gh-101907: Stop using _Py_OPCODE and _Py_OPARG macros (GH-101912)

* gh-101907: Removes use of non-standard C++ extension from Include/cpython/code.h

* Make cases_generator correct on Windows

Author: zooba

Include/cpython/code.h

@@ -19,21 +19,35 @@ extern "C" {
 typedef union {
     uint16_t cache;
     struct {
-        uint8_t opcode;
-        uint8_t oparg;
-    };
+        uint8_t code;
+        uint8_t arg;
+    } op;
 } _Py_CODEUNIT;
 
-#define _Py_OPCODE(word) ((word).opcode)
-#define _Py_OPARG(word) ((word).oparg)
+
+/* These macros only remain defined for compatibility. */
+#define _Py_OPCODE(word) ((word).op.code)
+#define _Py_OPARG(word) ((word).op.arg)
+
+static inline _Py_CODEUNIT
+_py_make_codeunit(uint8_t opcode, uint8_t oparg)
+{
+    // No designated initialisers because of C++ compat
+    _Py_CODEUNIT word;
+    word.op.code = opcode;
+    word.op.arg = oparg;
+    return word;
+}
 
 static inline void
 _py_set_opcode(_Py_CODEUNIT *word, uint8_t opcode)
 {
-    word->opcode = opcode;
+    word->op.code = opcode;
 }
 
-#define _Py_SET_OPCODE(word, opcode) _py_set_opocde(&(word), opcode)
+#define _Py_MAKE_CODEUNIT(opcode, oparg) _py_make_codeunit((opcode), (oparg))
+#define _Py_SET_OPCODE(word, opcode) _py_set_opcode(&(word), (opcode))
+
 
 typedef struct {
     PyObject *_co_code;

Misc/NEWS.d/next/C API/2023-02-14-15-53-01.gh-issue-101907.HgF1N2.rst

@@ -0,0 +1 @@
+Removes use of non-standard C++ extension in public header files.

Objects/codeobject.c

@@ -413,7 +413,7 @@ init_code(PyCodeObject *co, struct _PyCodeConstructor *con)
            PyBytes_GET_SIZE(con->code));
     int entry_point = 0;
     while (entry_point < Py_SIZE(co) &&
-        _Py_OPCODE(_PyCode_CODE(co)[entry_point]) != RESUME) {
+        _PyCode_CODE(co)[entry_point].op.code != RESUME) {
         entry_point++;
     }
     co->_co_firsttraceable = entry_point;
@@ -1505,12 +1505,12 @@ deopt_code(_Py_CODEUNIT *instructions, Py_ssize_t len)
 {
     for (int i = 0; i < len; i++) {
         _Py_CODEUNIT instruction = instructions[i];
-        int opcode = _PyOpcode_Deopt[_Py_OPCODE(instruction)];
+        int opcode = _PyOpcode_Deopt[instruction.op.code];
         int caches = _PyOpcode_Caches[opcode];
-        instructions[i].opcode = opcode;
+        instructions[i].op.code = opcode;
         while (caches--) {
-            instructions[++i].opcode = CACHE;
-            instructions[i].oparg = 0;
+            instructions[++i].op.code = CACHE;
+            instructions[i].op.arg = 0;
         }
     }
 }
@@ -1763,13 +1763,13 @@ code_richcompare(PyObject *self, PyObject *other, int op)
     for (int i = 0; i < Py_SIZE(co); i++) {
         _Py_CODEUNIT co_instr = _PyCode_CODE(co)[i];
         _Py_CODEUNIT cp_instr = _PyCode_CODE(cp)[i];
-        co_instr.opcode = _PyOpcode_Deopt[_Py_OPCODE(co_instr)];
-        cp_instr.opcode =_PyOpcode_Deopt[_Py_OPCODE(cp_instr)];
+        co_instr.op.code = _PyOpcode_Deopt[co_instr.op.code];
+        cp_instr.op.code = _PyOpcode_Deopt[cp_instr.op.code];
         eq = co_instr.cache == cp_instr.cache;
         if (!eq) {
             goto unequal;
         }
-        i += _PyOpcode_Caches[_Py_OPCODE(co_instr)];
+        i += _PyOpcode_Caches[co_instr.op.code];
     }
 
     /* compare constants */
@@ -1848,9 +1848,9 @@ code_hash(PyCodeObject *co)
     SCRAMBLE_IN(co->co_firstlineno);
     SCRAMBLE_IN(Py_SIZE(co));
     for (int i = 0; i < Py_SIZE(co); i++) {
-        int deop = _PyOpcode_Deopt[_Py_OPCODE(_PyCode_CODE(co)[i])];
+        int deop = _PyOpcode_Deopt[_PyCode_CODE(co)[i].op.code];
         SCRAMBLE_IN(deop);
-        SCRAMBLE_IN(_Py_OPARG(_PyCode_CODE(co)[i]));
+        SCRAMBLE_IN(_PyCode_CODE(co)[i].op.arg);
         i += _PyOpcode_Caches[deop];
     }
     if ((Py_hash_t)uhash == -1) {

Objects/frameobject.c

@@ -111,13 +111,13 @@ static unsigned int
 get_arg(const _Py_CODEUNIT *codestr, Py_ssize_t i)
 {
     _Py_CODEUNIT word;
-    unsigned int oparg = _Py_OPARG(codestr[i]);
-    if (i >= 1 && _Py_OPCODE(word = codestr[i-1]) == EXTENDED_ARG) {
-        oparg |= _Py_OPARG(word) << 8;
-        if (i >= 2 && _Py_OPCODE(word = codestr[i-2]) == EXTENDED_ARG) {
-            oparg |= _Py_OPARG(word) << 16;
-            if (i >= 3 && _Py_OPCODE(word = codestr[i-3]) == EXTENDED_ARG) {
-                oparg |= _Py_OPARG(word) << 24;
+    unsigned int oparg = codestr[i].op.arg;
+    if (i >= 1 && (word = codestr[i-1]).op.code == EXTENDED_ARG) {
+        oparg |= word.op.arg << 8;
+        if (i >= 2 && (word = codestr[i-2]).op.code == EXTENDED_ARG) {
+            oparg |= word.op.arg << 16;
+            if (i >= 3 && (word = codestr[i-3]).op.code == EXTENDED_ARG) {
+                oparg |= word.op.arg << 24;
             }
         }
     }
@@ -304,7 +304,7 @@ mark_stacks(PyCodeObject *code_obj, int len)
             if (next_stack == UNINITIALIZED) {
                 continue;
             }
-            opcode = _Py_OPCODE(code[i]);
+            opcode = code[i].op.code;
             switch (opcode) {
                 case JUMP_IF_FALSE_OR_POP:
                 case JUMP_IF_TRUE_OR_POP:
@@ -610,7 +610,7 @@ _PyFrame_GetState(PyFrameObject *frame)
             if (_PyInterpreterFrame_LASTI(frame->f_frame) < 0) {
                 return FRAME_CREATED;
             }
-            switch (_Py_OPCODE(*frame->f_frame->prev_instr))
+            switch (frame->f_frame->prev_instr->op.code)
             {
                 case COPY_FREE_VARS:
                 case MAKE_CELL:
@@ -1092,8 +1092,8 @@ _PyFrame_OpAlreadyRan(_PyInterpreterFrame *frame, int opcode, int oparg)
     for (_Py_CODEUNIT *instruction = _PyCode_CODE(frame->f_code);
          instruction < frame->prev_instr; instruction++)
     {
-        int check_opcode = _PyOpcode_Deopt[_Py_OPCODE(*instruction)];
-        check_oparg |= _Py_OPARG(*instruction);
+        int check_opcode = _PyOpcode_Deopt[instruction->op.code];
+        check_oparg |= instruction->op.arg;
         if (check_opcode == opcode && check_oparg == oparg) {
             return 1;
         }
@@ -1117,7 +1117,7 @@ frame_init_get_vars(_PyInterpreterFrame *frame)
     // here:
     PyCodeObject *co = frame->f_code;
     int lasti = _PyInterpreterFrame_LASTI(frame);
-    if (!(lasti < 0 && _Py_OPCODE(_PyCode_CODE(co)[0]) == COPY_FREE_VARS
+    if (!(lasti < 0 && _PyCode_CODE(co)->op.code == COPY_FREE_VARS
           && PyFunction_Check(frame->f_funcobj)))
     {
         /* Free vars are initialized */

Objects/genobject.c

@@ -332,11 +332,11 @@ _PyGen_yf(PyGenObject *gen)
             /* Return immediately if the frame didn't start yet. SEND
                always come after LOAD_CONST: a code object should not start
                with SEND */
-            assert(_Py_OPCODE(_PyCode_CODE(gen->gi_code)[0]) != SEND);
+            assert(_PyCode_CODE(gen->gi_code)[0].op.code != SEND);
             return NULL;
         }
         _Py_CODEUNIT next = frame->prev_instr[1];
-        if (_Py_OPCODE(next) != RESUME || _Py_OPARG(next) < 2)
+        if (next.op.code != RESUME || next.op.arg < 2)
         {
             /* Not in a yield from */
             return NULL;
@@ -371,9 +371,9 @@ gen_close(PyGenObject *gen, PyObject *args)
     _PyInterpreterFrame *frame = (_PyInterpreterFrame *)gen->gi_iframe;
     /* It is possible for the previous instruction to not be a
      * YIELD_VALUE if the debugger has changed the lineno. */
-    if (err == 0 && frame->prev_instr->opcode == YIELD_VALUE) {
-        assert(frame->prev_instr[1].opcode == RESUME);
-        int exception_handler_depth = frame->prev_instr->oparg;
+    if (err == 0 && frame->prev_instr[0].op.code == YIELD_VALUE) {
+        assert(frame->prev_instr[1].op.code == RESUME);
+        int exception_handler_depth = frame->prev_instr[0].op.code;
         assert(exception_handler_depth > 0);
         /* We can safely ignore the outermost try block
          * as it automatically generated to handle

Objects/typeobject.c

@@ -9507,8 +9507,8 @@ super_init_without_args(_PyInterpreterFrame *cframe, PyCodeObject *co,
         if (_PyInterpreterFrame_LASTI(cframe) >= 0) {
             // MAKE_CELL and COPY_FREE_VARS have no quickened forms, so no need
             // to use _PyOpcode_Deopt here:
-            assert(_Py_OPCODE(_PyCode_CODE(co)[0]) == MAKE_CELL ||
-                   _Py_OPCODE(_PyCode_CODE(co)[0]) == COPY_FREE_VARS);
+            assert(_PyCode_CODE(co)[0].op.code == MAKE_CELL ||
+                   _PyCode_CODE(co)[0].op.code == COPY_FREE_VARS);
             assert(PyCell_Check(firstarg));
             firstarg = PyCell_GET(firstarg);
         }

Python/bytecodes.c

@@ -246,9 +246,9 @@ dummy_func(
             DEOPT_IF(!PyUnicode_CheckExact(left), BINARY_OP);
             DEOPT_IF(Py_TYPE(right) != Py_TYPE(left), BINARY_OP);
             _Py_CODEUNIT true_next = next_instr[INLINE_CACHE_ENTRIES_BINARY_OP];
-            assert(_Py_OPCODE(true_next) == STORE_FAST ||
-                   _Py_OPCODE(true_next) == STORE_FAST__LOAD_FAST);
-            PyObject **target_local = &GETLOCAL(_Py_OPARG(true_next));
+            assert(true_next.op.code == STORE_FAST ||
+                   true_next.op.code == STORE_FAST__LOAD_FAST);
+            PyObject **target_local = &GETLOCAL(true_next.op.arg);
             DEOPT_IF(*target_local != left, BINARY_OP);
             STAT_INC(BINARY_OP, hit);
             /* Handle `left = left + right` or `left += right` for str.
@@ -1748,10 +1748,10 @@ dummy_func(
             Py_DECREF(left);
             Py_DECREF(right);
             ERROR_IF(cond == NULL, error);
-            assert(_Py_OPCODE(next_instr[1]) == POP_JUMP_IF_FALSE ||
-                   _Py_OPCODE(next_instr[1]) == POP_JUMP_IF_TRUE);
-            bool jump_on_true = _Py_OPCODE(next_instr[1]) == POP_JUMP_IF_TRUE;
-            int offset = _Py_OPARG(next_instr[1]);
+            assert(next_instr[1].op.code == POP_JUMP_IF_FALSE ||
+                   next_instr[1].op.code == POP_JUMP_IF_TRUE);
+            bool jump_on_true = next_instr[1].op.code == POP_JUMP_IF_TRUE;
+            int offset = next_instr[1].op.arg;
             int err = PyObject_IsTrue(cond);
             Py_DECREF(cond);
             if (err < 0) {
@@ -1774,7 +1774,7 @@ dummy_func(
             _Py_DECREF_SPECIALIZED(left, _PyFloat_ExactDealloc);
             _Py_DECREF_SPECIALIZED(right, _PyFloat_ExactDealloc);
             if (sign_ish & oparg) {
-                int offset = _Py_OPARG(next_instr[1]);
+                int offset = next_instr[1].op.arg;
                 JUMPBY(offset);
             }
         }
@@ -1795,7 +1795,7 @@ dummy_func(
             _Py_DECREF_SPECIALIZED(left, (destructor)PyObject_Free);
             _Py_DECREF_SPECIALIZED(right, (destructor)PyObject_Free);
             if (sign_ish & oparg) {
-                int offset = _Py_OPARG(next_instr[1]);
+                int offset = next_instr[1].op.arg;
                 JUMPBY(offset);
             }
         }
@@ -1814,7 +1814,7 @@ dummy_func(
             assert((oparg & 0xf) == COMPARISON_NOT_EQUALS || (oparg & 0xf) == COMPARISON_EQUALS);
             assert(COMPARISON_NOT_EQUALS + 1 == COMPARISON_EQUALS);
             if ((res + COMPARISON_NOT_EQUALS) & oparg) {
-                int offset = _Py_OPARG(next_instr[1]);
+                int offset = next_instr[1].op.arg;
                 JUMPBY(offset);
             }
         }
@@ -2122,7 +2122,7 @@ dummy_func(
                     _PyErr_Clear(tstate);
                 }
                 /* iterator ended normally */
-                assert(_Py_OPCODE(next_instr[INLINE_CACHE_ENTRIES_FOR_ITER + oparg]) == END_FOR);
+                assert(next_instr[INLINE_CACHE_ENTRIES_FOR_ITER + oparg].op.code == END_FOR);
                 Py_DECREF(iter);
                 STACK_SHRINK(1);
                 /* Jump forward oparg, then skip following END_FOR instruction */
@@ -2186,7 +2186,7 @@ dummy_func(
             DEOPT_IF(Py_TYPE(r) != &PyRangeIter_Type, FOR_ITER);
             STAT_INC(FOR_ITER, hit);
             _Py_CODEUNIT next = next_instr[INLINE_CACHE_ENTRIES_FOR_ITER];
-            assert(_PyOpcode_Deopt[_Py_OPCODE(next)] == STORE_FAST);
+            assert(_PyOpcode_Deopt[next.op.code] == STORE_FAST);
             if (r->len <= 0) {
                 STACK_SHRINK(1);
                 Py_DECREF(r);
@@ -2197,7 +2197,7 @@ dummy_func(
                 long value = r->start;
                 r->start = value + r->step;
                 r->len--;
-                if (_PyLong_AssignValue(&GETLOCAL(_Py_OPARG(next)), value) < 0) {
+                if (_PyLong_AssignValue(&GETLOCAL(next.op.arg), value) < 0) {
                     goto error;
                 }
                 // The STORE_FAST is already done.
@@ -2220,7 +2220,7 @@ dummy_func(
             gen->gi_exc_state.previous_item = tstate->exc_info;
             tstate->exc_info = &gen->gi_exc_state;
             JUMPBY(INLINE_CACHE_ENTRIES_FOR_ITER + oparg);
-            assert(_Py_OPCODE(*next_instr) == END_FOR);
+            assert(next_instr->op.code == END_FOR);
             DISPATCH_INLINED(gen_frame);
         }
 
@@ -2809,7 +2809,7 @@ dummy_func(
             STACK_SHRINK(3);
             // CALL + POP_TOP
             JUMPBY(INLINE_CACHE_ENTRIES_CALL + 1);
-            assert(_Py_OPCODE(next_instr[-1]) == POP_TOP);
+            assert(next_instr[-1].op.code == POP_TOP);
             DISPATCH();
         }
 
@@ -3118,8 +3118,8 @@ dummy_func(
         inst(EXTENDED_ARG, (--)) {
             assert(oparg);
             assert(cframe.use_tracing == 0);
-            opcode = _Py_OPCODE(*next_instr);
-            oparg = oparg << 8 | _Py_OPARG(*next_instr);
+            opcode = next_instr->op.code;
+            oparg = oparg << 8 | next_instr->op.arg;
             PRE_DISPATCH_GOTO();
             DISPATCH_GOTO();
         }

Python/ceval.c

@@ -132,8 +132,8 @@ lltrace_instruction(_PyInterpreterFrame *frame,
        objects enters the interpreter recursively. It is also slow.
        So you might want to comment it out. */
     dump_stack(frame, stack_pointer);
-    int oparg = _Py_OPARG(*next_instr);
-    int opcode = _Py_OPCODE(*next_instr);
+    int oparg = next_instr->op.arg;
+    int opcode = next_instr->op.code;
     const char *opname = _PyOpcode_OpName[opcode];
     assert(opname != NULL);
     int offset = (int)(next_instr - _PyCode_CODE(frame->f_code));
@@ -920,8 +920,8 @@ _PyEval_EvalFrameDefault(PyThreadState *tstate, _PyInterpreterFrame *frame, int
             // CPython hasn't ever traced the instruction after an EXTENDED_ARG.
             // Inline the EXTENDED_ARG here, so we can avoid branching there:
             INSTRUCTION_START(EXTENDED_ARG);
-            opcode = _Py_OPCODE(*next_instr);
-            oparg = oparg << 8 | _Py_OPARG(*next_instr);
+            opcode = next_instr->op.code;
+            oparg = oparg << 8 | next_instr->op.arg;
             // Make sure the next instruction isn't a RESUME, since that needs
             // to trace properly (and shouldn't have an EXTENDED_ARG, anyways):
             assert(opcode != RESUME);
@@ -946,7 +946,7 @@ _PyEval_EvalFrameDefault(PyThreadState *tstate, _PyInterpreterFrame *frame, int
 #endif
             /* Tell C compilers not to hold the opcode variable in the loop.
                next_instr points the current instruction without TARGET(). */
-            opcode = _Py_OPCODE(*next_instr);
+            opcode = next_instr->op.code;
             _PyErr_Format(tstate, PyExc_SystemError,
                           "%U:%d: unknown opcode %d",
                           frame->f_code->co_filename,
@@ -2196,7 +2196,7 @@ maybe_call_line_trace(Py_tracefunc func, PyObject *obj,
             (_PyInterpreterFrame_LASTI(frame) < instr_prev &&
              // SEND has no quickened forms, so no need to use _PyOpcode_Deopt
              // here:
-             _Py_OPCODE(*frame->prev_instr) != SEND);
+             frame->prev_instr->op.code != SEND);
         if (trace) {
             result = call_trace(func, obj, tstate, frame, PyTrace_LINE, Py_None);
         }

Python/ceval_macros.h

@@ -100,7 +100,7 @@
 
 #define DISPATCH_SAME_OPARG() \
     { \
-        opcode = _Py_OPCODE(*next_instr); \
+        opcode = next_instr->op.code; \
         PRE_DISPATCH_GOTO(); \
         opcode |= cframe.use_tracing OR_DTRACE_LINE; \
         DISPATCH_GOTO(); \
@@ -143,8 +143,8 @@ GETITEM(PyObject *v, Py_ssize_t i) {
 #define INSTR_OFFSET() ((int)(next_instr - _PyCode_CODE(frame->f_code)))
 #define NEXTOPARG()  do { \
         _Py_CODEUNIT word = *next_instr; \
-        opcode = _Py_OPCODE(word); \
-        oparg = _Py_OPARG(word); \
+        opcode = word.op.code; \
+        oparg = word.op.arg; \
     } while (0)
 #define JUMPTO(x)       (next_instr = _PyCode_CODE(frame->f_code) + (x))
 #define JUMPBY(x)       (next_instr += (x))
@@ -180,14 +180,14 @@ GETITEM(PyObject *v, Py_ssize_t i) {
 #if USE_COMPUTED_GOTOS
 #define PREDICT(op)             if (0) goto PREDICT_ID(op)
 #else
-#define PREDICT(op) \
+#define PREDICT(next_op) \
     do { \
         _Py_CODEUNIT word = *next_instr; \
-        opcode = _Py_OPCODE(word) | cframe.use_tracing OR_DTRACE_LINE; \
-        if (opcode == op) { \
-            oparg = _Py_OPARG(word); \
-            INSTRUCTION_START(op); \
-            goto PREDICT_ID(op); \
+        opcode = word.op.code | cframe.use_tracing OR_DTRACE_LINE; \
+        if (opcode == next_op) { \
+            oparg = word.op.arg; \
+            INSTRUCTION_START(next_op); \
+            goto PREDICT_ID(next_op); \
         } \
     } while(0)
 #endif