gh-117901: Add option for compiler's codegen to save nested instruction sequences for introspection (#118007)

This commit is contained in:
Irit Katriel 2024-04-24 10:46:17 +01:00 committed by GitHub
parent 692e902c74
commit 0aa0fc3d3c
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 115 additions and 1 deletions

View File

@ -61,6 +61,7 @@ _PyJumpTargetLabel _PyInstructionSequence_NewLabel(_PyInstructionSequence *seq);
int _PyInstructionSequence_ApplyLabelMap(_PyInstructionSequence *seq);
int _PyInstructionSequence_InsertInstruction(_PyInstructionSequence *seq, int pos,
int opcode, int oparg, _Py_SourceLocation loc);
int _PyInstructionSequence_AddNested(_PyInstructionSequence *seq, _PyInstructionSequence *nested);
void PyInstructionSequence_Fini(_PyInstructionSequence *seq);
extern PyTypeObject _PyInstructionSequence_Type;

View File

@ -1,4 +1,5 @@
import textwrap
from test.support.bytecode_helper import CodegenTestCase
# Tests for the code-generation stage of the compiler.
@ -6,11 +7,19 @@ from test.support.bytecode_helper import CodegenTestCase
class IsolatedCodeGenTests(CodegenTestCase):
def assertInstructionsMatch_recursive(self, insts, expected_insts):
expected_nested = [i for i in expected_insts if isinstance(i, list)]
expected_insts = [i for i in expected_insts if not isinstance(i, list)]
self.assertInstructionsMatch(insts, expected_insts)
self.assertEqual(len(insts.get_nested()), len(expected_nested))
for n_insts, n_expected in zip(insts.get_nested(), expected_nested):
self.assertInstructionsMatch_recursive(n_insts, n_expected)
def codegen_test(self, snippet, expected_insts):
import ast
a = ast.parse(snippet, "my_file.py", "exec")
insts = self.generate_code(a)
self.assertInstructionsMatch(insts, expected_insts)
self.assertInstructionsMatch_recursive(insts, expected_insts)
def test_if_expression(self):
snippet = "42 if True else 24"
@ -55,6 +64,91 @@ class IsolatedCodeGenTests(CodegenTestCase):
]
self.codegen_test(snippet, expected)
def test_function(self):
snippet = textwrap.dedent("""
def f(x):
return x + 42
""")
expected = [
# Function definition
('RESUME', 0),
('LOAD_CONST', 0),
('MAKE_FUNCTION', None),
('STORE_NAME', 0),
('LOAD_CONST', 1),
('RETURN_VALUE', None),
[
# Function body
('RESUME', 0),
('LOAD_FAST', 0),
('LOAD_CONST', 1),
('BINARY_OP', 0),
('RETURN_VALUE', None),
('LOAD_CONST', 0),
('RETURN_VALUE', None),
]
]
self.codegen_test(snippet, expected)
def test_nested_functions(self):
snippet = textwrap.dedent("""
def f():
def h():
return 12
def g():
x = 1
y = 2
z = 3
u = 4
return 42
""")
expected = [
# Function definition
('RESUME', 0),
('LOAD_CONST', 0),
('MAKE_FUNCTION', None),
('STORE_NAME', 0),
('LOAD_CONST', 1),
('RETURN_VALUE', None),
[
# Function body
('RESUME', 0),
('LOAD_CONST', 1),
('MAKE_FUNCTION', None),
('STORE_FAST', 0),
('LOAD_CONST', 2),
('MAKE_FUNCTION', None),
('STORE_FAST', 1),
('LOAD_CONST', 0),
('RETURN_VALUE', None),
[
('RESUME', 0),
('NOP', None),
('LOAD_CONST', 1),
('RETURN_VALUE', None),
('LOAD_CONST', 0),
('RETURN_VALUE', None),
],
[
('RESUME', 0),
('LOAD_CONST', 1),
('STORE_FAST', 0),
('LOAD_CONST', 2),
('STORE_FAST', 1),
('LOAD_CONST', 3),
('STORE_FAST', 2),
('LOAD_CONST', 4),
('STORE_FAST', 3),
('NOP', None),
('LOAD_CONST', 5),
('RETURN_VALUE', None),
('LOAD_CONST', 0),
('RETURN_VALUE', None),
],
],
]
self.codegen_test(snippet, expected)
def test_syntax_error__return_not_in_function(self):
snippet = "return 42"
with self.assertRaisesRegex(SyntaxError, "'return' outside function"):

View File

@ -0,0 +1 @@
Add option for compiler's codegen to save nested instruction sequences for introspection.

View File

@ -285,6 +285,10 @@ struct compiler {
struct compiler_unit *u; /* compiler state for current block */
PyObject *c_stack; /* Python list holding compiler_unit ptrs */
PyArena *c_arena; /* pointer to memory allocation arena */
bool c_save_nested_seqs; /* if true, construct recursive instruction sequences
* (including instructions for nested code objects)
*/
};
#define INSTR_SEQUENCE(C) ((C)->u->u_instr_sequence)
@ -402,6 +406,7 @@ compiler_setup(struct compiler *c, mod_ty mod, PyObject *filename,
c->c_flags = *flags;
c->c_optimize = (optimize == -1) ? _Py_GetConfig()->optimization_level : optimize;
c->c_nestlevel = 0;
c->c_save_nested_seqs = false;
if (!_PyAST_Optimize(mod, arena, c->c_optimize, merged)) {
return ERROR;
@ -1290,6 +1295,11 @@ compiler_exit_scope(struct compiler *c)
// Don't call PySequence_DelItem() with an exception raised
PyObject *exc = PyErr_GetRaisedException();
instr_sequence *nested_seq = NULL;
if (c->c_save_nested_seqs) {
nested_seq = c->u->u_instr_sequence;
Py_INCREF(nested_seq);
}
c->c_nestlevel--;
compiler_unit_free(c->u);
/* Restore c->u to the parent unit. */
@ -1303,10 +1313,17 @@ compiler_exit_scope(struct compiler *c)
PyErr_FormatUnraisable("Exception ignored on removing "
"the last compiler stack item");
}
if (nested_seq != NULL) {
if (_PyInstructionSequence_AddNested(c->u->u_instr_sequence, nested_seq) < 0) {
PyErr_FormatUnraisable("Exception ignored on appending "
"nested instruction sequence");
}
}
}
else {
c->u = NULL;
}
Py_XDECREF(nested_seq);
PyErr_SetRaisedException(exc);
}
@ -7734,6 +7751,7 @@ _PyCompile_CodeGen(PyObject *ast, PyObject *filename, PyCompilerFlags *pflags,
_PyArena_Free(arena);
return NULL;
}
c->c_save_nested_seqs = true;
metadata = PyDict_New();
if (metadata == NULL) {