mirror of
https://github.com/astral-sh/ruff.git
synced 2025-10-25 01:17:34 +00:00
3a77768f79
## Summary * Completely removes the concept of visibility constraints. Reachability constraints are now used to model the static visibility of bindings and declarations. Reachability constraints are *much* easier to reason about / work with, since they are applied at the beginning of a branch, and not applied retroactively. Removing the duplication between visibility and reachability constraints also leads to major code simplifications [^1]. For an overview of how the new constraint system works, see the updated doc comment in `reachability_constraints.rs`. * Fixes a [control-flow modeling bug (panic)](https://github.com/astral-sh/ty/issues/365) involving `break` statements in loops * Fixes a [bug where](https://github.com/astral-sh/ty/issues/624) where `elif` branches would have wrong reachability constraints * Fixes a [bug where](https://github.com/astral-sh/ty/issues/648) code after infinite loops would not be considered unreachble * Fixes a panic on the `pywin32` ecosystem project, which we should be able to move to `good.txt` once this has been merged. * Removes some false positives in unreachable code because we infer `Never` more often, due to the fact that reachability constraints now apply retroactively to *all* active bindings, not just to bindings inside a branch. * As one example, this removes the `division-by-zero` diagnostic from https://github.com/astral-sh/ty/issues/443 because we now infer `Never` for the divisor. * Supersedes and includes similar test changes as https://github.com/astral-sh/ruff/pull/18392 closes https://github.com/astral-sh/ty/issues/365 closes https://github.com/astral-sh/ty/issues/624 closes https://github.com/astral-sh/ty/issues/642 closes https://github.com/astral-sh/ty/issues/648 ## Benchmarks Benchmarks on black, pandas, and sympy showed that this is neither a performance improvement, nor a regression. ## Test Plan Regression tests for: - [x] https://github.com/astral-sh/ty/issues/365 - [x] https://github.com/astral-sh/ty/issues/624 - [x] https://github.com/astral-sh/ty/issues/642 - [x] https://github.com/astral-sh/ty/issues/648 [^1]: I'm afraid this is something that @carljm advocated for since the beginning, and I'm not sure anymore why we have never seriously tried this before. So I suggest we do *not* attempt to do a historical deep dive to find out exactly why this ever became so complicated, and just enjoy the fact that we eventually arrived here. --------- Co-authored-by: Carl Meyer <carl@astral.sh>
5.5 KiB
5.5 KiB
Narrowing for conditionals with boolean expressions
Narrowing in and conditional
class A: ...
class B: ...
def _(x: A | B):
if isinstance(x, A) and isinstance(x, B):
reveal_type(x) # revealed: A & B
else:
reveal_type(x) # revealed: (B & ~A) | (A & ~B)
Arms might not add narrowing constraints
class A: ...
class B: ...
def _(flag: bool, x: A | B):
if isinstance(x, A) and flag:
reveal_type(x) # revealed: A
else:
reveal_type(x) # revealed: A | B
if flag and isinstance(x, A):
reveal_type(x) # revealed: A
else:
reveal_type(x) # revealed: A | B
reveal_type(x) # revealed: A | B
Statically known arms
class A: ...
class B: ...
def _(x: A | B):
if isinstance(x, A) and True:
reveal_type(x) # revealed: A
else:
reveal_type(x) # revealed: B & ~A
if True and isinstance(x, A):
reveal_type(x) # revealed: A
else:
reveal_type(x) # revealed: B & ~A
if False and isinstance(x, A):
# TODO: should emit an `unreachable code` diagnostic
reveal_type(x) # revealed: Never
else:
reveal_type(x) # revealed: A | B
if False or isinstance(x, A):
reveal_type(x) # revealed: A
else:
reveal_type(x) # revealed: B & ~A
if True or isinstance(x, A):
reveal_type(x) # revealed: A | B
else:
# TODO: should emit an `unreachable code` diagnostic
reveal_type(x) # revealed: Never
reveal_type(x) # revealed: A | B
The type of multiple symbols can be narrowed down
class A: ...
class B: ...
def _(x: A | B, y: A | B):
if isinstance(x, A) and isinstance(y, B):
reveal_type(x) # revealed: A
reveal_type(y) # revealed: B
else:
# No narrowing: Only-one or both checks might have failed
reveal_type(x) # revealed: A | B
reveal_type(y) # revealed: A | B
reveal_type(x) # revealed: A | B
reveal_type(y) # revealed: A | B
Narrowing in or conditional
class A: ...
class B: ...
class C: ...
def _(x: A | B | C):
if isinstance(x, A) or isinstance(x, B):
reveal_type(x) # revealed: A | B
else:
reveal_type(x) # revealed: C & ~A & ~B
In or, all arms should add constraint in order to narrow
class A: ...
class B: ...
class C: ...
def _(flag: bool, x: A | B | C):
if isinstance(x, A) or isinstance(x, B) or flag:
reveal_type(x) # revealed: A | B | C
else:
reveal_type(x) # revealed: C & ~A & ~B
in or, all arms should narrow the same set of symbols
class A: ...
class B: ...
class C: ...
def _(x: A | B | C, y: A | B | C):
if isinstance(x, A) or isinstance(y, A):
# The predicate might be satisfied by the right side, so the type of `x` can’t be narrowed down here.
reveal_type(x) # revealed: A | B | C
# The same for `y`
reveal_type(y) # revealed: A | B | C
else:
reveal_type(x) # revealed: (B & ~A) | (C & ~A)
reveal_type(y) # revealed: (B & ~A) | (C & ~A)
if (isinstance(x, A) and isinstance(y, A)) or (isinstance(x, B) and isinstance(y, B)):
# Here, types of `x` and `y` can be narrowd since all `or` arms constraint them.
reveal_type(x) # revealed: A | B
reveal_type(y) # revealed: A | B
else:
reveal_type(x) # revealed: A | B | C
reveal_type(y) # revealed: A | B | C
mixing and and not
class A: ...
class B: ...
class C: ...
def _(x: A | B | C):
if isinstance(x, B) and not isinstance(x, C):
reveal_type(x) # revealed: B & ~C
else:
# ~(B & ~C) -> ~B | C -> (A & ~B) | (C & ~B) | C -> (A & ~B) | C
reveal_type(x) # revealed: (A & ~B) | C
mixing or and not
class A: ...
class B: ...
class C: ...
def _(x: A | B | C):
if isinstance(x, B) or not isinstance(x, C):
reveal_type(x) # revealed: B | (A & ~C)
else:
reveal_type(x) # revealed: C & ~B
or with nested and
class A: ...
class B: ...
class C: ...
def _(x: A | B | C):
if isinstance(x, A) or (isinstance(x, B) and not isinstance(x, C)):
reveal_type(x) # revealed: A | (B & ~C)
else:
# ~(A | (B & ~C)) -> ~A & ~(B & ~C) -> ~A & (~B | C) -> (~A & C) | (~A ~ B)
reveal_type(x) # revealed: C & ~A
and with nested or
class A: ...
class B: ...
class C: ...
def _(x: A | B | C):
if isinstance(x, A) and (isinstance(x, B) or not isinstance(x, C)):
# A & (B | ~C) -> (A & B) | (A & ~C)
reveal_type(x) # revealed: (A & B) | (A & ~C)
else:
# ~((A & B) | (A & ~C)) ->
# ~(A & B) & ~(A & ~C) ->
# (~A | ~B) & (~A | C) ->
# [(~A | ~B) & ~A] | [(~A | ~B) & C] ->
# ~A | (~A & C) | (~B & C) ->
# ~A | (C & ~B) ->
# ~A | (C & ~B) The positive side of ~A is A | B | C ->
reveal_type(x) # revealed: (B & ~A) | (C & ~A) | (C & ~B)
Boolean expression internal narrowing
def _(x: str | None, y: str | None):
if x is None and y is not x:
reveal_type(y) # revealed: str
# Neither of the conditions alone is sufficient for narrowing y's type:
if x is None:
reveal_type(y) # revealed: str | None
if y is not x:
reveal_type(y) # revealed: str | None
Assignment expressions
def f() -> bool:
return True
if x := f():
reveal_type(x) # revealed: Literal[True]
else:
reveal_type(x) # revealed: Literal[False]