fix(cometbft): Fix deserialization from serde_json::Value (#112)

Port-of: cometbft/tendermint-rs#1475

Closes #111

Author: melekes

.changelog/unreleased/bug-fixes/111-fix-deserialization-from-serde_json_Value.md

@@ -0,0 +1,4 @@
+- [cometbft-proto] Intoduce `CowStr` deserializer.
+  ([\#111](https://github.com/cometbft/cometbft-rs/issues/111))
+- [cometbft] Fix deserialization from `serde_json::Value`.
+  ([\#111](https://github.com/cometbft/cometbft-rs/issues/111))

cometbft/src/hash.rs

@@ -7,6 +7,7 @@ use core::{
 };
 
 use bytes::Bytes;
+use cometbft_proto::serializers::cow_str::CowStr;
 use cometbft_proto::Protobuf;
 use serde::{de::Error as _, Deserialize, Deserializer, Serialize, Serializer};
 use subtle_encoding::{base64, Encoding, Hex};
@@ -166,12 +167,12 @@ impl FromStr for Hash {
 
 impl<'de> Deserialize<'de> for Hash {
     fn deserialize<D: Deserializer<'de>>(deserializer: D) -> Result<Self, D::Error> {
-        let hex = <&str>::deserialize(deserializer)?;
+        let hex = CowStr::deserialize(deserializer)?;
 
         if hex.is_empty() {
             Err(D::Error::custom("empty hash"))
         } else {
-            Ok(Self::from_str(hex).map_err(|e| D::Error::custom(format!("{e}")))?)
+            Ok(Self::from_str(&hex).map_err(|e| D::Error::custom(format!("{e}")))?)
         }
     }
 }
@@ -200,8 +201,8 @@ pub mod allow_empty {
     where
         D: Deserializer<'de>,
     {
-        let hex = <&str>::deserialize(deserializer)?;
-        Hash::from_str(hex).map_err(serde::de::Error::custom)
+        let hex = CowStr::deserialize(deserializer)?;
+        Hash::from_str(&hex).map_err(serde::de::Error::custom)
     }
 }
 
@@ -301,15 +302,22 @@ mod tests {
     use super::*;
 
     #[derive(Debug, serde::Deserialize)]
-    struct Test {
+    struct AppHashTest {
         #[serde(default)]
         #[serde(with = "crate::serializers::apphash")]
         pub app_hash: AppHash,
     }
 
+    #[derive(Debug, serde::Deserialize)]
+    struct HashTest {
+        hash: Hash,
+        #[serde(with = "super::allow_empty")]
+        empty_hash: Hash,
+    }
+
     #[test]
     fn apphash_decode_base64() {
-        let test = serde_json::from_str::<Test>(
+        let test = serde_json::from_str::<AppHashTest>(
             r#"{"app_hash":"MfX9f+bYoI8IioRb4YT/8/VhPvtNjgWFgTi4mmMSkBc="}"#,
         )
         .unwrap();
@@ -326,7 +334,7 @@ mod tests {
 
     #[test]
     fn apphash_decode_hex() {
-        let test = serde_json::from_str::<Test>(
+        let test = serde_json::from_str::<AppHashTest>(
             r#"{"app_hash":"31F5FD7FE6D8A08F088A845BE184FFF3F5613EFB4D8E05858138B89A63129017"}"#,
         )
         .unwrap();
@@ -340,4 +348,28 @@ mod tests {
             ]
         );
     }
+
+    #[test]
+    fn hash_decode_hex() {
+        let s = r#"{
+            "hash": "9F86D081884C7D659A2FEAA0C55AD015A3BF4F1B2B0B822CD15D6C15B0F00A08",
+            "empty_hash": ""
+        }"#;
+
+        let expected_hash = &[
+            0x9F, 0x86, 0xD0, 0x81, 0x88, 0x4C, 0x7D, 0x65, 0x9A, 0x2F, 0xEA, 0xA0, 0xC5, 0x5A,
+            0xD0, 0x15, 0xA3, 0xBF, 0x4F, 0x1B, 0x2B, 0x0B, 0x82, 0x2C, 0xD1, 0x5D, 0x6C, 0x15,
+            0xB0, 0xF0, 0x0A, 0x08,
+        ];
+
+        let test = serde_json::from_str::<HashTest>(s).unwrap();
+        assert_eq!(test.hash.as_ref(), expected_hash);
+        assert_eq!(test.empty_hash, Hash::None);
+
+        // Test issue 1474
+        let json_value = serde_json::from_str::<serde_json::Value>(s).unwrap();
+        let test = serde_json::from_value::<HashTest>(json_value).unwrap();
+        assert_eq!(test.hash.as_ref(), expected_hash);
+        assert_eq!(test.empty_hash, Hash::None);
+    }
 }

proto/src/serializers.rs

@@ -55,6 +55,7 @@
 
 pub mod allow_null;
 pub mod bytes;
+pub mod cow_str;
 pub mod evidence;
 pub mod from_str;
 pub mod from_str_allow_null;

proto/src/serializers/bytes.rs

@@ -6,13 +6,14 @@ pub mod hexstring {
     use subtle_encoding::hex;
 
     use crate::prelude::*;
+    use crate::serializers::cow_str::CowStr;
 
     /// Deserialize a hex-encoded string into `Vec<u8>`
     pub fn deserialize<'de, D>(deserializer: D) -> Result<Vec<u8>, D::Error>
     where
         D: Deserializer<'de>,
     {
-        let string = Option::<String>::deserialize(deserializer)?.unwrap_or_default();
+        let string = Option::<CowStr<'_>>::deserialize(deserializer)?.unwrap_or_default();
         hex::decode_upper(&string)
             .or_else(|_| hex::decode(&string))
             .map_err(serde::de::Error::custom)
@@ -36,14 +37,15 @@ pub mod base64string {
     use subtle_encoding::base64;
 
     use crate::prelude::*;
+    use crate::serializers::cow_str::CowStr;
 
     /// Deserialize base64string into `Vec<u8>`
     pub fn deserialize<'de, D, T>(deserializer: D) -> Result<T, D::Error>
     where
         D: Deserializer<'de>,
         Vec<u8>: Into<T>,
     {
-        let s = Option::<String>::deserialize(deserializer)?.unwrap_or_default();
+        let s = Option::<CowStr<'_>>::deserialize(deserializer)?.unwrap_or_default();
         let v = base64::decode(s).map_err(serde::de::Error::custom)?;
         Ok(v.into())
     }
@@ -53,7 +55,7 @@ pub mod base64string {
     where
         D: Deserializer<'de>,
     {
-        let s = Option::<String>::deserialize(deserializer)?.unwrap_or_default();
+        let s = Option::<CowStr<'_>>::deserialize(deserializer)?.unwrap_or_default();
         String::from_utf8(base64::decode(s).map_err(serde::de::Error::custom)?)
             .map_err(serde::de::Error::custom)
     }
@@ -76,13 +78,14 @@ pub mod vec_base64string {
     use subtle_encoding::base64;
 
     use crate::prelude::*;
+    use crate::serializers::cow_str::CowStr;
 
     /// Deserialize array into `Vec<Vec<u8>>`
     pub fn deserialize<'de, D>(deserializer: D) -> Result<Vec<Vec<u8>>, D::Error>
     where
         D: Deserializer<'de>,
     {
-        Option::<Vec<String>>::deserialize(deserializer)?
+        Option::<Vec<CowStr<'_>>>::deserialize(deserializer)?
             .unwrap_or_default()
             .into_iter()
             .map(|s| base64::decode(s).map_err(serde::de::Error::custom))
@@ -111,13 +114,14 @@ pub mod option_base64string {
     use subtle_encoding::base64;
 
     use crate::prelude::*;
+    use crate::serializers::cow_str::CowStr;
 
     /// Deserialize `Option<base64string>` into `Vec<u8>` or null
     pub fn deserialize<'de, D>(deserializer: D) -> Result<Vec<u8>, D::Error>
     where
         D: Deserializer<'de>,
     {
-        let s = Option::<String>::deserialize(deserializer)?.unwrap_or_default();
+        let s = Option::<CowStr<'_>>::deserialize(deserializer)?.unwrap_or_default();
         base64::decode(s).map_err(serde::de::Error::custom)
     }
 
@@ -138,14 +142,15 @@ pub mod string {
     use serde::{Deserialize, Deserializer, Serializer};
 
     use crate::prelude::*;
+    use crate::serializers::cow_str::CowStr;
 
     /// Deserialize string into `Vec<u8>`
     #[allow(dead_code)]
     pub fn deserialize<'de, D>(deserializer: D) -> Result<Vec<u8>, D::Error>
     where
         D: Deserializer<'de>,
     {
-        let string = Option::<String>::deserialize(deserializer)?.unwrap_or_default();
+        let string = Option::<CowStr<'_>>::deserialize(deserializer)?.unwrap_or_default();
         Ok(string.as_bytes().to_vec())
     }
 

proto/src/serializers/cow_str.rs

@@ -0,0 +1,174 @@
+//! Wrapper `Cow<'_, str>` for deserializing without allocation.
+//!
+//! This is a workaround for [serde's issue 1852](https://github.com/serde-rs/serde/issues/1852).
+
+use alloc::borrow::{Cow, ToOwned};
+use alloc::string::String;
+use core::fmt::{self, Debug, Display, Formatter};
+use core::ops::Deref;
+use serde::{de, Deserialize, Deserializer, Serialize, Serializer};
+
+/// Wrapper `Cow<'_, str>` for deserializing without allocation.
+#[derive(Default)]
+pub struct CowStr<'a>(Cow<'a, str>);
+
+impl<'a> CowStr<'a> {
+    /// Convert into `Cow<'a, str>`.
+    pub fn into_inner(self) -> Cow<'a, str> {
+        self.0
+    }
+}
+
+impl<'de> Deserialize<'de> for CowStr<'de> {
+    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
+    where
+        D: Deserializer<'de>,
+    {
+        struct Visitor;
+
+        impl<'de> serde::de::Visitor<'de> for Visitor {
+            type Value = CowStr<'de>;
+
+            fn expecting(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result {
+                formatter.write_str("a string")
+            }
+
+            fn visit_borrowed_str<E>(self, value: &'de str) -> Result<Self::Value, E>
+            where
+                E: de::Error,
+            {
+                Ok(CowStr(Cow::Borrowed(value)))
+            }
+
+            fn visit_str<E>(self, value: &str) -> Result<Self::Value, E>
+            where
+                E: de::Error,
+            {
+                Ok(CowStr(Cow::Owned(value.to_owned())))
+            }
+
+            fn visit_string<E>(self, value: String) -> Result<Self::Value, E>
+            where
+                E: de::Error,
+            {
+                Ok(CowStr(Cow::Owned(value)))
+            }
+        }
+
+        deserializer.deserialize_str(Visitor)
+    }
+}
+
+impl Serialize for CowStr<'_> {
+    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
+    where
+        S: Serializer,
+    {
+        serializer.serialize_str(&self.0)
+    }
+}
+
+impl Debug for CowStr<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        <&str as Debug>::fmt(&&*self.0, f)
+    }
+}
+
+impl Display for CowStr<'_> {
+    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
+        <&str as Display>::fmt(&&*self.0, f)
+    }
+}
+
+impl Deref for CowStr<'_> {
+    type Target = str;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
+impl AsRef<str> for CowStr<'_> {
+    fn as_ref(&self) -> &str {
+        &self.0
+    }
+}
+
+impl AsRef<[u8]> for CowStr<'_> {
+    fn as_ref(&self) -> &[u8] {
+        self.0.as_bytes()
+    }
+}
+
+impl<'a> From<CowStr<'a>> for Cow<'a, str> {
+    fn from(value: CowStr<'a>) -> Self {
+        value.0
+    }
+}
+
+impl<'a> From<Cow<'a, str>> for CowStr<'a> {
+    fn from(value: Cow<'a, str>) -> Self {
+        CowStr(value)
+    }
+}
+
+/// Serialize `Cow<'_, str>`.
+#[allow(clippy::ptr_arg)]
+pub fn serialize<S>(value: &Cow<'_, str>, serializer: S) -> Result<S::Ok, S::Error>
+where
+    S: Serializer,
+{
+    serializer.serialize_str(value)
+}
+
+/// Deserialize `Cow<'_, str>`.
+pub fn deserialize<'de, D>(deserializer: D) -> Result<Cow<'de, str>, D::Error>
+where
+    D: Deserializer<'de>,
+{
+    CowStr::deserialize(deserializer).map(|value| value.into_inner())
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn borrowed() {
+        struct Test(u32);
+
+        impl<'de> Deserialize<'de> for Test {
+            fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
+            where
+                D: Deserializer<'de>,
+            {
+                let s = CowStr::deserialize(deserializer)?;
+                assert!(matches!(s.0, Cow::Borrowed(_)));
+                Ok(Test(s.parse().unwrap()))
+            }
+        }
+
+        let v = serde_json::from_str::<Test>("\"2\"").unwrap();
+        assert_eq!(v.0, 2);
+    }
+
+    #[test]
+    fn owned() {
+        struct Test(u32);
+
+        impl<'de> Deserialize<'de> for Test {
+            fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
+            where
+                D: Deserializer<'de>,
+            {
+                let s = CowStr::deserialize(deserializer)?;
+                assert!(matches!(s.0, Cow::Owned(_)));
+                Ok(Test(s.parse().unwrap()))
+            }
+        }
+
+        let json_value = serde_json::from_str::<serde_json::Value>("\"2\"").unwrap();
+        let v = serde_json::from_value::<Test>(json_value).unwrap();
+        assert_eq!(v.0, 2);
+    }
+}

proto/src/serializers/from_str.rs

@@ -2,13 +2,13 @@
 //! and [`Display`] to convert from or into string. Note this can be used for
 //! all primitive data types.
 
-use alloc::borrow::Cow;
 use core::fmt::Display;
 use core::str::FromStr;
 
 use serde::{de::Error as _, Deserialize, Deserializer, Serialize, Serializer};
 
 use crate::prelude::*;
+use crate::serializers::cow_str::CowStr;
 
 /// Deserialize string into T
 pub fn deserialize<'de, D, T>(deserializer: D) -> Result<T, D::Error>
@@ -17,7 +17,7 @@ where
     T: FromStr,
     <T as FromStr>::Err: Display,
 {
-    <Cow<'_, str>>::deserialize(deserializer)?
+    CowStr::deserialize(deserializer)?
         .parse::<T>()
         .map_err(D::Error::custom)
 }