Merge branch 'master' into bit-length

Author: BobTheBuidler

mypy/build.py

@@ -40,6 +40,7 @@
 from typing_extensions import TypeAlias as _TypeAlias
 
 import mypy.semanal_main
+from mypy.cache import Buffer
 from mypy.checker import TypeChecker
 from mypy.error_formatter import OUTPUT_CHOICES, ErrorFormatter
 from mypy.errors import CompileError, ErrorInfo, Errors, report_internal_error
@@ -116,6 +117,8 @@
     "abc",
 }
 
+# We are careful now, we can increase this in future if safe/useful.
+MAX_GC_FREEZE_CYCLES = 1
 
 Graph: _TypeAlias = dict[str, "State"]
 
@@ -707,6 +710,8 @@ def __init__(
         # new file can be processed O(n**2) times. This cache
         # avoids most of this redundant work.
         self.ast_cache: dict[str, tuple[MypyFile, list[ErrorInfo]]] = {}
+        # Number of times we used GC optimization hack for fresh SCCs.
+        self.gc_freeze_cycles = 0
 
     def dump_stats(self) -> None:
         if self.options.dump_build_stats:
@@ -1139,6 +1144,17 @@ def read_deps_cache(manager: BuildManager, graph: Graph) -> dict[str, FgDepMeta]
     return module_deps_metas
 
 
+def _load_ff_file(file: str, manager: BuildManager, log_error: str) -> bytes | None:
+    t0 = time.time()
+    try:
+        data = manager.metastore.read(file)
+    except OSError:
+        manager.log(log_error + file)
+        return None
+    manager.add_stats(metastore_read_time=time.time() - t0)
+    return data
+
+
 def _load_json_file(
     file: str, manager: BuildManager, log_success: str, log_error: str
 ) -> dict[str, Any] | None:
@@ -1259,7 +1275,11 @@ def get_cache_names(id: str, path: str, options: Options) -> tuple[str, str, str
     deps_json = None
     if options.cache_fine_grained:
         deps_json = prefix + ".deps.json"
-    return (prefix + ".meta.json", prefix + ".data.json", deps_json)
+    if options.fixed_format_cache:
+        data_suffix = ".data.ff"
+    else:
+        data_suffix = ".data.json"
+    return (prefix + ".meta.json", prefix + data_suffix, deps_json)
 
 
 def find_cache_meta(id: str, path: str, manager: BuildManager) -> CacheMeta | None:
@@ -1559,8 +1579,13 @@ def write_cache(
         tree.path = path
 
     # Serialize data and analyze interface
-    data = tree.serialize()
-    data_bytes = json_dumps(data, manager.options.debug_cache)
+    if manager.options.fixed_format_cache:
+        data_io = Buffer()
+        tree.write(data_io)
+        data_bytes = data_io.getvalue()
+    else:
+        data = tree.serialize()
+        data_bytes = json_dumps(data, manager.options.debug_cache)
     interface_hash = hash_digest(data_bytes)
 
     plugin_data = manager.plugin.report_config_data(ReportConfigContext(id, path, is_check=False))
@@ -2085,15 +2110,23 @@ def load_tree(self, temporary: bool = False) -> None:
             self.meta is not None
         ), "Internal error: this method must be called only for cached modules"
 
-        data = _load_json_file(
-            self.meta.data_json, self.manager, "Load tree ", "Could not load tree: "
-        )
+        data: bytes | dict[str, Any] | None
+        if self.options.fixed_format_cache:
+            data = _load_ff_file(self.meta.data_json, self.manager, "Could not load tree: ")
+        else:
+            data = _load_json_file(
+                self.meta.data_json, self.manager, "Load tree ", "Could not load tree: "
+            )
         if data is None:
             return
 
         t0 = time.time()
         # TODO: Assert data file wasn't changed.
-        self.tree = MypyFile.deserialize(data)
+        if isinstance(data, bytes):
+            data_io = Buffer(data)
+            self.tree = MypyFile.read(data_io)
+        else:
+            self.tree = MypyFile.deserialize(data)
         t1 = time.time()
         self.manager.add_stats(deserialize_time=t1 - t0)
         if not temporary:
@@ -2481,7 +2514,11 @@ def write_cache(self) -> None:
         ):
             if self.options.debug_serialize:
                 try:
-                    self.tree.serialize()
+                    if self.manager.options.fixed_format_cache:
+                        data = Buffer()
+                        self.tree.write(data)
+                    else:
+                        self.tree.serialize()
                 except Exception:
                     print(f"Error serializing {self.id}", file=self.manager.stdout)
                     raise  # Propagate to display traceback
@@ -3326,8 +3363,29 @@ def process_graph(graph: Graph, manager: BuildManager) -> None:
                 #
                 # TODO: see if it's possible to determine if we need to process only a
                 # _subset_ of the past SCCs instead of having to process them all.
+                if (
+                    platform.python_implementation() == "CPython"
+                    and manager.gc_freeze_cycles < MAX_GC_FREEZE_CYCLES
+                ):
+                    # When deserializing cache we create huge amount of new objects, so even
+                    # with our generous GC thresholds, GC is still doing a lot of pointless
+                    # work searching for garbage. So, we temporarily disable it when
+                    # processing fresh SCCs, and then move all the new objects to the oldest
+                    # generation with the freeze()/unfreeze() trick below. This is arguably
+                    # a hack, but it gives huge performance wins for large third-party
+                    # libraries, like torch.
+                    gc.collect()
+                    gc.disable()
                 for prev_scc in fresh_scc_queue:
                     process_fresh_modules(graph, prev_scc, manager)
+                if (
+                    platform.python_implementation() == "CPython"
+                    and manager.gc_freeze_cycles < MAX_GC_FREEZE_CYCLES
+                ):
+                    manager.gc_freeze_cycles += 1
+                    gc.freeze()
+                    gc.unfreeze()
+                    gc.enable()
                 fresh_scc_queue = []
             size = len(scc)
             if size == 1:

mypy/cache.py

@@ -0,0 +1,153 @@
+from __future__ import annotations
+
+from collections.abc import Sequence
+from typing import TYPE_CHECKING, Final
+
+try:
+    from native_internal import (
+        Buffer as Buffer,
+        read_bool as read_bool,
+        read_float as read_float,
+        read_int as read_int,
+        read_str as read_str,
+        write_bool as write_bool,
+        write_float as write_float,
+        write_int as write_int,
+        write_str as write_str,
+    )
+except ImportError:
+    # TODO: temporary, remove this after we publish mypy-native on PyPI.
+    if not TYPE_CHECKING:
+
+        class Buffer:
+            def __init__(self, source: bytes = b"") -> None:
+                raise NotImplementedError
+
+            def getvalue(self) -> bytes:
+                raise NotImplementedError
+
+        def read_int(data: Buffer) -> int:
+            raise NotImplementedError
+
+        def write_int(data: Buffer, value: int) -> None:
+            raise NotImplementedError
+
+        def read_str(data: Buffer) -> str:
+            raise NotImplementedError
+
+        def write_str(data: Buffer, value: str) -> None:
+            raise NotImplementedError
+
+        def read_bool(data: Buffer) -> bool:
+            raise NotImplementedError
+
+        def write_bool(data: Buffer, value: bool) -> None:
+            raise NotImplementedError
+
+        def read_float(data: Buffer) -> float:
+            raise NotImplementedError
+
+        def write_float(data: Buffer, value: float) -> None:
+            raise NotImplementedError
+
+
+LITERAL_INT: Final = 1
+LITERAL_STR: Final = 2
+LITERAL_BOOL: Final = 3
+LITERAL_FLOAT: Final = 4
+LITERAL_COMPLEX: Final = 5
+LITERAL_NONE: Final = 6
+
+
+def read_literal(data: Buffer, marker: int) -> int | str | bool | float:
+    if marker == LITERAL_INT:
+        return read_int(data)
+    elif marker == LITERAL_STR:
+        return read_str(data)
+    elif marker == LITERAL_BOOL:
+        return read_bool(data)
+    elif marker == LITERAL_FLOAT:
+        return read_float(data)
+    assert False, f"Unknown literal marker {marker}"
+
+
+def write_literal(data: Buffer, value: int | str | bool | float | complex | None) -> None:
+    if isinstance(value, bool):
+        write_int(data, LITERAL_BOOL)
+        write_bool(data, value)
+    elif isinstance(value, int):
+        write_int(data, LITERAL_INT)
+        write_int(data, value)
+    elif isinstance(value, str):
+        write_int(data, LITERAL_STR)
+        write_str(data, value)
+    elif isinstance(value, float):
+        write_int(data, LITERAL_FLOAT)
+        write_float(data, value)
+    elif isinstance(value, complex):
+        write_int(data, LITERAL_COMPLEX)
+        write_float(data, value.real)
+        write_float(data, value.imag)
+    else:
+        write_int(data, LITERAL_NONE)
+
+
+def read_int_opt(data: Buffer) -> int | None:
+    if read_bool(data):
+        return read_int(data)
+    return None
+
+
+def write_int_opt(data: Buffer, value: int | None) -> None:
+    if value is not None:
+        write_bool(data, True)
+        write_int(data, value)
+    else:
+        write_bool(data, False)
+
+
+def read_str_opt(data: Buffer) -> str | None:
+    if read_bool(data):
+        return read_str(data)
+    return None
+
+
+def write_str_opt(data: Buffer, value: str | None) -> None:
+    if value is not None:
+        write_bool(data, True)
+        write_str(data, value)
+    else:
+        write_bool(data, False)
+
+
+def read_int_list(data: Buffer) -> list[int]:
+    size = read_int(data)
+    return [read_int(data) for _ in range(size)]
+
+
+def write_int_list(data: Buffer, value: list[int]) -> None:
+    write_int(data, len(value))
+    for item in value:
+        write_int(data, item)
+
+
+def read_str_list(data: Buffer) -> list[str]:
+    size = read_int(data)
+    return [read_str(data) for _ in range(size)]
+
+
+def write_str_list(data: Buffer, value: Sequence[str]) -> None:
+    write_int(data, len(value))
+    for item in value:
+        write_str(data, item)
+
+
+def read_str_opt_list(data: Buffer) -> list[str | None]:
+    size = read_int(data)
+    return [read_str_opt(data) for _ in range(size)]
+
+
+def write_str_opt_list(data: Buffer, value: list[str | None]) -> None:
+    write_int(data, len(value))
+    for item in value:
+        write_str_opt(data, item)

mypy/checker.py

@@ -171,6 +171,7 @@
 from mypy.typeanal import check_for_explicit_any, has_any_from_unimported_type, make_optional_type
 from mypy.typeops import (
     bind_self,
+    can_have_shared_disjoint_base,
     coerce_to_literal,
     custom_special_method,
     erase_def_to_union_or_bound,
@@ -2658,6 +2659,8 @@ def visit_class_def(self, defn: ClassDef) -> None:
         for base in typ.mro[1:]:
             if base.is_final:
                 self.fail(message_registry.CANNOT_INHERIT_FROM_FINAL.format(base.name), defn)
+        if not can_have_shared_disjoint_base(typ.bases):
+            self.fail(message_registry.INCOMPATIBLE_DISJOINT_BASES.format(typ.name), defn)
         with self.tscope.class_scope(defn.info), self.enter_partial_types(is_class=True):
             old_binder = self.binder
             self.binder = ConditionalTypeBinder(self.options)
@@ -5826,6 +5829,10 @@ def _make_fake_typeinfo_and_full_name(
             format_type_distinctly(*base_classes, options=self.options, bare=True), "and"
         )
 
+        if not can_have_shared_disjoint_base(base_classes):
+            errors.append((pretty_names_list, "have distinct disjoint bases"))
+            return None
+
         new_errors = []
         for base in base_classes:
             if base.type.is_final:

mypy/checkexpr.py

@@ -582,7 +582,10 @@ def visit_call_expr_inner(self, e: CallExpr, allow_none_return: bool = False) ->
                     and not node.node.no_args
                     and not (
                         isinstance(union_target := get_proper_type(node.node.target), UnionType)
-                        and union_target.uses_pep604_syntax
+                        and (
+                            union_target.uses_pep604_syntax
+                            or self.chk.options.python_version >= (3, 10)
+                        )
                     )
                 ):
                     self.msg.type_arguments_not_allowed(e)

mypy/checkmember.py

@@ -1484,13 +1484,7 @@ def analyze_decorator_or_funcbase_access(
     if isinstance(defn, Decorator):
         return analyze_var(name, defn.var, itype, mx)
     typ = function_type(defn, mx.chk.named_type("builtins.function"))
-    is_trivial_self = False
-    if isinstance(defn, Decorator):
-        # Use fast path if there are trivial decorators like @classmethod or @property
-        is_trivial_self = defn.func.is_trivial_self and not defn.decorators
-    elif isinstance(defn, (FuncDef, OverloadedFuncDef)):
-        is_trivial_self = defn.is_trivial_self
-    if is_trivial_self:
+    if isinstance(defn, (FuncDef, OverloadedFuncDef)) and defn.is_trivial_self:
         return bind_self_fast(typ, mx.self_type)
     typ = check_self_arg(typ, mx.self_type, defn.is_class, mx.context, name, mx.msg)
     return bind_self(typ, original_type=mx.self_type, is_classmethod=defn.is_class)

mypy/checkpattern.py

@@ -192,7 +192,7 @@ def visit_or_pattern(self, o: OrPattern) -> PatternType:
         for capture_list in capture_types.values():
             typ = UninhabitedType()
             for _, other in capture_list:
-                typ = join_types(typ, other)
+                typ = make_simplified_union([typ, other])
 
             captures[capture_list[0][0]] = typ
 
@@ -539,12 +539,12 @@ def visit_class_pattern(self, o: ClassPattern) -> PatternType:
         #
         type_info = o.class_ref.node
         if type_info is None:
-            return PatternType(AnyType(TypeOfAny.from_error), AnyType(TypeOfAny.from_error), {})
-        if isinstance(type_info, TypeAlias) and not type_info.no_args:
+            typ: Type = AnyType(TypeOfAny.from_error)
+        elif isinstance(type_info, TypeAlias) and not type_info.no_args:
             self.msg.fail(message_registry.CLASS_PATTERN_GENERIC_TYPE_ALIAS, o)
             return self.early_non_match()
-        if isinstance(type_info, TypeInfo):
-            typ: Type = fill_typevars_with_any(type_info)
+        elif isinstance(type_info, TypeInfo):
+            typ = fill_typevars_with_any(type_info)
         elif isinstance(type_info, TypeAlias):
             typ = type_info.target
         elif (