parity with orjson (#47)

* update dependencies

* more parity

CHANGELOG.md

@@ -1,5 +1,12 @@
 # Changelog
 ## Next Version
+## 1.0.1 - 13/10/2021
+### Fixed
+- Decrement refcount for numpy `PyArrayInterface`. by [@ilj](https://github.com/ijl/orjson/commit/4c312a82f5215ff71eed5bd09d28fa004999299b).
+- Fix serialization of dataclass inheriting from `abc.ABC` and using `__slots__`. by [@ilj](https://github.com/ijl/orjson/commit/4c312a82f5215ff71eed5bd09d28fa004999299b)
+### Changed
+- Updated dependencies.
+- `find_str_kind` test for 4-byte before latin1. by [@ilj](https://github.com/ijl/orjson/commit/05860e1a2ea3e8f90823d6a59e5fc9929a8692b5)
 ## 1.0.0 - 31/8/2021
 ### Changed
 -  Aligned to orjson's flags and features of SIMD. Didn't include the stable compilation part as seems unnecessary.

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "ormsgpack"
-version = "1.0.0"
+version = "1.0.1"
 authors = ["Aviram Hassan <[email protected]>"]
 description = "Fast, correct Python msgpack library supporting dataclasses, datetimes, and numpy"
 edition = "2018"
@@ -44,11 +44,11 @@ encoding_rs = { version = "0.8", default_features = false }
 inlinable_string = { version = "0.1" }
 itoa = { version = "0.4", default_features = false }
 once_cell = { version = "1", default_features = false }
-pyo3 = { version = "^0.14.1", default_features = false, features = ["extension-module"]}
+pyo3 = { version = "^0.14.5", default_features = false, features = ["extension-module"]}
 ryu = { version = "1", default_features = false }
 serde = { version = "1", default_features = false }
 simdutf8 = { version = "0.1", default_features = false, features = ["std"] }
-smallvec = { version = "^1.6", default_features = false, features = ["union", "write"] }
+smallvec = { version = "^1.7", default_features = false, features = ["union", "write"] }
 rmp = { version = "^0.8.10"}
 rmp-serde = {version = "^0.15.4"}
 

src/serialize/numpy.rs

@@ -207,7 +207,8 @@ impl<'a> NumpyArray {
 impl Drop for NumpyArray {
     fn drop(&mut self) {
         if self.depth == 0 {
-            ffi!(Py_XDECREF(self.capsule as *mut pyo3::ffi::PyObject))
+            ffi!(Py_DECREF(self.array as *mut pyo3::ffi::PyObject));
+            ffi!(Py_DECREF(self.capsule as *mut pyo3::ffi::PyObject));
         }
     }
 }

src/serialize/serializer.rs

@@ -249,7 +249,10 @@ impl<'p> Serialize for PyObjectSerializer {
                     err!(RECURSION_LIMIT_REACHED)
                 }
                 let dict = ffi!(PyObject_GetAttr(self.ptr, DICT_STR));
-                if unlikely!(dict.is_null()) {
+                let ob_type = ob_type!(self.ptr);
+                if unlikely!(
+                    dict.is_null() || ffi!(PyDict_Contains((*ob_type).tp_dict, SLOTS_STR)) == 1
+                ) {
                     unsafe { pyo3::ffi::PyErr_Clear() };
                     DataclassFallbackSerializer::new(
                         self.ptr,

src/typeref.rs

@@ -52,6 +52,7 @@ pub static mut EMPTY_UNICODE: *mut PyObject = 0 as *mut PyObject;
 pub static mut DST_STR: *mut PyObject = 0 as *mut PyObject;
 pub static mut DICT_STR: *mut PyObject = 0 as *mut PyObject;
 pub static mut DATACLASS_FIELDS_STR: *mut PyObject = 0 as *mut PyObject;
+pub static mut SLOTS_STR: *mut PyObject = 0 as *mut PyObject;
 pub static mut PYDANTIC_FIELDS_STR: *mut PyObject = 0 as *mut PyObject;
 pub static mut FIELD_TYPE_STR: *mut PyObject = 0 as *mut PyObject;
 pub static mut ARRAY_STRUCT_STR: *mut PyObject = 0 as *mut PyObject;
@@ -120,6 +121,7 @@ pub fn init_typerefs() {
         DICT_STR = PyUnicode_InternFromString("__dict__\0".as_ptr() as *const c_char);
         DATACLASS_FIELDS_STR =
             PyUnicode_InternFromString("__dataclass_fields__\0".as_ptr() as *const c_char);
+        SLOTS_STR = PyUnicode_InternFromString("__slots__\0".as_ptr() as *const c_char);
         PYDANTIC_FIELDS_STR = PyUnicode_InternFromString("__fields__\0".as_ptr() as *const c_char);
         FIELD_TYPE_STR = PyUnicode_InternFromString("_field_type\0".as_ptr() as *const c_char);
         ARRAY_STRUCT_STR =

src/unicode.rs

@@ -51,11 +51,11 @@ enum PyUnicodeKind {
 fn find_str_kind(buf: &str, num_chars: usize) -> PyUnicodeKind {
     if buf.len() == num_chars {
         PyUnicodeKind::Ascii
+    } else if is_four_byte(buf) {
+        PyUnicodeKind::FourByte
     } else if encoding_rs::mem::is_str_latin1(buf) {
         // fails fast, no obvious effect on CJK
         PyUnicodeKind::OneByte
-    } else if is_four_byte(buf) {
-        PyUnicodeKind::FourByte
     } else {
         PyUnicodeKind::TwoByte
     }

tests/test_dataclass.py

@@ -1,5 +1,5 @@
 # SPDX-License-Identifier: (Apache-2.0 OR MIT)
-
+import abc
 import uuid
 from dataclasses import InitVar, asdict, dataclass, field
 from enum import Enum
@@ -96,6 +96,22 @@ def __post_init__(self, a: str, b: str):
         self._other = 1
         self.ab = f"{a} {b}"
 
+class AbstractBase(abc.ABC):
+    @abc.abstractmethod
+    def key(self):
+        raise NotImplementedError
+
+
+@dataclass(frozen=True)
+class ConcreteAbc(AbstractBase):
+
+    __slots__ = ("attr",)
+
+    attr: float
+
+    def key(self):
+        return "dkjf"
+
 
 def test_dataclass():
     """
@@ -260,3 +276,8 @@ def default(obj):
     assert ormsgpack.packb(
         obj, option=ormsgpack.OPT_PASSTHROUGH_DATACLASS, default=default
     ) == msgpack.packb({"name": "a", "number": 1})
+
+
+def test_dataclass_abc():
+    obj = ConcreteAbc(1.0)
+    assert ormsgpack.packb(obj) == msgpack.packb({"attr": 1.0})