Accept __index__ for the remaining int-bound arguments (#8999)

#8828 made bin(), chr() and several other arguments that were bound
as PyIntRef go through __index__, but these still downcast to int and
reject an object that only defines __index__: the start and end of the
str/bytes/bytearray find family, enumerate(start=),
itertools.combinations/combinations_with_replacement/batched,
_thread.stack_size, resource.getrlimit/setrlimit, pwd.getpwuid,
grp.getgrgid, fcntl.fcntl/lockf and mmap.move/madvise. CPython converts
all of them with _PyNumber_Index or _PyEval_SliceIndex.

Bind them as ArgIndex, as #8828 did for the others.

test_mmap.test_madvise and test_mmap_closed_by_int_scenarios now pass,
so drop their expectedFailure markers.

Assisted-by: Claude Code:claude-opus-5-5

Lib/test/test_mmap.py

@@ -897,7 +897,6 @@ class MmapTests(unittest.TestCase):
                     match = closed_mmap_repr_pat.match(repr(mm))
                     self.assertIsNotNone(match)
 
-    @unittest.expectedFailure  # TODO: RUSTPYTHON
     @unittest.skipUnless(hasattr(mmap.mmap, 'madvise'), 'needs madvise')
     def test_madvise(self):
         size = 2 * PAGESIZE
@@ -1039,7 +1038,6 @@ class MmapTests(unittest.TestCase):
         self.assertEqual(m1[:data_length], data)
         self.assertEqual(m2[:data_length], data)
 
-    @unittest.expectedFailure  # TODO: RUSTPYTHON
     def test_mmap_closed_by_int_scenarios(self):
         """
         gh-103987: Test that mmap objects raise ValueError

crates/stdlib/src/fcntl.rs

@@ -8,9 +8,8 @@ mod fcntl {
 
     use crate::vm::{
         PyResult, VirtualMachine,
-        builtins::PyIntRef,
         convert::ToPyException,
-        function::{ArgMemoryBuffer, ArgStrOrBytesLike, Either, OptionalArg},
+        function::{ArgIndex, ArgMemoryBuffer, ArgStrOrBytesLike, Either, OptionalArg},
         stdlib::_io,
     };
     use rustpython_host_env::io::is_interrupted_error;
@@ -86,7 +85,7 @@ mod fcntl {
     #[derive(FromArgs)]
     struct FcntlArg {
         #[pyarg(positional, default = 0)]
-        arg: Either<ArgStrOrBytesLike, PyIntRef>,
+        arg: Either<ArgStrOrBytesLike, ArgIndex>,
     }
 
     #[pyfunction]
@@ -114,7 +113,7 @@ mod fcntl {
                 )?;
                 return Ok(vm.ctx.new_bytes(buf[..arg_len].to_vec()).into());
             }
-            Either::B(i) => i.as_u32_mask(),
+            Either::B(i) => i.as_ref().as_u32_mask(),
         };
         let ret = retry_on_eintr(
             vm,
@@ -213,9 +212,9 @@ mod fcntl {
     #[derive(FromArgs)]
     struct LockfArgs {
         #[pyarg(positional, default = 0)]
-        len: PyIntRef,
+        len: ArgIndex,
         #[pyarg(positional, default = 0)]
-        start: PyIntRef,
+        start: ArgIndex,
         #[pyarg(positional, default)]
         whence: i32,
     }
@@ -229,8 +228,8 @@ mod fcntl {
         LockfArgs { len, start, whence }: LockfArgs,
         vm: &VirtualMachine,
     ) -> PyResult {
-        let start = start.try_to_primitive(vm)?;
-        let len = len.try_to_primitive(vm)?;
+        let start = start.as_ref().try_to_primitive(vm)?;
+        let len = len.as_ref().try_to_primitive(vm)?;
         // F_LOCK and F_TLOCK differ in exactly this: the first one waits.
         let ret = retry_on_eintr(
             vm,

crates/stdlib/src/grp.rs

@@ -5,9 +5,10 @@ pub(crate) use grp::module_def;
 mod grp {
     use crate::vm::{
         PyObjectRef, PyResult, VirtualMachine,
-        builtins::{PyIntRef, PyListRef, PyUtf8StrRef},
+        builtins::{PyListRef, PyUtf8StrRef},
         convert::{IntoPyException, ToPyObject},
         exceptions,
+        function::ArgIndex,
         types::PyStructSequence,
     };
     use core::hint::cold_path;
@@ -44,12 +45,12 @@ mod grp {
     #[derive(FromArgs)]
     struct GetGrGidArgs {
         #[pyarg(any)]
-        id: PyIntRef,
+        id: ArgIndex,
     }
 
     #[pyfunction]
     fn getgrgid(id: GetGrGidArgs, vm: &VirtualMachine) -> PyResult<GroupData> {
-        let gid = id.id;
+        let gid = id.id.into_int_ref();
         let gr_gid = gid.as_bigint();
         let gid = host_grp::gid_t::try_from(gr_gid).ok();
         let group = gid

crates/stdlib/src/mmap.rs

@@ -14,7 +14,7 @@ mod mmap {
         builtins::{PyBytes, PyBytesRef, PyInt, PyIntRef, PyType, PyTypeRef},
         byte::{bytes_from_object, value_from_object},
         convert::ToPyException,
-        function::ArgBytesLike,
+        function::{ArgBytesLike, ArgIndex},
         protocol::{
             BufferDescriptor, BufferMethods, PyBuffer, PyMappingMethods, PySequenceMethods,
         },
@@ -330,9 +330,9 @@ mod mmap {
         option: core::ffi::c_int,
         // None means 0.
         #[pyarg(positional, default, py_default = "0")]
-        start: Option<PyIntRef>,
+        start: Option<ArgIndex>,
         #[pyarg(positional, optional)]
-        length: Option<PyIntRef>,
+        length: Option<ArgIndex>,
     }
 
     #[cfg(all(unix, not(target_os = "redox")))]
@@ -345,7 +345,8 @@ mod mmap {
             let start = self
                 .start
                 .map(|s| {
-                    s.try_to_primitive::<usize>(vm)
+                    s.as_ref()
+                        .try_to_primitive::<usize>(vm)
                         .ok()
                         .filter(|s| *s < len)
                         .ok_or_else(|| vm.new_value_error("madvise start out of bounds"))
@@ -355,7 +356,8 @@ mod mmap {
             let length = self
                 .length
                 .map(|s| {
-                    s.try_to_primitive::<usize>(vm)
+                    s.as_ref()
+                        .try_to_primitive::<usize>(vm)
                         .map_err(|_| vm.new_value_error("madvise length invalid"))
                 })
                 .transpose()?
@@ -920,9 +922,9 @@ mod mmap {
         #[pymethod(name = "move")]
         fn move_(
             zelf: &Py<Self>,
-            dest: PyIntRef,
-            src: PyIntRef,
-            count: PyIntRef,
+            dest: ArgIndex,
+            src: ArgIndex,
+            count: ArgIndex,
             vm: &VirtualMachine,
         ) -> PyResult<()> {
             fn args(
@@ -948,7 +950,7 @@ mod mmap {
             }
 
             let size = zelf.__len__();
-            let (dest, src, cnt) = args(&dest, &src, &count, size, vm)
+            let (dest, src, cnt) = args(dest.as_ref(), src.as_ref(), count.as_ref(), size, vm)
                 .ok_or_else(|| vm.new_value_error("source, destination, or count out of range"))?;
 
             let dest_end = dest + cnt;

crates/stdlib/src/resource.rs

@@ -6,8 +6,9 @@ pub(crate) use resource::module_def;
 mod resource {
     use crate::vm::{
         Py, PyObject, PyObjectRef, PyResult, TryFromBorrowedObject, VirtualMachine,
-        builtins::{PyInt, PyIntRef},
+        builtins::PyInt,
         convert::{ToPyException, ToPyObject},
+        function::ArgIndex,
         types::PyStructSequence,
     };
     use num_traits::Signed;
@@ -185,8 +186,8 @@ mod resource {
     }
 
     #[pyfunction]
-    fn getrlimit(resource: PyIntRef, vm: &VirtualMachine) -> PyResult<Limits> {
-        let resource = py2rlim(&resource, vm)?;
+    fn getrlimit(resource: ArgIndex, vm: &VirtualMachine) -> PyResult<Limits> {
+        let resource = py2rlim(resource.as_ref(), vm)?;
 
         if resource >= RLIM_NLIMITS as host_resource::rlim_t {
             return Err(vm.new_value_error("invalid resource specified"));
@@ -197,8 +198,8 @@ mod resource {
     }
 
     #[pyfunction]
-    fn setrlimit(resource: PyIntRef, limits: Limits, vm: &VirtualMachine) -> PyResult<()> {
-        let resource = py2rlim(&resource, vm)?;
+    fn setrlimit(resource: ArgIndex, limits: Limits, vm: &VirtualMachine) -> PyResult<()> {
+        let resource = py2rlim(resource.as_ref(), vm)?;
 
         if resource >= RLIM_NLIMITS as host_resource::rlim_t {
             return Err(vm.new_value_error("invalid resource specified"));

crates/vm/src/anystr.rs

@@ -5,9 +5,9 @@ use rustpython_unicode::case;
 
 use crate::{
     AsObject, Py, PyObject, PyObjectRef, PyResult, TryFromObject, VirtualMachine,
-    builtins::{PyInt, PyIntRef, PyTuple},
+    builtins::{PyInt, PyTuple},
     convert::TryFromBorrowedObject,
-    function::PySsize,
+    function::{ArgIndex, PySsize},
 };
 
 #[derive(FromArgs)]
@@ -41,17 +41,17 @@ pub(crate) struct StartsEndsWithArgs {
     #[pyarg(positional)]
     affix: PyObjectRef,
     #[pyarg(positional, default)]
-    start: Option<PyIntRef>,
+    start: Option<ArgIndex>,
     #[pyarg(positional, default)]
-    end: Option<PyIntRef>,
+    end: Option<ArgIndex>,
 }
 
 impl StartsEndsWithArgs {
     pub(crate) fn get_value(self, len: usize) -> (PyObjectRef, Option<Range<usize>>) {
         let range = if self.start.is_some() || self.end.is_some() {
             Some(adjust_indices(
-                self.start.as_deref(),
-                self.end.as_deref(),
+                self.start.as_ref().map(AsRef::as_ref),
+                self.end.as_ref().map(AsRef::as_ref),
                 len,
             ))
         } else {

crates/vm/src/builtins/enumerate.rs

@@ -1,11 +1,12 @@
 use super::{
-    IterStatus, PositionIterInternal, PyGenericAlias, PyIntRef, PyTupleRef, PyType, PyTypeRef,
+    IterStatus, PositionIterInternal, PyGenericAlias, PyTupleRef, PyType, PyTypeRef,
     iter::builtins_reversed, locked_rev_next,
 };
 use crate::common::lock::{PyMutex, PyRwLock};
 use crate::{
     AsObject, Context, Py, PyObjectRef, PyPayload, PyResult, VirtualMachine,
     class::PyClassImpl,
+    function::ArgIndex,
     protocol::{PyIter, PyIterReturn},
     raise_if_stop,
     types::{Constructor, IterNext, Iterable, SelfIter},
@@ -56,7 +57,7 @@ pub struct EnumerateArgs {
     #[pyarg(any)]
     iterable: PyIter,
     #[pyarg(any, default = 0)]
-    start: PyIntRef,
+    start: ArgIndex,
 }
 
 impl Constructor for PyEnumerate {
@@ -67,6 +68,7 @@ impl Constructor for PyEnumerate {
         Self::Args { iterable, start }: Self::Args,
         _vm: &VirtualMachine,
     ) -> PyResult<Self> {
+        let start = start.into_int_ref();
         let counter = match start.as_bigint().to_usize() {
             Some(n) => Counter::Small(n),
             None => Counter::Big(start.as_bigint().clone()),

crates/vm/src/builtins/str.rs

@@ -1,6 +1,6 @@
 use super::{
     PositionIterInternal, PyBytesRef, PyDict, PyList, PyTuple, PyTupleRef, PyType, PyTypeRef,
-    int::{PyInt, PyIntRef},
+    int::PyInt,
     iter::{IterStatus, builtins_iter},
 };
 use crate::{
@@ -17,7 +17,7 @@ use crate::{
     },
     convert::{ToPyException, ToPyObject, ToPyResult},
     format::{format, format_map},
-    function::{ArgIterable, FuncArgs, OptionalArg, PyComparisonValue, PySsize},
+    function::{ArgIndex, ArgIterable, FuncArgs, OptionalArg, PyComparisonValue, PySsize},
     intern::PyInterned,
     object::{MaybeTraverse, Traverse, TraverseFn},
     protocol::{
@@ -1871,14 +1871,18 @@ pub(crate) struct FindArgs {
     #[pyarg(positional)]
     sub: PyStrRef,
     #[pyarg(positional, default)]
-    start: Option<PyIntRef>,
+    start: Option<ArgIndex>,
     #[pyarg(positional, default)]
-    end: Option<PyIntRef>,
+    end: Option<ArgIndex>,
 }
 
 impl FindArgs {
     fn get_value(self, len: usize) -> (PyStrRef, core::ops::Range<usize>) {
-        let range = adjust_indices(self.start.as_deref(), self.end.as_deref(), len);
+        let range = adjust_indices(
+            self.start.as_ref().map(AsRef::as_ref),
+            self.end.as_ref().map(AsRef::as_ref),
+            len,
+        );
         (self.sub, range)
     }
 }

crates/vm/src/bytes_inner.rs

@@ -10,7 +10,7 @@ use crate::{
     cformat::cformat_bytes,
     common::wtf8::is_py_ascii_whitespace,
     common::{borrow::BorrowedValue, hash},
-    function::{ArgIterable, Either, OptionalArg, PyComparisonValue},
+    function::{ArgIndex, ArgIterable, Either, OptionalArg, PyComparisonValue},
     literal::escape::Escape,
     protocol::{BufferFlags, PyBuffer},
     sequence::{SequenceExt, SequenceMutExt},
@@ -315,9 +315,9 @@ pub struct ByteInnerFindOptions {
     #[pyarg(positional)]
     sub: ByteInnerSub,
     #[pyarg(positional, default)]
-    start: Option<PyIntRef>,
+    start: Option<ArgIndex>,
     #[pyarg(positional, default)]
-    end: Option<PyIntRef>,
+    end: Option<ArgIndex>,
 }
 
 impl ByteInnerFindOptions {
@@ -327,7 +327,11 @@ impl ByteInnerFindOptions {
         vm: &VirtualMachine,
     ) -> PyResult<(Vec<u8>, core::ops::Range<usize>)> {
         let sub = self.sub.into_vec(vm)?;
-        let range = anystr::adjust_indices(self.start.as_deref(), self.end.as_deref(), len);
+        let range = anystr::adjust_indices(
+            self.start.as_ref().map(AsRef::as_ref),
+            self.end.as_ref().map(AsRef::as_ref),
+            len,
+        );
         Ok((sub, range))
     }
 }

crates/vm/src/stdlib/_thread.rs

@@ -27,8 +27,8 @@ pub(crate) mod _thread {
         convert::ToPyException,
         frame::FrameObjectRef,
         function::{
-            ArgCallable, ArgumentError, DefaultRepr, FromArgs, FuncArgs, KwArgs, NameExcInfo,
-            OptionalArg, Param, ParamKind, PosArgs, PySetterValue, TimeoutSeconds,
+            ArgCallable, ArgIndex, ArgumentError, DefaultRepr, FromArgs, FuncArgs, KwArgs,
+            NameExcInfo, OptionalArg, Param, ParamKind, PosArgs, PySetterValue, TimeoutSeconds,
         },
         object::{Traverse, TraverseFn},
         types::{Constructor, GetAttr, PyStructSequence, Representable, SetAttr},
@@ -847,7 +847,7 @@ pub(crate) mod _thread {
     #[derive(FromArgs)]
     struct StackSizeArgs {
         #[pyarg(positional, default = 0)]
-        size: PyIntRef,
+        size: ArgIndex,
     }
 
     #[pyfunction]
@@ -855,7 +855,7 @@ pub(crate) mod _thread {
         const MIN_SIZE: usize = PY_OS_MIN_STACK_SIZE + SYSTEM_PAGE_SIZE;
         let StackSizeArgs { size } = args;
 
-        let Ok(size) = size.try_to_primitive(vm) else {
+        let Ok(size) = size.as_ref().try_to_primitive(vm) else {
             return Err(vm.new_value_error(format!("size must be at least {MIN_SIZE} bytes")));
         };
 

crates/vm/src/stdlib/itertools.rs

@@ -5,13 +5,11 @@ mod decl {
     use crate::{
         AsObject, Py, PyObject, PyObjectRef, PyPayload, PyRef, PyResult, PyWeakRef, TryFromObject,
         VirtualMachine,
-        builtins::{
-            PyGenericAlias, PyInt, PyIntRef, PyList, PyTuple, PyTupleRef, PyType, PyTypeRef, int,
-        },
+        builtins::{PyGenericAlias, PyInt, PyList, PyTuple, PyTupleRef, PyType, PyTypeRef, int},
         class::PyClassDef,
         common::lock::{PyMutex, PyRwLock, PyRwLockWriteGuard},
         convert::ToPyObject,
-        function::{FuncArgs, NameIterables, PosArgs},
+        function::{ArgIndex, FuncArgs, NameIterables, PosArgs},
         protocol::{PyIter, PyIterReturn, PyNumber},
         raise_if_stop,
         stdlib::sys,
@@ -1329,7 +1327,7 @@ mod decl {
         #[pyarg(any)]
         iterable: PyObjectRef,
         #[pyarg(any)]
-        r: PyIntRef,
+        r: ArgIndex,
     }
 
     impl Constructor for PyItertoolsCombinations {
@@ -1342,6 +1340,7 @@ mod decl {
         ) -> PyResult<Self> {
             let pool: Vec<_> = iterable.try_to_value(vm)?;
 
+            let r = r.into_int_ref();
             let r = r.as_bigint();
             if r.is_negative() {
                 return Err(vm.new_value_error("r must be non-negative"));
@@ -1454,6 +1453,7 @@ mod decl {
             vm: &VirtualMachine,
         ) -> PyResult<Self> {
             let pool: Vec<_> = iterable.try_to_value(vm)?;
+            let r = r.into_int_ref();
             let r = r.as_bigint();
             if r.is_negative() {
                 return Err(vm.new_value_error("r must be non-negative"));
@@ -1800,7 +1800,7 @@ mod decl {
         #[pyarg(any)]
         iterable: PyObjectRef,
         #[pyarg(any)]
-        n: PyIntRef,
+        n: ArgIndex,
         #[pyarg(named, default)]
         strict: bool,
     }
@@ -1817,6 +1817,7 @@ mod decl {
             }: Self::Args,
             vm: &VirtualMachine,
         ) -> PyResult<Self> {
+            let n = n.into_int_ref();
             let n = n.as_bigint();
             if n.lt(&BigInt::one()) {
                 return Err(vm.new_value_error("n must be at least one"));

crates/vm/src/stdlib/pwd.rs

@@ -5,11 +5,8 @@ pub(crate) use pwd::module_def;
 #[pymodule]
 mod pwd {
     use crate::{
-        PyResult, VirtualMachine,
-        builtins::{PyIntRef, PyUtf8StrRef},
-        convert::IntoPyException,
-        exceptions,
-        types::PyStructSequence,
+        PyResult, VirtualMachine, builtins::PyUtf8StrRef, convert::IntoPyException, exceptions,
+        function::ArgIndex, types::PyStructSequence,
     };
     use core::hint::cold_path;
     use rustpython_host_env::pwd as host_pwd;
@@ -68,7 +65,8 @@ mod pwd {
     }
 
     #[pyfunction]
-    fn getpwuid(uidobj: PyIntRef, vm: &VirtualMachine) -> PyResult<PasswdData> {
+    fn getpwuid(uidobj: ArgIndex, vm: &VirtualMachine) -> PyResult<PasswdData> {
+        let uidobj = uidobj.into_int_ref();
         let uid_t = libc::uid_t::try_from(uidobj.as_bigint()).ok();
         let user = uid_t
             .map(host_pwd::getpwuid)

extra_tests/snippets/protocol_index_argument.py

@@ -312,3 +312,65 @@ class HintIndex:
 
 with assert_raises(TypeError):
     operator.length_hint(HintIndex())
+
+
+# Arguments bound as `int` objects accept `__index__` as well.
+assert "abcabc".find("a", Index(1)) == 3
+assert "abcabc".rfind("a", 0, Index(3)) == 0
+assert "abcabc".index("c", Index(3)) == 5
+assert "abcabc".rindex("a", Index(1), Index(4)) == 3
+assert "aaaaaa".count("a", Index(2), Index(-1)) == 3
+assert "abcdef".startswith("d", Index(3))
+assert "abcdef".endswith("c", 0, Index(3))
+assert b"abcabc".find(b"a", Index(1)) == 3
+assert b"abcabc".count(b"b", Index(2)) == 1
+assert b"abcdef".startswith(b"d", Index(3))
+assert bytearray(b"abcabc").rindex(b"a", 0, Index(3)) == 0
+assert bytearray(b"abcdef").endswith(b"c", 0, Index(3))
+
+assert list(enumerate("ab", Index(3))) == [(3, "a"), (4, "b")]
+assert list(enumerate("ab", start=Index(3))) == [(3, "a"), (4, "b")]
+
+import itertools
+
+assert list(itertools.combinations("abc", Index(2))) == [
+    ("a", "b"),
+    ("a", "c"),
+    ("b", "c"),
+]
+assert list(itertools.combinations_with_replacement("ab", Index(2))) == [
+    ("a", "a"),
+    ("a", "b"),
+    ("b", "b"),
+]
+assert list(itertools.batched("abcde", Index(2))) == [("a", "b"), ("c", "d"), ("e",)]
+
+assert _thread.stack_size(Index(0)) == 0
+
+import mmap
+
+mapped = mmap.mmap(-1, 6)
+mapped.write(b"abcdef")
+mapped.move(Index(0), Index(3), Index(3))
+assert mapped[:] == b"defdef"
+if hasattr(mapped, "madvise"):
+    mapped.madvise(mmap.MADV_NORMAL, Index(0), Index(6))
+
+if sys.platform != "win32":
+    import fcntl
+    import grp
+    import pwd
+    import resource
+    import tempfile
+
+    assert pwd.getpwuid(Index(0)).pw_uid == 0
+    assert grp.getgrgid(Index(0)).gr_gid == 0
+
+    nofile = resource.getrlimit(resource.RLIMIT_NOFILE)
+    assert resource.getrlimit(Index(resource.RLIMIT_NOFILE)) == nofile
+    resource.setrlimit(Index(resource.RLIMIT_NOFILE), nofile)
+
+    with tempfile.TemporaryFile() as f:
+        assert fcntl.fcntl(f, fcntl.F_SETFD, Index(fcntl.FD_CLOEXEC)) == 0
+        assert fcntl.fcntl(f, fcntl.F_GETFD) == fcntl.FD_CLOEXEC
+        fcntl.lockf(f, fcntl.LOCK_SH | fcntl.LOCK_NB, Index(0), Index(0))