mirror of
https://github.com/astral-sh/ruff.git
synced 2025-10-27 02:16:54 +00:00
2a00eca66b
## Summary
Implements proper reachability analysis and — in effect — exhaustiveness
checking for `match` statements. This allows us to check the following
code without any errors (leads to *"can implicitly return `None`"* on
`main`):
```py
from enum import Enum, auto
class Color(Enum):
RED = auto()
GREEN = auto()
BLUE = auto()
def hex(color: Color) -> str:
match color:
case Color.RED:
return "#ff0000"
case Color.GREEN:
return "#00ff00"
case Color.BLUE:
return "#0000ff"
```
Note that code like this already worked fine if there was a
`assert_never(color)` statement in a catch-all case, because we would
then consider that `assert_never` call terminal. But now this also works
without the wildcard case. Adding a member to the enum would still lead
to an error here, if that case would not be handled in `hex`.
What needed to happen to support this is a new way of evaluating match
pattern constraints. Previously, we would simply compare the type of the
subject expression against the patterns. For the last case here, the
subject type would still be `Color` and the value type would be
`Literal[Color.BLUE]`, so we would infer an ambiguous truthiness.
Now, before we compare the subject type against the pattern, we first
generate a union type that corresponds to the set of all values that
would have *definitely been matched* by previous patterns. Then, we
build a "narrowed" subject type by computing `subject_type &
~already_matched_type`, and compare *that* against the pattern type. For
the example here, `already_matched_type = Literal[Color.RED] |
Literal[Color.GREEN]`, and so we have a narrowed subject type of `Color
& ~(Literal[Color.RED] | Literal[Color.GREEN]) = Literal[Color.BLUE]`,
which allows us to infer a reachability of `AlwaysTrue`.
<details>
<summary>A note on negated reachability constraints</summary>
It might seem that we now perform duplicate work, because we also record
*negated* reachability constraints. But that is still important for
cases like the following (and possibly also for more realistic
scenarios):
```py
from typing import Literal
def _(x: int | str):
match x:
case None:
pass # never reachable
case _:
y = 1
y
```
</details>
closes https://github.com/astral-sh/ty/issues/99
## Test Plan
* I verified that this solves all examples from the linked ticket (the
first example needs a PEP 695 type alias, because we don't support
legacy type aliases yet)
* Verified that the ecosystem changes are all because of removed false
positives
* Updated tests
6.4 KiB
6.4 KiB
Pattern matching
[environment]
python-version = "3.10"
With wildcard
def _(target: int):
match target:
case 1:
y = 2
case _:
y = 3
reveal_type(y) # revealed: Literal[2, 3]
Without wildcard
def _(target: int):
match target:
case 1:
y = 2
case 2:
y = 3
# revealed: Literal[2, 3]
# error: [possibly-unresolved-reference]
reveal_type(y)
Basic match
def _(target: int):
y = 1
y = 2
match target:
case 1:
y = 3
case 2:
y = 4
reveal_type(y) # revealed: Literal[2, 3, 4]
Value match
A value pattern matches based on equality: the first case branch here will be taken if subject
is equal to 2, even if subject is not an instance of int. We can't know whether C here has a
custom __eq__ implementation that might cause it to compare equal to 2, so we have to consider
the possibility that the case branch might be taken even though the type C is disjoint from the
type Literal[2].
This leads us to infer Literal[1, 3] as the type of y after the match statement, rather than
Literal[1]:
from typing import final
@final
class C:
pass
def _(subject: C):
y = 1
match subject:
case 2:
y = 3
reveal_type(y) # revealed: Literal[1, 3]
Class match
A case branch with a class pattern is taken if the subject is an instance of the given class, and
all subpatterns in the class pattern match.
Without arguments
from typing import final
class Foo:
pass
class FooSub(Foo):
pass
class Bar:
pass
@final
class Baz:
pass
def _(target: FooSub):
y = 1
match target:
case Baz():
y = 2
case Foo():
y = 3
case Bar():
y = 4
reveal_type(y) # revealed: Literal[3]
def _(target: FooSub):
y = 1
match target:
case Baz():
y = 2
case Bar():
y = 3
case Foo():
y = 4
reveal_type(y) # revealed: Literal[3, 4]
def _(target: FooSub | str):
y = 1
match target:
case Baz():
y = 2
case Foo():
y = 3
case Bar():
y = 4
reveal_type(y) # revealed: Literal[1, 3, 4]
With arguments
from typing_extensions import assert_never
from dataclasses import dataclass
@dataclass
class Point:
x: int
y: int
class Other: ...
def _(target: Point):
y = 1
match target:
case Point(0, 0):
y = 2
case Point(x=0, y=1):
y = 3
case Point(x=1, y=0):
y = 4
reveal_type(y) # revealed: Literal[1, 2, 3, 4]
def _(target: Point):
match target:
case Point(x, y): # irrefutable sub-patterns
pass
case _:
assert_never(target)
def _(target: Point | Other):
match target:
case Point(0, 0):
reveal_type(target) # revealed: Point
case Point(x=0, y=1):
reveal_type(target) # revealed: Point
case Point(x=1, y=0):
reveal_type(target) # revealed: Point
case Other():
reveal_type(target) # revealed: Other
Singleton match
Singleton patterns are matched based on identity, not equality comparisons or isinstance() checks.
from typing import Literal
def _(target: Literal[True, False]):
y = 1
match target:
case True:
y = 2
case False:
y = 3
case None:
y = 4
reveal_type(y) # revealed: Literal[2, 3]
def _(target: bool):
y = 1
match target:
case True:
y = 2
case False:
y = 3
case None:
y = 4
reveal_type(y) # revealed: Literal[2, 3]
def _(target: None):
y = 1
match target:
case True:
y = 2
case False:
y = 3
case None:
y = 4
reveal_type(y) # revealed: Literal[4]
def _(target: None | Literal[True]):
y = 1
match target:
case True:
y = 2
case False:
y = 3
case None:
y = 4
reveal_type(y) # revealed: Literal[2, 4]
# bool is an int subclass
def _(target: int):
y = 1
match target:
case True:
y = 2
case False:
y = 3
case None:
y = 4
reveal_type(y) # revealed: Literal[1, 2, 3]
def _(target: str):
y = 1
match target:
case True:
y = 2
case False:
y = 3
case None:
y = 4
reveal_type(y) # revealed: Literal[1]
Matching on enums
from enum import Enum
class Answer(Enum):
NO = 0
YES = 1
def _(answer: Answer):
y = 0
match answer:
case Answer.YES:
reveal_type(answer) # revealed: Literal[Answer.YES]
y = 1
case Answer.NO:
reveal_type(answer) # revealed: Literal[Answer.NO]
y = 2
reveal_type(y) # revealed: Literal[1, 2]
Or match
A | pattern matches if any of the subpatterns match.
from typing import Literal, final
def _(target: Literal["foo", "baz"]):
y = 1
match target:
case "foo" | "bar":
y = 2
case "baz":
y = 3
reveal_type(y) # revealed: Literal[2, 3]
def _(target: None):
y = 1
match target:
case None | 3:
y = 2
case "foo" | 4 | True:
y = 3
reveal_type(y) # revealed: Literal[2]
@final
class Baz:
pass
def _(target: int | None | float):
y = 1
match target:
case None | 3:
y = 2
case Baz():
y = 3
reveal_type(y) # revealed: Literal[1, 2]
class Foo: ...
def _(target: None | Foo):
y = 1
match target:
case Baz() | True | False:
y = 2
case int():
y = 3
reveal_type(y) # revealed: Literal[1, 3]
Guard with object that implements __bool__ incorrectly
class NotBoolable:
__bool__: int = 3
def _(target: int, flag: NotBoolable):
y = 1
match target:
# error: [unsupported-bool-conversion] "Boolean conversion is unsupported for type `NotBoolable`"
case 1 if flag:
y = 2
case 2:
y = 3
reveal_type(y) # revealed: Literal[1, 2, 3]