gh-105481: remove regen-opcode. Generated _PyOpcode_Caches in regen-cases. (#108367)

Include/internal/pycore_opcode.h

@@ -1,40 +0,0 @@
-// Auto-generated by Tools/build/generate_opcode_h.py from Lib/opcode.py
-
-#ifndef Py_INTERNAL_OPCODE_H
-#define Py_INTERNAL_OPCODE_H
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-#ifndef Py_BUILD_CORE
-#  error "this header requires Py_BUILD_CORE define"
-#endif
-
-#include "opcode.h"
-
-extern const uint8_t _PyOpcode_Caches[256];
-
-#ifdef NEED_OPCODE_TABLES
-
-const uint8_t _PyOpcode_Caches[256] = {
-    [LOAD_GLOBAL] = 4,
-    [BINARY_OP] = 1,
-    [UNPACK_SEQUENCE] = 1,
-    [COMPARE_OP] = 1,
-    [BINARY_SUBSCR] = 1,
-    [FOR_ITER] = 1,
-    [LOAD_SUPER_ATTR] = 1,
-    [LOAD_ATTR] = 9,
-    [STORE_ATTR] = 4,
-    [CALL] = 3,
-    [STORE_SUBSCR] = 1,
-    [SEND] = 1,
-    [JUMP_BACKWARD] = 1,
-    [TO_BOOL] = 3,
-};
-#endif   // NEED_OPCODE_TABLES
-
-#ifdef __cplusplus
-}
-#endif
-#endif  // !Py_INTERNAL_OPCODE_H

Include/internal/pycore_opcode_metadata.h

@@ -1814,6 +1814,26 @@ const char *const _PyOpcode_OpName[268] = {
 };
 #endif // NEED_OPCODE_METADATA
 
+extern const uint8_t _PyOpcode_Caches[256];
+#ifdef NEED_OPCODE_METADATA
+const uint8_t _PyOpcode_Caches[256] = {
+    [TO_BOOL] = 3,
+    [BINARY_OP] = 1,
+    [BINARY_SUBSCR] = 1,
+    [STORE_SUBSCR] = 1,
+    [SEND] = 1,
+    [UNPACK_SEQUENCE] = 1,
+    [STORE_ATTR] = 4,
+    [LOAD_GLOBAL] = 4,
+    [LOAD_SUPER_ATTR] = 1,
+    [LOAD_ATTR] = 9,
+    [COMPARE_OP] = 1,
+    [FOR_ITER] = 1,
+    [CALL] = 3,
+    [JUMP_BACKWARD] = 1,
+};
+#endif // NEED_OPCODE_METADATA
+
 extern const uint8_t _PyOpcode_Deopt[256];
 #ifdef NEED_OPCODE_METADATA
 const uint8_t _PyOpcode_Deopt[256] = {

Include/internal/pycore_opcode_utils.h

@@ -8,8 +8,7 @@ extern "C" {
 #  error "this header requires Py_BUILD_CORE define"
 #endif
 
-#include "pycore_opcode.h"        // JUMP_FORWARD
+#include "opcode_ids.h"
-
 
 #define MAX_REAL_OPCODE 254
 

Lib/opcode.py

@@ -5,48 +5,40 @@ operate on bytecodes (e.g. peephole optimizers).
 """
 
 
-# Note that __all__ is further extended below
+__all__ = ["cmp_op", "stack_effect", "hascompare", "opname", "opmap",
-__all__ = ["cmp_op", "stack_effect", "hascompare"]
+           "HAVE_ARGUMENT", "EXTENDED_ARG", "hasarg", "hasconst", "hasname",
+           "hasjump", "hasjrel", "hasjabs", "hasfree", "haslocal", "hasexc"]
 
 import _opcode
 from _opcode import stack_effect
 
-import sys
+from _opcode_metadata import (_specializations, _specialized_opmap, opmap,
-# The build uses older versions of Python which do not have _opcode_metadata
+                              HAVE_ARGUMENT, MIN_INSTRUMENTED_OPCODE)
-if sys.version_info[:2] >= (3, 13):
+EXTENDED_ARG = opmap['EXTENDED_ARG']
-    from _opcode_metadata import _specializations, _specialized_opmap
-    from _opcode_metadata import opmap, HAVE_ARGUMENT, MIN_INSTRUMENTED_OPCODE
-    EXTENDED_ARG = opmap['EXTENDED_ARG']
 
-    opname = ['<%r>' % (op,) for op in range(max(opmap.values()) + 1)]
+opname = ['<%r>' % (op,) for op in range(max(opmap.values()) + 1)]
-    for op, i in opmap.items():
+for op, i in opmap.items():
     opname[i] = op
 
-    __all__.extend(["opname", "opmap", "HAVE_ARGUMENT", "EXTENDED_ARG"])
-
 cmp_op = ('<', '<=', '==', '!=', '>', '>=')
 
-# The build uses older versions of Python which do not have _opcode.has_* functions
+# These lists are documented as part of the dis module's API
-if sys.version_info[:2] >= (3, 13):
+hasarg = [op for op in opmap.values() if _opcode.has_arg(op)]
-    # These lists are documented as part of the dis module's API
+hasconst = [op for op in opmap.values() if _opcode.has_const(op)]
-    hasarg = [op for op in opmap.values() if _opcode.has_arg(op)]
+hasname = [op for op in opmap.values() if _opcode.has_name(op)]
-    hasconst = [op for op in opmap.values() if _opcode.has_const(op)]
+hasjump = [op for op in opmap.values() if _opcode.has_jump(op)]
-    hasname = [op for op in opmap.values() if _opcode.has_name(op)]
+hasjrel = hasjump  # for backward compatibility
-    hasjump = [op for op in opmap.values() if _opcode.has_jump(op)]
+hasjabs = []
-    hasjrel = hasjump  # for backward compatibility
+hasfree = [op for op in opmap.values() if _opcode.has_free(op)]
-    hasjabs = []
+haslocal = [op for op in opmap.values() if _opcode.has_local(op)]
-    hasfree = [op for op in opmap.values() if _opcode.has_free(op)]
+hasexc = [op for op in opmap.values() if _opcode.has_exc(op)]
-    haslocal = [op for op in opmap.values() if _opcode.has_local(op)]
-    hasexc = [op for op in opmap.values() if _opcode.has_exc(op)]
 
-    __all__.extend(["hasarg", "hasconst", "hasname", "hasjump", "hasjrel",
-                    "hasjabs", "hasfree", "haslocal", "hasexc"])
 
-    _intrinsic_1_descs = _opcode.get_intrinsic1_descs()
+_intrinsic_1_descs = _opcode.get_intrinsic1_descs()
-    _intrinsic_2_descs = _opcode.get_intrinsic2_descs()
+_intrinsic_2_descs = _opcode.get_intrinsic2_descs()
-    _nb_ops = _opcode.get_nb_ops()
+_nb_ops = _opcode.get_nb_ops()
 
-    hascompare = [opmap["COMPARE_OP"]]
+hascompare = [opmap["COMPARE_OP"]]
 
 _cache_format = {
     "LOAD_GLOBAL": {

Makefile.pre.in

@@ -1321,7 +1321,7 @@ regen-limited-abi: all
 # Regenerate all generated files
 
 .PHONY: regen-all
-regen-all: regen-cases regen-opcode regen-typeslots \
+regen-all: regen-cases regen-typeslots \
 	regen-token regen-ast regen-keyword regen-sre regen-frozen clinic \
 	regen-pegen-metaparser regen-pegen regen-test-frozenmain \
 	regen-test-levenshtein regen-global-objects
@@ -1425,15 +1425,6 @@ regen-ast:
 	$(UPDATE_FILE) $(srcdir)/Include/internal/pycore_ast_state.h $(srcdir)/Include/internal/pycore_ast_state.h.new
 	$(UPDATE_FILE) $(srcdir)/Python/Python-ast.c $(srcdir)/Python/Python-ast.c.new
 
-.PHONY: regen-opcode
-regen-opcode:
-	# Regenerate Include/internal/pycore_opcode.h from Lib/opcode.py
-	# using Tools/build/generate_opcode_h.py
-	$(PYTHON_FOR_REGEN) $(srcdir)/Tools/build/generate_opcode_h.py \
-		$(srcdir)/Lib/opcode.py \
-		$(srcdir)/Include/internal/pycore_opcode.h.new
-	$(UPDATE_FILE) $(srcdir)/Include/internal/pycore_opcode.h $(srcdir)/Include/internal/pycore_opcode.h.new
-
 .PHONY: regen-token
 regen-token:
 	# Regenerate Doc/library/token-list.inc from Grammar/Tokens
@@ -1651,6 +1642,7 @@ PYTHON_HEADERS= \
 		$(srcdir)/Include/object.h \
 		$(srcdir)/Include/objimpl.h \
 		$(srcdir)/Include/opcode.h \
+		$(srcdir)/Include/opcode_ids.h \
 		$(srcdir)/Include/osdefs.h \
 		$(srcdir)/Include/osmodule.h \
 		$(srcdir)/Include/patchlevel.h \
@@ -1790,7 +1782,7 @@ PYTHON_HEADERS= \
 		$(srcdir)/Include/internal/pycore_object_state.h \
 		$(srcdir)/Include/internal/pycore_obmalloc.h \
 		$(srcdir)/Include/internal/pycore_obmalloc_init.h \
-		$(srcdir)/Include/internal/pycore_opcode.h \
+		$(srcdir)/Include/internal/pycore_opcode_metadata.h \
 		$(srcdir)/Include/internal/pycore_opcode_utils.h \
 		$(srcdir)/Include/internal/pycore_optimizer.h \
 		$(srcdir)/Include/internal/pycore_pathconfig.h \

Misc/NEWS.d/next/Core and Builtins/2023-08-23-14-54-15.gh-issue-105481.40q-c4.rst

@@ -0,0 +1,2 @@
+The regen-opcode build stage was removed and its work is now done in
+regen-cases.

Objects/codeobject.c

@@ -6,8 +6,7 @@
 #include "pycore_code.h"          // _PyCodeConstructor
 #include "pycore_frame.h"         // FRAME_SPECIALS_SIZE
 #include "pycore_interp.h"        // PyInterpreterState.co_extra_freefuncs
-#include "pycore_opcode.h"        // _PyOpcode_Caches
+#include "pycore_opcode_metadata.h" // _PyOpcode_Deopt, _PyOpcode_Caches
-#include "pycore_opcode_metadata.h" // _PyOpcode_Deopt
 #include "pycore_pystate.h"       // _PyInterpreterState_GET()
 #include "pycore_setobject.h"     // _PySet_NextEntry()
 #include "pycore_tuple.h"         // _PyTuple_ITEMS()

Objects/frameobject.c

@@ -6,8 +6,7 @@
 #include "pycore_function.h"      // _PyFunction_FromConstructor()
 #include "pycore_moduleobject.h"  // _PyModule_GetDict()
 #include "pycore_object.h"        // _PyObject_GC_UNTRACK()
-#include "pycore_opcode.h"        // _PyOpcode_Caches
+#include "pycore_opcode_metadata.h" // _PyOpcode_Deopt, _PyOpcode_Caches
-#include "pycore_opcode_metadata.h" // _PyOpcode_Deopt
 
 
 #include "frameobject.h"          // PyFrameObject

PCbuild/regen.targets

@@ -13,8 +13,6 @@
     <_ASTOutputs Include="$(PySourcePath)Python\Python-ast.c">
       <Argument>-C</Argument>
     </_ASTOutputs>
-    <_OpcodeSources Include="$(PySourcePath)Tools\build\generate_opcode_h.py;$(PySourcePath)Lib\opcode.py" />
-    <_OpcodeOutputs Include="$(PySourcePath)Include\internal\pycore_opcode.h" />
     <_TokenSources Include="$(PySourcePath)Grammar\Tokens" />
     <_TokenOutputs Include="$(PySourcePath)Doc\library\token-list.inc">
       <Format>rst</Format>
@@ -34,7 +32,7 @@
 
   <Target Name="_TouchRegenSources" Condition="$(ForceRegen) == 'true'">
     <Message Text="Touching source files to force regeneration" Importance="high" />
-    <Touch Files="@(_PegenSources);@(_ASTSources);@(_OpcodeSources);@(_TokenSources);@(_KeywordOutputs)"
+    <Touch Files="@(_PegenSources);@(_ASTSources);@(_TokenSources);@(_KeywordOutputs)"
            AlwaysCreate="False" />
   </Target>
 
@@ -55,14 +53,6 @@
           WorkingDirectory="$(PySourcePath)" />
   </Target>
 
-  <Target Name="_RegenOpcodes"
-          Inputs="@(_OpcodeSources)" Outputs="@(_OpcodeOutputs)"
-          DependsOnTargets="FindPythonForBuild">
-    <Message Text="Regenerate @(_OpcodeOutputs->'%(Filename)%(Extension)',' ')" Importance="high" />
-    <Exec Command="$(PythonForBuild) Tools\build\generate_opcode_h.py Lib\opcode.py Include\internal\pycore_opcode.h Include\internal\pycore_intrinsics.h"
-          WorkingDirectory="$(PySourcePath)" />
-  </Target>
-
   <Target Name="_RegenTokens"
           Inputs="@(_TokenSources)" Outputs="@(_TokenOutputs)"
           DependsOnTargets="FindPythonForBuild">
@@ -89,7 +79,7 @@
 
   <Target Name="Regen"
           Condition="$(Configuration) != 'PGUpdate'"
-          DependsOnTargets="_TouchRegenSources;_RegenPegen;_RegenAST_H;_RegenOpcodes;_RegenTokens;_RegenKeywords;_RegenGlobalObjects">
+          DependsOnTargets="_TouchRegenSources;_RegenPegen;_RegenAST_H;_RegenTokens;_RegenKeywords;_RegenGlobalObjects">
     <Message Text="Generated sources are up to date" Importance="high" />
   </Target>
 

Python/assemble.c

@@ -3,9 +3,8 @@
 #include "Python.h"
 #include "pycore_code.h"            // write_location_entry_start()
 #include "pycore_compile.h"
-#include "pycore_opcode.h"          // _PyOpcode_Caches[] and opcode category macros
 #include "pycore_opcode_utils.h"    // IS_BACKWARDS_JUMP_OPCODE
-#include "pycore_opcode_metadata.h" // IS_PSEUDO_INSTR
+#include "pycore_opcode_metadata.h" // IS_PSEUDO_INSTR, _PyOpcode_Caches
 
 
 #define DEFAULT_CODE_SIZE 128

Python/ceval.c

@@ -14,8 +14,7 @@
 #include "pycore_long.h"          // _PyLong_GetZero()
 #include "pycore_moduleobject.h"  // PyModuleObject
 #include "pycore_object.h"        // _PyObject_GC_TRACK()
-#include "pycore_opcode.h"        // EXTRA_CASES
+#include "pycore_opcode_metadata.h" // EXTRA_CASES
-#include "pycore_opcode_metadata.h"
 #include "pycore_opcode_utils.h"  // MAKE_FUNCTION_*
 #include "pycore_pyerrors.h"      // _PyErr_GetRaisedException()
 #include "pycore_pystate.h"       // _PyInterpreterState_GET()

Python/compile.c

@@ -24,6 +24,7 @@
 #include <stdbool.h>
 
 #include "Python.h"
+#include "opcode.h"
 #include "pycore_ast.h"           // _PyAST_GetDocString()
 #define NEED_OPCODE_TABLES
 #include "pycore_opcode_utils.h"

Python/executor.c

@@ -1,5 +1,7 @@
 #include "Python.h"
 
+#include "opcode.h"
+
 #include "pycore_call.h"
 #include "pycore_ceval.h"
 #include "pycore_dict.h"

Python/instrumentation.c

@@ -1,4 +1,7 @@
 #include "Python.h"
+
+#include "opcode_ids.h"
+
 #include "pycore_call.h"
 #include "pycore_frame.h"
 #include "pycore_interp.h"
@@ -6,8 +9,7 @@
 #include "pycore_modsupport.h"    // _PyModule_CreateInitialized()
 #include "pycore_namespace.h"
 #include "pycore_object.h"
-#include "pycore_opcode.h"
+#include "pycore_opcode_metadata.h" // IS_VALID_OPCODE, _PyOpcode_Caches
-#include "pycore_opcode_metadata.h" // IS_VALID_OPCODE
 #include "pycore_pyerrors.h"
 #include "pycore_pystate.h"       // _PyInterpreterState_GET()
 

Python/optimizer.c

@@ -1,7 +1,6 @@
 #include "Python.h"
 #include "opcode.h"
 #include "pycore_interp.h"
-#include "pycore_opcode.h"
 #include "pycore_opcode_metadata.h"
 #include "pycore_opcode_utils.h"
 #include "pycore_optimizer.h"

Python/optimizer_analysis.c

@@ -1,7 +1,6 @@
 #include "Python.h"
 #include "opcode.h"
 #include "pycore_interp.h"
-#include "pycore_opcode.h"
 #include "pycore_opcode_metadata.h"
 #include "pycore_opcode_utils.h"
 #include "pycore_pystate.h"       // _PyInterpreterState_GET()

Python/specialize.c

@@ -1,4 +1,7 @@
 #include "Python.h"
+
+#include "opcode.h"
+
 #include "pycore_code.h"
 #include "pycore_descrobject.h"   // _PyMethodWrapper_Type
 #include "pycore_dict.h"
@@ -7,7 +10,7 @@
 #include "pycore_long.h"
 #include "pycore_moduleobject.h"
 #include "pycore_object.h"
-#include "pycore_opcode.h"        // _PyOpcode_Caches
+#include "pycore_opcode_metadata.h" // _PyOpcode_Caches
 #include "pycore_pylifecycle.h"   // _PyOS_URandomNonblock()
 
 

Tools/build/generate_opcode_h.py

@@ -1,67 +0,0 @@
-# This script generates the pycore_opcode.h header file.
-
-import sys
-import tokenize
-
-SCRIPT_NAME = "Tools/build/generate_opcode_h.py"
-PYTHON_OPCODE = "Lib/opcode.py"
-
-internal_header = f"""
-// Auto-generated by {SCRIPT_NAME} from {PYTHON_OPCODE}
-
-#ifndef Py_INTERNAL_OPCODE_H
-#define Py_INTERNAL_OPCODE_H
-#ifdef __cplusplus
-extern "C" {{
-#endif
-
-#ifndef Py_BUILD_CORE
-#  error "this header requires Py_BUILD_CORE define"
-#endif
-
-#include "opcode.h"
-""".lstrip()
-
-internal_footer = """
-#ifdef __cplusplus
-}
-#endif
-#endif  // !Py_INTERNAL_OPCODE_H
-"""
-
-DEFINE = "#define {:<38} {:>3}\n"
-
-UINT32_MASK = (1<<32)-1
-
-def get_python_module_dict(filename):
-    mod = {}
-    with tokenize.open(filename) as fp:
-        code = fp.read()
-    exec(code, mod)
-    return mod
-
-def main(opcode_py,
-         internal_opcode_h='Include/internal/pycore_opcode.h'):
-
-    opcode = get_python_module_dict(opcode_py)
-
-    with open(internal_opcode_h, 'w') as iobj:
-        iobj.write(internal_header)
-
-        iobj.write("\nextern const uint8_t _PyOpcode_Caches[256];\n")
-        iobj.write("\n#ifdef NEED_OPCODE_TABLES\n")
-
-        iobj.write("\nconst uint8_t _PyOpcode_Caches[256] = {\n")
-        for name, entries in opcode["_inline_cache_entries"].items():
-            iobj.write(f"    [{name}] = {entries},\n")
-        iobj.write("};\n")
-
-        iobj.write("#endif   // NEED_OPCODE_TABLES\n")
-
-        iobj.write(internal_footer)
-
-    print(f"{internal_opcode_h} regenerated from {opcode_py}")
-
-
-if __name__ == '__main__':
-    main(sys.argv[1], sys.argv[2])

Tools/cases_generator/generate_cases.py

@@ -540,6 +540,18 @@ class Generator(Analyzer):
                 for name in self.opmap:
                     self.out.emit(f'[{name}] = "{name}",')
 
+            with self.metadata_item(
+                f"const uint8_t _PyOpcode_Caches[256]",
+                "=",
+                ";",
+            ):
+                for name, _ in self.families.items():
+                    instr = self.instrs[name]
+                    if instr.cache_offset > 0:
+                        self.out.emit(f'[{name}] = {instr.cache_offset},')
+                # Irregular case:
+                self.out.emit('[JUMP_BACKWARD] = 1,')
+
             deoptcodes = {}
             for name, op in self.opmap.items():
                 if op < 256: