Python 3.14 text signatures, sizeof, and SyntaxError offsets

majit/majit-translate/src/codewriter/assembler.rs

@@ -3283,6 +3283,8 @@ fn type_flag_from_str(
     use majit_ir::descr::ArrayFlag;
     let word = crate::layout::target_word_size();
     match type_str {
+        // descr.py:241-254 raw Ptr parity; see call.rs::get_type_flag.
+        "*const u8" => (ArrayFlag::Unsigned, majit_ir::value::Type::Int, word),
         s if s.starts_with('&')
             || s.starts_with("Box<")
             || s.starts_with("Arc<")

majit/majit-translate/src/codewriter/call.rs

@@ -7487,6 +7487,16 @@ pub(crate) fn get_type_flag(
 ) -> (majit_ir::descr::ArrayFlag, majit_ir::value::Type, usize) {
     use majit_ir::descr::ArrayFlag;
     match type_str {
+        // descr.py:241-254: Ptr whose pointee has `_gckind == 'raw'` is an
+        // unsigned, int-banked word.  Pyre's mapdict shape pointers are erased
+        // to `*const u8`; they are immortal raw identities, not GC references.
+        // Keep `*mut PyObject` and other erased managed pointers on the
+        // conservative Ref fallback below.
+        "*const u8" => (
+            ArrayFlag::Unsigned,
+            majit_ir::value::Type::Int,
+            crate::layout::target_word_size(),
+        ),
         // RPython: isinstance(TYPE, lltype.Ptr) and TYPE.TO._gckind == 'gc' → FLAG_POINTER
         s if s.starts_with('&')
             || s.starts_with("Box<")
@@ -9617,6 +9627,17 @@ mod tests {
         assert!(!cc.is_known_struct("*mut PyObject"));
     }
 
+    #[test]
+    fn raw_byte_identity_pointer_is_int_banked() {
+        use majit_ir::descr::ArrayFlag;
+        use majit_ir::value::Type;
+
+        let (flag, field_type, size) = get_type_flag("*const u8");
+        assert_eq!(flag, ArrayFlag::Unsigned);
+        assert_eq!(field_type, Type::Int);
+        assert_eq!(size, crate::layout::target_word_size());
+    }
+
     #[derive(Debug)]
     struct StubVInfo {
         vtypeptr_id: usize,

pyre/extra_tests/parity_tests/bool_text_signatures_python314.py

@@ -0,0 +1,25 @@
+"""CPython 3.14 text signatures for bool's own descriptors."""
+
+import inspect
+
+
+EXPECTED = {
+    "__new__": "($type, *args, **kwargs)",
+    "__repr__": "($self, /)",
+    "__invert__": "($self, /)",
+    "__and__": "($self, value, /)",
+    "__rand__": "($self, value, /)",
+    "__or__": "($self, value, /)",
+    "__ror__": "($self, value, /)",
+    "__xor__": "($self, value, /)",
+    "__rxor__": "($self, value, /)",
+}
+
+for name, signature in EXPECTED.items():
+    descriptor = bool.__dict__[name]
+    assert descriptor.__text_signature__ == signature, name
+
+assert str(inspect.signature(bool.__repr__)) == "(self, /)"
+assert str(inspect.signature(bool.__and__)) == "(self, value, /)"
+
+print("OK")

pyre/extra_tests/parity_tests/builtin_text_signatures_python314.py

@@ -0,0 +1,65 @@
+"""CPython 3.14 text signatures for functions in the builtins module."""
+
+import builtins
+import inspect
+
+
+EXPECTED = {
+    "__import__": "($module, /, name, globals=None, locals=None, fromlist=(),\n           level=0)",
+    "abs": "($module, x, /)",
+    "aiter": "($module, async_iterable, /)",
+    "all": "($module, iterable, /)",
+    "anext": "($module, aiterator, default=<unrepresentable>, /)",
+    "any": "($module, iterable, /)",
+    "ascii": "($module, obj, /)",
+    "bin": "($module, number, /)",
+    "breakpoint": "($module, /, *args, **kws)",
+    "callable": "($module, obj, /)",
+    "chr": "($module, i, /)",
+    "compile": "($module, /, source, filename, mode, flags=0,\n        dont_inherit=False, optimize=-1, *, _feature_version=-1)",
+    "delattr": "($module, obj, name, /)",
+    "divmod": "($module, x, y, /)",
+    "eval": "($module, source, /, globals=None, locals=None)",
+    "exec": "($module, source, /, globals=None, locals=None, *, closure=None)",
+    "format": "($module, value, format_spec='', /)",
+    "globals": "($module, /)",
+    "hasattr": "($module, obj, name, /)",
+    "hash": "($module, obj, /)",
+    "hex": "($module, number, /)",
+    "id": "($module, obj, /)",
+    "input": "($module, prompt='', /)",
+    "isinstance": "($module, obj, class_or_tuple, /)",
+    "issubclass": "($module, cls, class_or_tuple, /)",
+    "len": "($module, obj, /)",
+    "locals": "($module, /)",
+    "oct": "($module, number, /)",
+    "open": "($module, /, file, mode='r', buffering=-1, encoding=None,\n     errors=None, newline=None, closefd=True, opener=None)",
+    "ord": "($module, character, /)",
+    "pow": "($module, /, base, exp, mod=None)",
+    "print": "($module, /, *args, sep=' ', end='\\n', file=None, flush=False)",
+    "repr": "($module, obj, /)",
+    "round": "($module, /, number, ndigits=None)",
+    "setattr": "($module, obj, name, value, /)",
+    "sorted": "($module, iterable, /, *, key=None, reverse=False)",
+    "sum": "($module, iterable, /, start=0)",
+}
+
+for name, signature in EXPECTED.items():
+    assert getattr(builtins, name).__text_signature__ == signature, name
+
+for name in ("__build_class__", "dir", "getattr", "iter", "max", "min", "next", "vars"):
+    assert getattr(builtins, name).__text_signature__ is None, name
+
+assert str(inspect.signature(len)) == "(obj, /)"
+assert str(inspect.signature(sorted)) == (
+    "(iterable, /, *, key=None, reverse=False)"
+)
+assert str(inspect.signature(open)) == (
+    "(file, mode='r', buffering=-1, encoding=None, errors=None, newline=None, "
+    "closefd=True, opener=None)"
+)
+assert str(inspect.signature(print)) == (
+    "(*args, sep=' ', end='\\n', file=None, flush=False)"
+)
+
+print("OK")

pyre/extra_tests/parity_tests/bytearray_text_signatures_python314.py

@@ -0,0 +1,101 @@
+"""CPython 3.14 text signatures for bytearray descriptors."""
+
+import inspect
+
+
+EXPECTED = {
+    "__new__": "($type, *args, **kwargs)",
+    "__repr__": "($self, /)",
+    "__str__": "($self, /)",
+    "__lt__": "($self, value, /)",
+    "__le__": "($self, value, /)",
+    "__eq__": "($self, value, /)",
+    "__ne__": "($self, value, /)",
+    "__gt__": "($self, value, /)",
+    "__ge__": "($self, value, /)",
+    "__iter__": "($self, /)",
+    "__init__": "($self, /, *args, **kwargs)",
+    "__buffer__": "($self, flags, /)",
+    "__release_buffer__": "($self, buffer, /)",
+    "__mod__": "($self, value, /)",
+    "__rmod__": "($self, value, /)",
+    "__len__": "($self, /)",
+    "__getitem__": "($self, key, /)",
+    "__setitem__": "($self, key, value, /)",
+    "__delitem__": "($self, key, /)",
+    "__add__": "($self, value, /)",
+    "__mul__": "($self, value, /)",
+    "__rmul__": "($self, value, /)",
+    "__contains__": "($self, key, /)",
+    "__iadd__": "($self, value, /)",
+    "__imul__": "($self, value, /)",
+    "__alloc__": "($self, /)",
+    "__reduce__": "($self, /)",
+    "__reduce_ex__": "($self, proto=0, /)",
+    "__sizeof__": "($self, /)",
+    "append": "($self, item, /)",
+    "capitalize": "($self, /)",
+    "center": "($self, width, fillchar=b' ', /)",
+    "clear": "($self, /)",
+    "copy": "($self, /)",
+    "count": "($self, sub[, start[, end]], /)",
+    "decode": "($self, /, encoding='utf-8', errors='strict')",
+    "endswith": "($self, suffix[, start[, end]], /)",
+    "expandtabs": "($self, /, tabsize=8)",
+    "extend": "($self, iterable_of_ints, /)",
+    "find": "($self, sub[, start[, end]], /)",
+    "hex": "($self, /, sep=<unrepresentable>, bytes_per_sep=1)",
+    "index": "($self, sub[, start[, end]], /)",
+    "insert": "($self, index, item, /)",
+    "isalnum": "($self, /)",
+    "isalpha": "($self, /)",
+    "isascii": "($self, /)",
+    "isdigit": "($self, /)",
+    "islower": "($self, /)",
+    "isspace": "($self, /)",
+    "istitle": "($self, /)",
+    "isupper": "($self, /)",
+    "join": "($self, iterable_of_bytes, /)",
+    "ljust": "($self, width, fillchar=b' ', /)",
+    "lower": "($self, /)",
+    "lstrip": "($self, bytes=None, /)",
+    "partition": "($self, sep, /)",
+    "pop": "($self, index=-1, /)",
+    "remove": "($self, value, /)",
+    "replace": "($self, old, new, count=-1, /)",
+    "removeprefix": "($self, prefix, /)",
+    "removesuffix": "($self, suffix, /)",
+    "resize": "($self, size, /)",
+    "reverse": "($self, /)",
+    "rfind": "($self, sub[, start[, end]], /)",
+    "rindex": "($self, sub[, start[, end]], /)",
+    "rjust": "($self, width, fillchar=b' ', /)",
+    "rpartition": "($self, sep, /)",
+    "rsplit": "($self, /, sep=None, maxsplit=-1)",
+    "rstrip": "($self, bytes=None, /)",
+    "split": "($self, /, sep=None, maxsplit=-1)",
+    "splitlines": "($self, /, keepends=False)",
+    "startswith": "($self, prefix[, start[, end]], /)",
+    "strip": "($self, bytes=None, /)",
+    "swapcase": "($self, /)",
+    "title": "($self, /)",
+    "translate": "($self, table, /, delete=b'')",
+    "upper": "($self, /)",
+    "zfill": "($self, width, /)",
+}
+
+for name, signature in EXPECTED.items():
+    assert bytearray.__dict__[name].__text_signature__ == signature, name
+
+raw_maketrans = bytearray.__dict__["maketrans"]
+assert not hasattr(raw_maketrans, "__text_signature__")
+assert bytearray.maketrans.__text_signature__ == "(frm, to, /)"
+assert bytearray.__dict__["fromhex"].__text_signature__ == "($type, string, /)"
+
+assert str(inspect.signature(bytearray.decode)) == (
+    "(self, /, encoding='utf-8', errors='strict')"
+)
+assert str(inspect.signature(bytearray.resize)) == "(self, size, /)"
+assert str(inspect.signature(bytearray.fromhex)) == "(string, /)"
+
+print("OK")

pyre/extra_tests/parity_tests/bytes_text_signatures_python314.py

@@ -0,0 +1,85 @@
+"""CPython 3.14 text signatures for bytes descriptors."""
+
+import inspect
+
+
+EXPECTED = {
+    "__new__": "($type, *args, **kwargs)",
+    "__repr__": "($self, /)",
+    "__hash__": "($self, /)",
+    "__str__": "($self, /)",
+    "__lt__": "($self, value, /)",
+    "__le__": "($self, value, /)",
+    "__eq__": "($self, value, /)",
+    "__ne__": "($self, value, /)",
+    "__gt__": "($self, value, /)",
+    "__ge__": "($self, value, /)",
+    "__iter__": "($self, /)",
+    "__buffer__": "($self, flags, /)",
+    "__mod__": "($self, value, /)",
+    "__rmod__": "($self, value, /)",
+    "__len__": "($self, /)",
+    "__getitem__": "($self, key, /)",
+    "__add__": "($self, value, /)",
+    "__mul__": "($self, value, /)",
+    "__rmul__": "($self, value, /)",
+    "__contains__": "($self, key, /)",
+    "__getnewargs__": "($self, /)",
+    "__bytes__": "($self, /)",
+    "capitalize": "($self, /)",
+    "center": "($self, width, fillchar=b' ', /)",
+    "count": "($self, sub[, start[, end]], /)",
+    "decode": "($self, /, encoding='utf-8', errors='strict')",
+    "endswith": "($self, suffix[, start[, end]], /)",
+    "expandtabs": "($self, /, tabsize=8)",
+    "find": "($self, sub[, start[, end]], /)",
+    "hex": "($self, /, sep=<unrepresentable>, bytes_per_sep=1)",
+    "index": "($self, sub[, start[, end]], /)",
+    "isalnum": "($self, /)",
+    "isalpha": "($self, /)",
+    "isascii": "($self, /)",
+    "isdigit": "($self, /)",
+    "islower": "($self, /)",
+    "isspace": "($self, /)",
+    "istitle": "($self, /)",
+    "isupper": "($self, /)",
+    "join": "($self, iterable_of_bytes, /)",
+    "ljust": "($self, width, fillchar=b' ', /)",
+    "lower": "($self, /)",
+    "lstrip": "($self, bytes=None, /)",
+    "partition": "($self, sep, /)",
+    "replace": "($self, old, new, count=-1, /)",
+    "removeprefix": "($self, prefix, /)",
+    "removesuffix": "($self, suffix, /)",
+    "rfind": "($self, sub[, start[, end]], /)",
+    "rindex": "($self, sub[, start[, end]], /)",
+    "rjust": "($self, width, fillchar=b' ', /)",
+    "rpartition": "($self, sep, /)",
+    "rsplit": "($self, /, sep=None, maxsplit=-1)",
+    "rstrip": "($self, bytes=None, /)",
+    "split": "($self, /, sep=None, maxsplit=-1)",
+    "splitlines": "($self, /, keepends=False)",
+    "startswith": "($self, prefix[, start[, end]], /)",
+    "strip": "($self, bytes=None, /)",
+    "swapcase": "($self, /)",
+    "title": "($self, /)",
+    "translate": "($self, table, /, delete=b'')",
+    "upper": "($self, /)",
+    "zfill": "($self, width, /)",
+}
+
+for name, signature in EXPECTED.items():
+    assert bytes.__dict__[name].__text_signature__ == signature, name
+
+raw_maketrans = bytes.__dict__["maketrans"]
+assert not hasattr(raw_maketrans, "__text_signature__")
+assert bytes.maketrans.__text_signature__ == "(frm, to, /)"
+assert bytes.__dict__["fromhex"].__text_signature__ == "($type, string, /)"
+
+assert str(inspect.signature(bytes.decode)) == (
+    "(self, /, encoding='utf-8', errors='strict')"
+)
+assert str(inspect.signature(bytes.replace)) == "(self, old, new, count=-1, /)"
+assert str(inspect.signature(bytes.fromhex)) == "(string, /)"
+
+print("OK")

pyre/extra_tests/parity_tests/complex_text_signatures_python314.py

@@ -0,0 +1,38 @@
+"""CPython 3.14 text signatures for complex descriptors."""
+
+import inspect
+
+
+VALUE_BINARY = {
+    name: "($self, value, /)"
+    for name in (
+        "__lt__", "__le__", "__eq__", "__ne__", "__gt__", "__ge__",
+        "__add__", "__radd__", "__sub__", "__rsub__", "__mul__", "__rmul__",
+        "__truediv__", "__rtruediv__",
+    )
+}
+SELF_ONLY = {
+    name: "($self, /)"
+    for name in (
+        "__repr__", "__hash__", "__neg__", "__pos__", "__abs__", "__bool__",
+        "conjugate", "__complex__", "__getnewargs__",
+    )
+}
+EXPECTED = {
+    "__new__": "($type, *args, **kwargs)",
+    **VALUE_BINARY,
+    **SELF_ONLY,
+    "__pow__": "($self, value, mod=None, /)",
+    "__rpow__": "($self, value, mod=None, /)",
+    "from_number": "($type, number, /)",
+    "__format__": "($self, format_spec, /)",
+}
+
+for name, signature in EXPECTED.items():
+    assert complex.__dict__[name].__text_signature__ == signature, name
+
+assert str(inspect.signature(complex.from_number)) == "(number, /)"
+assert str(inspect.signature(complex.__pow__)) == "(self, value, mod=None, /)"
+assert str(inspect.signature(complex.conjugate)) == "(self, /)"
+
+print("OK")

pyre/extra_tests/parity_tests/dict_set_sizeof_python314.py

@@ -0,0 +1,18 @@
+"""Python 3.14 ``__sizeof__`` surface for dict and set-like types."""
+
+for typ in (dict, set, frozenset):
+    assert "__sizeof__" in typ.__dict__
+    assert typ.__sizeof__.__text_signature__ == "($self, /)"
+
+assert dict().__sizeof__() == 48
+assert {0: None}.__sizeof__() == 208
+assert {str(i): None for i in range(6)}.__sizeof__() == 256
+assert dict.fromkeys(range(11)).__sizeof__() == 616
+
+for typ in (set, frozenset):
+    assert typ().__sizeof__() == 200
+    assert typ(range(4)).__sizeof__() == 200
+    assert typ(range(5)).__sizeof__() == 712
+    assert typ(range(19)).__sizeof__() == 2248
+
+print("OK")