gh-106706: Streamline family syntax in cases generator DSL (#106716)

From `family(opname, STRUCTSIZE) = OPNAME + SPEC1 + ... +  SPECn;`
to `family(OPNAME, STRUCTSIZE) = SPEC1 + ... + SPECn;`

diff --git a/Misc/NEWS.d/next/Tools-Demos/2023-07-13-12-08-35.gh-issue-106706.29zp8E.rst b/Misc/NEWS.d/next/Tools-Demos/2023-07-13-12-08-35.gh-issue-106706.29zp8E.rst
new file mode 100644 (file)
index 0000000..bbd8e8e
--- /dev/null
+++ b/Misc/NEWS.d/next/Tools-Demos/2023-07-13-12-08-35.gh-issue-106706.29zp8E.rst
@@ -0,0 +1,3 @@
+Change bytecode syntax for families\r
+to remove redundant name matching\r
+pseudo syntax.\r

diff --git a/Python/bytecodes.c b/Python/bytecodes.c
index 3432b027713462e1905d78e0f58550e6593dffc7..3c3992c068b063c61bc10d4d9e9ece6894338b6a 100644 (file)
--- a/Python/bytecodes.c
+++ b/Python/bytecodes.c
@@ -284,8 +284,7 @@ dummy_func(
             res = Py_IsFalse(value) ? Py_True : Py_False;
         }
 
-        family(to_bool, INLINE_CACHE_ENTRIES_TO_BOOL) = {
-            TO_BOOL,
+        family(TO_BOOL, INLINE_CACHE_ENTRIES_TO_BOOL) = {
             TO_BOOL_ALWAYS_TRUE,
             TO_BOOL_BOOL,
             TO_BOOL_INT,
@@ -372,8 +371,7 @@ dummy_func(
             ERROR_IF(res == NULL, error);
         }
 
-        family(binary_op, INLINE_CACHE_ENTRIES_BINARY_OP) = {
-            BINARY_OP,
+        family(BINARY_OP, INLINE_CACHE_ENTRIES_BINARY_OP) = {
             BINARY_OP_MULTIPLY_INT,
             BINARY_OP_ADD_INT,
             BINARY_OP_SUBTRACT_INT,
@@ -507,8 +505,7 @@ dummy_func(
         macro(BINARY_OP_INPLACE_ADD_UNICODE) =
             _GUARD_BOTH_UNICODE + _BINARY_OP_INPLACE_ADD_UNICODE;
 
-        family(binary_subscr, INLINE_CACHE_ENTRIES_BINARY_SUBSCR) = {
-            BINARY_SUBSCR,
+        family(BINARY_SUBSCR, INLINE_CACHE_ENTRIES_BINARY_SUBSCR) = {
             BINARY_SUBSCR_DICT,
             BINARY_SUBSCR_GETITEM,
             BINARY_SUBSCR_LIST_INT,
@@ -643,8 +640,7 @@ dummy_func(
             ERROR_IF(err, error);
         }
 
-        family(store_subscr, INLINE_CACHE_ENTRIES_STORE_SUBSCR) = {
-            STORE_SUBSCR,
+        family(STORE_SUBSCR, INLINE_CACHE_ENTRIES_STORE_SUBSCR) = {
             STORE_SUBSCR_DICT,
             STORE_SUBSCR_LIST_INT,
         };
@@ -921,8 +917,7 @@ dummy_func(
             ERROR_IF(iter == NULL, error);
         }
 
-        family(send, INLINE_CACHE_ENTRIES_SEND) = {
-            SEND,
+        family(SEND, INLINE_CACHE_ENTRIES_SEND) = {
             SEND_GEN,
         };
 
@@ -1134,8 +1129,7 @@ dummy_func(
             }
         }
 
-        family(unpack_sequence, INLINE_CACHE_ENTRIES_UNPACK_SEQUENCE) = {
-            UNPACK_SEQUENCE,
+        family(UNPACK_SEQUENCE, INLINE_CACHE_ENTRIES_UNPACK_SEQUENCE) = {
             UNPACK_SEQUENCE_TWO_TUPLE,
             UNPACK_SEQUENCE_TUPLE,
             UNPACK_SEQUENCE_LIST,
@@ -1198,8 +1192,7 @@ dummy_func(
             ERROR_IF(res == 0, error);
         }
 
-        family(store_attr, INLINE_CACHE_ENTRIES_STORE_ATTR) = {
-            STORE_ATTR,
+        family(STORE_ATTR, INLINE_CACHE_ENTRIES_STORE_ATTR) = {
             STORE_ATTR_INSTANCE_VALUE,
             STORE_ATTR_SLOT,
             STORE_ATTR_WITH_HINT,
@@ -1298,8 +1291,7 @@ dummy_func(
 
         macro(LOAD_FROM_DICT_OR_GLOBALS) = _LOAD_FROM_DICT_OR_GLOBALS;
 
-        family(load_global, INLINE_CACHE_ENTRIES_LOAD_GLOBAL) = {
-            LOAD_GLOBAL,
+        family(LOAD_GLOBAL, INLINE_CACHE_ENTRIES_LOAD_GLOBAL) = {
             LOAD_GLOBAL_MODULE,
             LOAD_GLOBAL_BUILTIN,
         };
@@ -1647,8 +1639,7 @@ dummy_func(
             GO_TO_INSTRUCTION(LOAD_SUPER_ATTR);
         }
 
-        family(load_super_attr, INLINE_CACHE_ENTRIES_LOAD_SUPER_ATTR) = {
-            LOAD_SUPER_ATTR,
+        family(LOAD_SUPER_ATTR, INLINE_CACHE_ENTRIES_LOAD_SUPER_ATTR) = {
             LOAD_SUPER_ATTR_ATTR,
             LOAD_SUPER_ATTR_METHOD,
         };
@@ -1750,8 +1741,7 @@ dummy_func(
             }
         }
 
-        family(load_attr, INLINE_CACHE_ENTRIES_LOAD_ATTR) = {
-            LOAD_ATTR,
+        family(LOAD_ATTR, INLINE_CACHE_ENTRIES_LOAD_ATTR) = {
             LOAD_ATTR_INSTANCE_VALUE,
             LOAD_ATTR_MODULE,
             LOAD_ATTR_WITH_HINT,
@@ -2048,8 +2038,7 @@ dummy_func(
             Py_DECREF(owner);
         }
 
-        family(compare_op, INLINE_CACHE_ENTRIES_COMPARE_OP) = {
-            COMPARE_OP,
+        family(COMPARE_OP, INLINE_CACHE_ENTRIES_COMPARE_OP) = {
             COMPARE_OP_FLOAT,
             COMPARE_OP_INT,
             COMPARE_OP_STR,
@@ -2350,8 +2339,7 @@ dummy_func(
         // This is optimized by skipping that instruction and combining
         // its effect (popping 'iter' instead of pushing 'next'.)
 
-        family(for_iter, INLINE_CACHE_ENTRIES_FOR_ITER) = {
-            FOR_ITER,
+        family(FOR_ITER, INLINE_CACHE_ENTRIES_FOR_ITER) = {
             FOR_ITER_LIST,
             FOR_ITER_TUPLE,
             FOR_ITER_RANGE,
@@ -2810,8 +2798,7 @@ dummy_func(
 
         // Cache layout: counter/1, func_version/2
         // Neither CALL_INTRINSIC_1/2 nor CALL_FUNCTION_EX are members!
-        family(call, INLINE_CACHE_ENTRIES_CALL) = {
-            CALL,
+        family(CALL, INLINE_CACHE_ENTRIES_CALL) = {
             CALL_BOUND_METHOD_EXACT_ARGS,
             CALL_PY_EXACT_ARGS,
             CALL_PY_WITH_DEFAULTS,

diff --git a/Tools/cases_generator/generate_cases.py b/Tools/cases_generator/generate_cases.py
index 6589289121863b57da8b96fcc14756b6b8abd670..3edd8ee51ba64e06f88fc020b0ebaa34d007d3c6 100644 (file)
--- a/Tools/cases_generator/generate_cases.py
+++ b/Tools/cases_generator/generate_cases.py
@@ -438,7 +438,7 @@ class Instruction:
         """Write one instruction, sans prologue and epilogue."""
         # Write a static assertion that a family's cache size is correct
         if family := self.family:
-            if self.name == family.members[0]:
+            if self.name == family.name:
                 if cache_size := family.size:
                     out.emit(
                         f"static_assert({cache_size} == "
@@ -831,7 +831,7 @@ class Analyzer:
     def map_families(self) -> None:
         """Link instruction names back to their family, if they have one."""
         for family in self.families.values():
-            for member in family.members:
+            for member in [family.name] + family.members:
                 if member_instr := self.instrs.get(member):
                     if member_instr.family not in (family, None):
                         self.error(
@@ -855,8 +855,11 @@ class Analyzer:
         - All members must have the same cache, input and output effects
         """
         for family in self.families.values():
-            if len(family.members) < 2:
-                self.error(f"Family {family.name!r} has insufficient members", family)
+            if family.name not in self.macro_instrs and family.name not in self.instrs:
+                self.error(
+                    f"Family {family.name!r} has unknown instruction {family.name!r}",
+                    family,
+                )
             members = [
                 member
                 for member in family.members
@@ -867,10 +870,8 @@ class Analyzer:
                 self.error(
                     f"Family {family.name!r} has unknown members: {unknown}", family
                 )
-            if len(members) < 2:
-                continue
-            expected_effects = self.effect_counts(members[0])
-            for member in members[1:]:
+            expected_effects = self.effect_counts(family.name)
+            for member in members:
                 member_effects = self.effect_counts(member)
                 if member_effects != expected_effects:
                     self.error(
@@ -1311,11 +1312,10 @@ class Analyzer:
             self.out.emit("")
             self.out.emit("_specializations = {")
             for name, family in self.families.items():
-                assert len(family.members) > 1
                 with self.out.indent():
-                    self.out.emit(f"\"{family.members[0]}\": [")
+                    self.out.emit(f"\"{family.name}\": [")
                     with self.out.indent():
-                        for m in family.members[1:]:
+                        for m in family.members:
                             self.out.emit(f"\"{m}\",")
                     self.out.emit(f"],")
             self.out.emit("}")
@@ -1551,9 +1551,8 @@ class Analyzer:
                 self.out.emit(f"next_instr += {cache_adjust};")
 
             if (
-                last_instr
-                and (family := last_instr.family)
-                and mac.name == family.members[0]
+                (family := self.families.get(mac.name))
+                and mac.name == family.name
                 and (cache_size := family.size)
             ):
                 self.out.emit(

diff --git a/Tools/cases_generator/interpreter_definition.md b/Tools/cases_generator/interpreter_definition.md
index c03870ef59eb49fada63c1a36392fa9c85c08d71..f141848631d04ab2d4c29f7bd8d6d6428f0c9c31 100644 (file)
--- a/Tools/cases_generator/interpreter_definition.md
+++ b/Tools/cases_generator/interpreter_definition.md
@@ -347,7 +347,7 @@ For explanations see "Generating the interpreter" below.)
 
 ### Defining an instruction family
 
-A _family_ represents a specializable instruction and its specializations.
+A _family_ maps a specializable instruction to its specializations.
 
 Example: These opcodes all share the same instruction format):
 ```C

diff --git a/Tools/cases_generator/test_generator.py b/Tools/cases_generator/test_generator.py
index e374ac41e6a94d70bfe12bb15a8af6dc4c40b136..e44273429b74058ea77f523c0b1effed9ff1ff90 100644 (file)
--- a/Tools/cases_generator/test_generator.py
+++ b/Tools/cases_generator/test_generator.py
@@ -287,7 +287,7 @@ def test_macro_instruction():
         inst(OP3, (unused/5, arg2, left, right -- res)) {
             res = op3(arg2, left, right);
         }
-        family(op, INLINE_CACHE_ENTRIES_OP) = { OP, OP3 };
+        family(OP, INLINE_CACHE_ENTRIES_OP) = { OP3 };
     """
     output = """
         TARGET(OP1) {