Log message:
py-attrs: update to 26.1.0.

## [26.1.0]

### Backwards-incompatible Changes

- Field aliases are now resolved *before* calling `field_transformer`, so \ 
transformers receive fully populated `Attribute` objects with usable `alias` \ 
values instead of `None`.
  The new `Attribute.alias_is_default` flag indicates whether the alias was \ 
auto-generated
  [#1509]

### Changes

- Fix type annotations for `attrs.validators.optional()`, so it no longer \ 
rejects tuples with more than one validator.
  [#1496]
- The `attrs.validators.disabled()` contextmanager can now be nested.
  [#1513]
- Frozen classes can set `on_setattr=attrs.setters.NO_OP` in addition to `None`.
  [#1515]
- It's now possible to pass *attrs* **instances** in addition to *attrs* \ 
**classes** to `attrs.fields()`.
  [#1529]

Log message:
py-attrs: update to 25.4.0.

## [25.4.0]

### Backwards-incompatible Changes

- Class-level `kw_only=True` behavior is now consistent with `dataclasses`.

  Previously, a class that sets `kw_only=True` makes all attributes \ 
keyword-only, including those from base classes.
  If an attribute sets `kw_only=False`, that setting is ignored, and it is still \ 
made keyword-only.

  Now, only the attributes defined in that class that doesn't explicitly set \ 
`kw_only=False` are made keyword-only.

  This shouldn't be a problem for most users, unless you have a pattern like this:

  ```python
  @attrs.define(kw_only=True)
  class Base:
      a: int
      b: int = attrs.field(default=1, kw_only=False)

  @attrs.define
  class Subclass(Base):
      c: int
  ```

  Here, we have a `kw_only=True` *attrs* class (`Base`) with an attribute that \ 
sets `kw_only=False` and has a default (`Base.b`), and then create a subclass \ 
(`Subclass`) with required arguments (`Subclass.c`).
  Previously this would work, since it would make `Base.b` keyword-only, but now \ 
this fails since `Base.b` is positional, and we have a required positional \ 
argument (`Subclass.c`) following another argument with defaults.

### Changes

- Values passed to the `__init__()` method of `attrs` classes are now correctly \ 
passed to `__attrs_pre_init__()` instead of their default values (in cases where \ 
*kw_only* was not specified).
- Added support for Python 3.14 and [PEP 749]
- `attrs.validators.deep_mapping()` now allows to leave out either \ 
*key_validator* xor *value_validator*.
- `attrs.validators.deep_iterator()` and `attrs.validators.deep_mapping()` now \ 
accept lists and tuples for all validators and wrap them into a \ 
`attrs.validators.and_()`.
- Added a new **experimental** way to inspect classes:

  `attrs.inspect(cls)` returns the _effective_ class-wide parameters that were \ 
used by *attrs* to construct the class.

  The returned class is the same data structure that *attrs* uses internally to \ 
decide how to construct the final class.
- Fixed annotations for `attrs.field(converter=...)`.
  Previously, a `tuple` of converters was only accepted if it had exactly one \ 
element.
- The performance of `attrs.asdict()` has been improved by 45–260%.
- The performance of `attrs.astuple()` has been improved by 49–270%.
- The type annotation for `attrs.validators.or_()` now allows for different \ 
types of validators.

  This was only an issue on Pyright.

Log message:
py-attrs: updated to 25.3.0

25.3.0

Changes
- Restore support for generator-based `field_transformer`s.

Log message:
py-attrs: updated to 25.2.0

25.2.0

Changes

- Checking mandatory vs non-mandatory attribute order is now performed after the \ 
field transformer, since the field transformer may change attributes and/or \ 
their order.
- `attrs.make_class()` now allows for Unicode class names.
- Speed up class creation by 30%-50% by compiling methods only once and using a \ 
variety of other techniques.
- The error message if an attribute has both an annotation and a type argument \ 
will now disclose _what_ attribute seems to be the problem.

Log message:
py-attrs: updated to 25.1.0

25.1.0
This release only ensures correct PyPI licensing metadata.

Log message:
py-attrs: updated to 24.3.0

24.3.0

Backwards-incompatible Changes

- Python 3.7 has been dropped.

Changes

- Introduce `attrs.NothingType`, for annotating types consistent with \ 
`attrs.NOTHING`.
- Allow mutating `__suppress_context__` and `__notes__` on frozen exceptions.
- `attrs.converters.optional()` works again when taking \ 
`attrs.converters.pipe()` or another Converter as its argument.
- *attrs* instances now support \ 
[`copy.replace()`](https://docs.python.org/3/library/copy.html#copy.replace).
- `attrs.validators.instance_of()`'s type hints now allow for union types.
  For example: `instance_of(str | int)`

Log message:
py-attrs: updated to 24.2.0

24.2.0

Deprecations

- Given the amount of warnings raised in the broader ecosystem, we've decided to \ 
only soft-deprecate the *hash* argument to `@define` / `@attr.s`.
 Please don't use it in new code, but we don't intend to remove it anymore.

Changes

- `attrs.converters.pipe()` (and its syntactic sugar of passing a list for \ 
`attrs.field()`'s / `attr.ib()`'s *converter* argument) works again when passing \ 
`attrs.setters.convert` to *on_setattr* (which is default for `attrs.define`).
- Restored support for PEP [649](https://peps.python.org/pep-0649/) / \ 
[749](https://peps.python.org/pep-0749/)-implementing Pythons -- currently \ 
3.14-dev.

Log message:
py-attrs: updated to 24.1.0

24.1.0

Backwards-incompatible Changes

- `attrs.evolve()` doesn't accept the *inst* argument as a keyword argument anymore.
  Pass it as the first positional argument instead.
- `attrs.validators.provides()` has been removed.
  The removed code is available as a \ 
[gist](https://gist.github.com/hynek/9eaaaeb659808f3519870dfa16d2b6b2) for \ 
convenient copy and pasting.
- All packaging metadata except from `__version__` and `__version_info__` has \ 
been removed from the `attr` and `attrs` modules (for example, `attrs.__url__`).

  Please use \ 
[`importlib.metadata`](https://docs.python.org/3/library/importlib.metadata.html) \ 
or [*importlib_metadata*](https://pypi.org/project/importlib-metadata/) instead.
- Speed up the generated `__eq__` methods significantly by generating a chain of \ 
attribute comparisons instead of constructing and comparing tuples.
  This change arguably makes the behavior more correct,
  but changes it if an attribute compares equal by identity but not value, like \ 
`float('nan')`.

Deprecations

- The *repr_ns* argument to `attr.s` is now deprecated.
  It was a workaround for nested classes in Python 2 and is pointless in Python 3.
- The *hash* argument to `@attr.s`, `@attrs.define`, and `make_class()` is now \ 
deprecated in favor of *unsafe_hash*, as defined by PEP 681.

Changes

- Allow original slotted `functools.cached_property` classes to be cleaned by \ 
garbage collection.
  Allow `super()` calls in slotted cached properties.
- Our type stubs now use modern type notation and are organized such that VS \ 
Code's quick-fix prefers the `attrs` namespace.
- Preserve `AttributeError` raised by properties of slotted classes with \ 
`functools.cached_properties`.
- It is now possible to wrap a converter into an `attrs.Converter` and get the \ 
current instance and/or the current field definition passed into the converter \ 
callable.

  Note that this is not supported by any type checker, yet.
- `attrs.make_class()` now populates the `__annotations__` dict of the generated \ 
class, so that `attrs.resolve_types()` can resolve them.
- Added the `attrs.validators.or_()` validator.
- The combination of a `__attrs_pre_init__` that takes arguments, a kw-only \ 
field, and a default on that field does not crash anymore.
- `attrs.validators.in_()` now transforms certain unhashable options to tuples \ 
to keep the field hashable.

  This allows fields that use this validator to be used with, for example, \ 
`attrs.filters.include()`.
- If a class has an *inherited* method called `__attrs_init_subclass__`, it is \ 
now called once the class is done assembling.

  This is a replacement for Python's `__init_subclass__` and useful for \ 
registering classes, and similar.