Add wrapper for PyASCIIObject state bitfield accesses based on bindgen

Author: decathorpe

newsfragments/3015.added.md

@@ -0,0 +1 @@
+Support `PyASCIIObject` / `PyUnicode` and associated methods on big-endian architectures.

pyo3-ffi/src/cpython/unicodeobject.rs

@@ -30,6 +30,184 @@ use std::os::raw::{c_char, c_int, c_uint, c_void};
 // skipped Py_UNICODE_HIGH_SURROGATE
 // skipped Py_UNICODE_LOW_SURROGATE
 
+// generated by bindgen v0.63.0 (with small adaptations)
+#[repr(C)]
+struct BitfieldUnit<Storage> {
+    storage: Storage,
+}
+
+impl<Storage> BitfieldUnit<Storage> {
+    #[inline]
+    pub const fn new(storage: Storage) -> Self {
+        Self { storage }
+    }
+}
+
+impl<Storage> BitfieldUnit<Storage>
+where
+    Storage: AsRef<[u8]> + AsMut<[u8]>,
+{
+    #[inline]
+    fn get_bit(&self, index: usize) -> bool {
+        debug_assert!(index / 8 < self.storage.as_ref().len());
+        let byte_index = index / 8;
+        let byte = self.storage.as_ref()[byte_index];
+        let bit_index = if cfg!(target_endian = "big") {
+            7 - (index % 8)
+        } else {
+            index % 8
+        };
+        let mask = 1 << bit_index;
+        byte & mask == mask
+    }
+
+    #[inline]
+    fn set_bit(&mut self, index: usize, val: bool) {
+        debug_assert!(index / 8 < self.storage.as_ref().len());
+        let byte_index = index / 8;
+        let byte = &mut self.storage.as_mut()[byte_index];
+        let bit_index = if cfg!(target_endian = "big") {
+            7 - (index % 8)
+        } else {
+            index % 8
+        };
+        let mask = 1 << bit_index;
+        if val {
+            *byte |= mask;
+        } else {
+            *byte &= !mask;
+        }
+    }
+
+    #[inline]
+    fn get(&self, bit_offset: usize, bit_width: u8) -> u64 {
+        debug_assert!(bit_width <= 64);
+        debug_assert!(bit_offset / 8 < self.storage.as_ref().len());
+        debug_assert!((bit_offset + (bit_width as usize)) / 8 <= self.storage.as_ref().len());
+        let mut val = 0;
+        for i in 0..(bit_width as usize) {
+            if self.get_bit(i + bit_offset) {
+                let index = if cfg!(target_endian = "big") {
+                    bit_width as usize - 1 - i
+                } else {
+                    i
+                };
+                val |= 1 << index;
+            }
+        }
+        val
+    }
+
+    #[inline]
+    fn set(&mut self, bit_offset: usize, bit_width: u8, val: u64) {
+        debug_assert!(bit_width <= 64);
+        debug_assert!(bit_offset / 8 < self.storage.as_ref().len());
+        debug_assert!((bit_offset + (bit_width as usize)) / 8 <= self.storage.as_ref().len());
+        for i in 0..(bit_width as usize) {
+            let mask = 1 << i;
+            let val_bit_is_set = val & mask == mask;
+            let index = if cfg!(target_endian = "big") {
+                bit_width as usize - 1 - i
+            } else {
+                i
+            };
+            self.set_bit(index + bit_offset, val_bit_is_set);
+        }
+    }
+}
+
+// generated by bindgen v0.63.0 (with small adaptations)
+// The same code is generated for Python 3.7, 3.8, 3.9, 3.10, and 3.11, but the "ready" field
+// has been removed from Python 3.12.
+
+/// Wrapper around the `PyASCIIObject.state` bitfield with getters and setters that work
+/// on most little- and big-endian architectures.
+///
+/// Memory layout of C bitfields is implementation defined, so these functions are still
+/// unsafe. Users must verify that they work as expected on the architectures they target.
+#[repr(C)]
+#[repr(align(4))]
+struct PyASCIIObjectState {
+    _bitfield_align: [u8; 0],
+    _bitfield: BitfieldUnit<[u8; 4usize]>,
+}
+
+// c_uint and u32 are not necessarily the same type on all targets / architectures
+#[allow(clippy::useless_transmute)]
+impl PyASCIIObjectState {
+    #[inline]
+    unsafe fn interned(&self) -> c_uint {
+        std::mem::transmute(self._bitfield.get(0usize, 2u8) as u32)
+    }
+
+    #[inline]
+    unsafe fn set_interned(&mut self, val: c_uint) {
+        let val: u32 = std::mem::transmute(val);
+        self._bitfield.set(0usize, 2u8, val as u64)
+    }
+
+    #[inline]
+    unsafe fn kind(&self) -> c_uint {
+        std::mem::transmute(self._bitfield.get(2usize, 3u8) as u32)
+    }
+
+    #[inline]
+    unsafe fn set_kind(&mut self, val: c_uint) {
+        let val: u32 = std::mem::transmute(val);
+        self._bitfield.set(2usize, 3u8, val as u64)
+    }
+
+    #[inline]
+    unsafe fn compact(&self) -> c_uint {
+        std::mem::transmute(self._bitfield.get(5usize, 1u8) as u32)
+    }
+
+    #[inline]
+    unsafe fn set_compact(&mut self, val: c_uint) {
+        let val: u32 = std::mem::transmute(val);
+        self._bitfield.set(5usize, 1u8, val as u64)
+    }
+
+    #[inline]
+    unsafe fn ascii(&self) -> c_uint {
+        std::mem::transmute(self._bitfield.get(6usize, 1u8) as u32)
+    }
+
+    #[inline]
+    unsafe fn set_ascii(&mut self, val: c_uint) {
+        let val: u32 = std::mem::transmute(val);
+        self._bitfield.set(6usize, 1u8, val as u64)
+    }
+
+    #[inline]
+    unsafe fn ready(&self) -> c_uint {
+        std::mem::transmute(self._bitfield.get(7usize, 1u8) as u32)
+    }
+
+    #[inline]
+    unsafe fn set_ready(&mut self, val: c_uint) {
+        let val: u32 = std::mem::transmute(val);
+        self._bitfield.set(7usize, 1u8, val as u64)
+    }
+}
+
+impl From<u32> for PyASCIIObjectState {
+    #[inline]
+    fn from(value: u32) -> Self {
+        PyASCIIObjectState {
+            _bitfield_align: [],
+            _bitfield: BitfieldUnit::new(value.to_ne_bytes()),
+        }
+    }
+}
+
+impl From<PyASCIIObjectState> for u32 {
+    #[inline]
+    fn from(value: PyASCIIObjectState) -> Self {
+        u32::from_ne_bytes(value._bitfield.storage)
+    }
+}
+
 #[repr(C)]
 pub struct PyASCIIObject {
     pub ob_base: PyObject,
@@ -52,34 +230,98 @@ pub struct PyASCIIObject {
 }
 
 /// Interacting with the bitfield is not actually well-defined, so we mark these APIs unsafe.
-///
-/// In addition, they are disabled on big-endian architectures to restrict this to most "common"
-/// platforms, which are at least tested on CI and appear to be sound.
-#[cfg(target_endian = "little")]
 impl PyASCIIObject {
+    /// Get the `interned` field of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Returns one of: [`SSTATE_NOT_INTERNED`], [`SSTATE_INTERNED_MORTAL`], [`SSTATE_INTERNED_IMMORTAL`]
     #[inline]
     pub unsafe fn interned(&self) -> c_uint {
-        self.state & 3
+        PyASCIIObjectState::from(self.state).interned()
     }
 
+    /// Set the `interned` field of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Calling this function with an argument that is not [`SSTATE_NOT_INTERNED`],
+    /// [`SSTATE_INTERNED_MORTAL`], or [`SSTATE_INTERNED_IMMORTAL`] is invalid.
+    #[inline]
+    pub unsafe fn set_interned(&mut self, val: c_uint) {
+        let mut state = PyASCIIObjectState::from(self.state);
+        state.set_interned(val);
+        self.state = u32::from(state);
+    }
+
+    /// Get the `kind` field of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Returns one of: [`PyUnicode_WCHAR_KIND`], [`PyUnicode_1BYTE_KIND`], [`PyUnicode_2BYTE_KIND`],
+    /// [`PyUnicode_4BYTE_KIND`]
     #[inline]
     pub unsafe fn kind(&self) -> c_uint {
-        (self.state >> 2) & 7
+        PyASCIIObjectState::from(self.state).kind()
     }
 
+    /// Set the `kind` field of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Calling this function with an argument that is not [`PyUnicode_WCHAR_KIND`], [`PyUnicode_1BYTE_KIND`],
+    /// [`PyUnicode_2BYTE_KIND`], or [`PyUnicode_4BYTE_KIND`] is invalid.
+    #[inline]
+    pub unsafe fn set_kind(&mut self, val: c_uint) {
+        let mut state = PyASCIIObjectState::from(self.state);
+        state.set_kind(val);
+        self.state = u32::from(state);
+    }
+
+    /// Get the `compact` field of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Returns either `0` or `1`.
     #[inline]
     pub unsafe fn compact(&self) -> c_uint {
-        (self.state >> 5) & 1
+        PyASCIIObjectState::from(self.state).compact()
+    }
+
+    /// Set the `compact` flag of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Calling this function with an argument that is neither `0` nor `1` is invalid.
+    #[inline]
+    pub unsafe fn set_compact(&mut self, val: c_uint) {
+        let mut state = PyASCIIObjectState::from(self.state);
+        state.set_compact(val);
+        self.state = u32::from(state);
     }
 
+    /// Get the `ascii` field of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Returns either `0` or `1`.
     #[inline]
     pub unsafe fn ascii(&self) -> c_uint {
-        (self.state >> 6) & 1
+        PyASCIIObjectState::from(self.state).ascii()
     }
 
+    /// Set the `ascii` flag of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Calling this function with an argument that is neither `0` nor `1` is invalid.
+    #[inline]
+    pub unsafe fn set_ascii(&mut self, val: c_uint) {
+        let mut state = PyASCIIObjectState::from(self.state);
+        state.set_ascii(val);
+        self.state = u32::from(state);
+    }
+
+    /// Get the `ready` field of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Returns either `0` or `1`.
     #[inline]
     pub unsafe fn ready(&self) -> c_uint {
-        (self.state >> 7) & 1
+        PyASCIIObjectState::from(self.state).ready()
+    }
+
+    /// Set the `ready` flag of the [`PyASCIIObject`] state bitfield.
+    ///
+    /// Calling this function with an argument that is neither `0` nor `1` is invalid.
+    #[inline]
+    pub unsafe fn set_ready(&mut self, val: c_uint) {
+        let mut state = PyASCIIObjectState::from(self.state);
+        state.set_ready(val);
+        self.state = u32::from(state);
     }
 }
 
@@ -120,7 +362,6 @@ pub const SSTATE_INTERNED_MORTAL: c_uint = 1;
 pub const SSTATE_INTERNED_IMMORTAL: c_uint = 2;
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_IS_ASCII(op: *mut PyObject) -> c_uint {
     debug_assert!(crate::PyUnicode_Check(op) != 0);
     debug_assert!(PyUnicode_IS_READY(op) != 0);
@@ -129,13 +370,11 @@ pub unsafe fn PyUnicode_IS_ASCII(op: *mut PyObject) -> c_uint {
 }
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_IS_COMPACT(op: *mut PyObject) -> c_uint {
     (*(op as *mut PyASCIIObject)).compact()
 }
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_IS_COMPACT_ASCII(op: *mut PyObject) -> c_uint {
     ((*(op as *mut PyASCIIObject)).ascii() != 0 && PyUnicode_IS_COMPACT(op) != 0).into()
 }
@@ -149,25 +388,21 @@ pub const PyUnicode_2BYTE_KIND: c_uint = 2;
 pub const PyUnicode_4BYTE_KIND: c_uint = 4;
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_1BYTE_DATA(op: *mut PyObject) -> *mut Py_UCS1 {
     PyUnicode_DATA(op) as *mut Py_UCS1
 }
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_2BYTE_DATA(op: *mut PyObject) -> *mut Py_UCS2 {
     PyUnicode_DATA(op) as *mut Py_UCS2
 }
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_4BYTE_DATA(op: *mut PyObject) -> *mut Py_UCS4 {
     PyUnicode_DATA(op) as *mut Py_UCS4
 }
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_KIND(op: *mut PyObject) -> c_uint {
     debug_assert!(crate::PyUnicode_Check(op) != 0);
     debug_assert!(PyUnicode_IS_READY(op) != 0);
@@ -176,7 +411,6 @@ pub unsafe fn PyUnicode_KIND(op: *mut PyObject) -> c_uint {
 }
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn _PyUnicode_COMPACT_DATA(op: *mut PyObject) -> *mut c_void {
     if PyUnicode_IS_ASCII(op) != 0 {
         (op as *mut PyASCIIObject).offset(1) as *mut c_void
@@ -186,15 +420,13 @@ pub unsafe fn _PyUnicode_COMPACT_DATA(op: *mut PyObject) -> *mut c_void {
 }
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn _PyUnicode_NONCOMPACT_DATA(op: *mut PyObject) -> *mut c_void {
     debug_assert!(!(*(op as *mut PyUnicodeObject)).data.any.is_null());
 
     (*(op as *mut PyUnicodeObject)).data.any
 }
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_DATA(op: *mut PyObject) -> *mut c_void {
     debug_assert!(crate::PyUnicode_Check(op) != 0);
 
@@ -210,7 +442,6 @@ pub unsafe fn PyUnicode_DATA(op: *mut PyObject) -> *mut c_void {
 // skipped PyUnicode_READ_CHAR
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_GET_LENGTH(op: *mut PyObject) -> Py_ssize_t {
     debug_assert!(crate::PyUnicode_Check(op) != 0);
     debug_assert!(PyUnicode_IS_READY(op) != 0);
@@ -219,15 +450,13 @@ pub unsafe fn PyUnicode_GET_LENGTH(op: *mut PyObject) -> Py_ssize_t {
 }
 
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_IS_READY(op: *mut PyObject) -> c_uint {
     (*(op as *mut PyASCIIObject)).ready()
 }
 
 #[cfg(not(Py_3_12))]
 #[cfg_attr(Py_3_10, deprecated(note = "Python 3.10"))]
 #[inline]
-#[cfg(target_endian = "little")]
 pub unsafe fn PyUnicode_READY(op: *mut PyObject) -> c_int {
     debug_assert!(crate::PyUnicode_Check(op) != 0);