Add special casing for deserializing empty objects, lists and strings

Author: akx

src/deserialize/deserializer.rs

@@ -1,7 +1,7 @@
 // SPDX-License-Identifier: (Apache-2.0 OR MIT)
 
 use crate::deserialize::cache::*;
-use crate::deserialize::utf8::read_input_to_str;
+use crate::deserialize::utf8::{read_buf_to_str, read_input_to_buf};
 use crate::deserialize::DeserializeError;
 use crate::typeref::*;
 use crate::unicode::*;
@@ -14,7 +14,19 @@ use std::ptr::NonNull;
 pub fn deserialize(
     ptr: *mut pyo3::ffi::PyObject,
 ) -> Result<NonNull<pyo3::ffi::PyObject>, DeserializeError<'static>> {
-    let buffer_str = read_input_to_str(ptr)?;
+    let buffer = read_input_to_buf(ptr)?;
+    if unlikely!(buffer.len() == 2) {
+        if buffer == b"[]" {
+            return Ok(nonnull!(ffi!(PyList_New(0))));
+        } else if buffer == b"{}" {
+            return Ok(nonnull!(ffi!(PyDict_New())));
+        } else if buffer == b"\"\"" {
+            ffi!(Py_INCREF(EMPTY_UNICODE));
+            unsafe { return Ok(nonnull!(EMPTY_UNICODE)) }
+        }
+    }
+    let buffer_str = read_buf_to_str(buffer)?;
+
     let mut deserializer = serde_json::Deserializer::from_str(buffer_str);
     let seed = JsonValue {};
     match seed.deserialize(&mut deserializer) {

src/deserialize/utf8.rs

@@ -31,9 +31,9 @@ fn is_valid_utf8(buf: &[u8]) -> bool {
     std::str::from_utf8(buf).is_ok()
 }
 
-pub fn read_input_to_str(
+pub fn read_input_to_buf(
     ptr: *mut pyo3::ffi::PyObject,
-) -> Result<&'static str, DeserializeError<'static>> {
+) -> Result<&'static [u8], DeserializeError<'static>> {
     let obj_type_ptr = ob_type!(ptr);
     let contents: &[u8];
     if is_type!(obj_type_ptr, STR_TYPE) {
@@ -73,9 +73,13 @@ pub fn read_input_to_str(
             ));
         }
         contents = unsafe { std::slice::from_raw_parts(buffer, length) };
-        if !is_valid_utf8(contents) {
-            return Err(DeserializeError::new(Cow::Borrowed(INVALID_STR), 0, 0, ""));
-        }
+    }
+    Ok(contents)
+}
+
+pub fn read_buf_to_str(contents: &[u8]) -> Result<&str, DeserializeError> {
+    if !is_valid_utf8(contents) {
+        return Err(DeserializeError::new(Cow::Borrowed(INVALID_STR), 0, 0, ""));
     }
     Ok(unsafe { std::str::from_utf8_unchecked(contents) })
 }