Add Pcd::to_sorted_key_values, and hash values with a counter

src/pcd.rs

@@ -6,7 +6,6 @@ use aes_gcm::{
     aead::{Aead, Payload},
     AeadCore, Aes256Gcm, KeyInit,
 };
-use log::debug;
 use rand::thread_rng;
 use serde::{Deserialize, Serialize};
 use serde_json::{Map, Value};
@@ -103,15 +102,29 @@ pub trait Pcd<'a>: Serialize + Deserialize<'a> {
         }
     }
 
-    fn hash_all_fields(&self, commited_in: OutPoint) -> Result<Map<String, Value>> {
+    fn to_sorted_key_values(&self) -> Result<Map<String, Value>> {
         let map = self.to_value_object()?;
 
+        let mut sorted_key_values: Vec<(String, Value)> = map.into_iter().map(|(key, value)| (key, value)).collect(); 
+        sorted_key_values.sort_by_key(|(key, _)| key.to_owned());
+
+        Ok(sorted_key_values.into_iter().collect())
+    }
+
+    /// Create a hashed commitments for all values in the pcd
+    /// We need the commited_in outpoint to prevent same data producing the same hashes between different processes or differents states of the same process
+    fn hash_all_fields(&self, commited_in: OutPoint) -> Result<Map<String, Value>> {
         let outpoint = serialize(&commited_in);
 
-        let mut field2hash = Map::with_capacity(map.len());
+        // To prevent fields with identical data having identical commitments, we order the map in alphabetical order and append a counter to the value
+        let sorted_key_values = self.to_sorted_key_values()?;
+
+        let mut field2hash = Map::with_capacity(sorted_key_values.len());
         // this could be optimised since there's a midstate we're reusing
-        for (field, value) in map {
-            let tagged_hash = AnkPcdHash::from_value_with_outpoint(&value, &outpoint);
+        for (i, (field, value)) in sorted_key_values.into_iter().enumerate() {
+            let mut value_bin = serde_json::to_string(&value)?.into_bytes();
+            value_bin.push(i.try_into()?);
+            let tagged_hash = AnkPcdHash::from_value_with_outpoint(&value_bin, &outpoint);
             field2hash.insert(field, Value::String(tagged_hash.to_string()));
         }
 
@@ -204,9 +217,9 @@ pub trait Pcd<'a>: Serialize + Deserialize<'a> {
         fields2keys: &Map<String, Value>,
         fields2plain: &mut Map<String, Value>,
     ) -> Result<()> {
-        let map = self.to_value_object()?;
+        let sorted_key_values = self.to_sorted_key_values()?;
 
-        for (field, encrypted_value) in map.iter() {
+        for (i, (field, encrypted_value)) in sorted_key_values.iter().enumerate() {
             if let Some(aes_key) = fields2keys.get(field) {
                 let key_buf = Vec::from_hex(&aes_key.to_string().trim_matches('\"'))?;
 
@@ -239,7 +252,9 @@ pub trait Pcd<'a>: Serialize + Deserialize<'a> {
                 fields2plain.insert(field.to_owned(), serde_json::from_str(&decrypted_value)?);
             } else if let Some(commitment) = fields2commit.get(field) { // We should always have a commitment
                 // We check if the hashed value is the commitment
-                let hashed_value = AnkPcdHash::from_value_with_outpoint(encrypted_value, &serialize(&commited_in));
+                let mut value_bin = encrypted_value.to_string().into_bytes();
+                value_bin.push(i.try_into()?);
+                let hashed_value = AnkPcdHash::from_value_with_outpoint(&value_bin, &serialize(&commited_in));
                 if commitment.as_str().unwrap() != &hashed_value.to_string() {
                     // The value is encrypted, and we don't have the key
                     // We put the commitment instead of the encrypted value
@@ -258,7 +273,7 @@ pub trait Pcd<'a>: Serialize + Deserialize<'a> {
 
         match value {
             Value::Object(map) => Ok(map),
-            _ => Err(Error::msg("self is not a valid json object"))
+            _ => Err(Error::msg("not a valid json object"))
         }
     }
 
@@ -544,6 +559,83 @@ mod tests {
         .unwrap()
     }
 
+    #[test]
+    fn test_sort_map() {
+        let pcd = json!({
+            "z": 1, 
+            "b": 2, 
+            "a": 3
+        });
+
+        let expected = json!({
+            "a": 3,
+            "b": 2, 
+            "z": 1
+        });
+
+        let sorted_map = pcd.to_sorted_key_values().unwrap();
+
+        assert_eq!(Value::Object(sorted_map), expected);
+    }
+
+    #[test]
+    fn test_sort_empty_map() {
+        let empty_map = json!({});
+        let expected = json!({});
+
+        let actual_sorted_map = empty_map.to_sorted_key_values().expect("Failed to sort keys");
+
+        assert_eq!(
+            Value::Object(actual_sorted_map),
+            expected,
+            "Sorting failed for an empty map"
+        );
+    }
+
+    #[test]
+    fn test_sort_already_sorted_map() {
+        let sorted_map = json!({
+            "a": 1,
+            "b": 2,
+            "c": 3
+        });
+
+        let expected = sorted_map.clone(); // Expected result is the same
+
+        let actual_sorted_map = sorted_map.to_sorted_key_values().expect("Failed to sort keys");
+
+        assert_eq!(
+            Value::Object(actual_sorted_map),
+            expected,
+            "Sorting failed for an already sorted map"
+        );
+    }
+
+
+    #[test]
+    fn test_sort_mixed_value_map() {
+        let mixed_map = json!({
+            "z": [1, 2, 3],
+            "b": { "nested": true },
+            "a": 42
+        });
+
+        let expected = json!({
+            "a": 42,
+            "b": { "nested": true },
+            "z": [1, 2, 3]
+        });
+
+        let actual_sorted_map = mixed_map.to_sorted_key_values().expect("Failed to sort keys");
+
+        assert_eq!(
+            Value::Object(actual_sorted_map),
+            expected,
+            "Sorting failed for a map with mixed value types"
+        );
+    }
+
+
     #[test]
     fn test_validation_rule_new() {
         // Valid input