            tag = "latest";
+#            copyToRoot = pkgs.buildEnv {
+#              name = "androscalpel_root_img";
+#              paths = [ self.packages.${system}.androscalpel ];
+#              pathsToLink = [ "/bin" ];
+#            };
+#          };
+#          default = self.packages.${system}.androscalpel;
+#        };
+
+        devShells.default = pkgs.mkShell {
+          packages = [ pkgs.rustc pkgs.cargo ];
+        };
+      });
+}