New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

17.3.0 changes the order how inherited class attributes are defined if they have the same base class #285

Closed
Emantor opened this Issue Nov 9, 2017 · 4 comments

Comments

Projects
None yet
4 participants
@Emantor

Emantor commented Nov 9, 2017

Minimal working example:

import attr

@attr.s()
class A:
    name = attr.ib()
    cookie = attr.ib(default="Cookie")

@attr.s()
class B:
    cake = attr.ib(default="Cake")
    cereal = attr.ib(default="Cereal")

@attr.s()
class C(B, A):
    breakfast = attr.ib(default="All")

@attr.s()
class D(A):
    food = attr.ib(default="food")
    melon = attr.ib(default="melon")

@attr.s()
class E(D, C):
    apple = attr.ib(default="apple")

This works with attrs==17.2.0 and breaks with attrs==17.3.0. In 17.3.0 the list of non_super_names in base.py: ['cake', 'cereal', 'breakfast', 'name', 'cookie', 'food', 'melon', 'apple'] breaks because attr is trying to create a non-default attribute in between default attributes.

@Emantor Emantor changed the title from 17.3.0 changes the order how inherited class attributes are defined if the have the same base class to 17.3.0 changes the order how inherited class attributes are defined if they have the same base class Nov 9, 2017

@jluebbe

This comment has been minimized.

jluebbe commented Nov 9, 2017

When looking at what _transform_attrs is doing for class E, we see what's going wrong. The MRO for E is E, D, C, B, A.
It first collects all (not just locally defined) attributes from D.__attrs_attrs__ in reverse order: melon, food, cookie, name. Then it looks at C.__attrs_attrs__ and collects the attributes which are not yet defined: breakfast, cereal, cake.
When reversing this again, cake, cerealand breakfast end up before the name attribute.
It seems that this should work if it only looked at the locally defined attributes when iterating over the MRO, as A.name would be processed last.

@hynek hynek added this to the 17.4 milestone Nov 9, 2017

@jluebbe

This comment has been minimized.

jluebbe commented Nov 9, 2017

As a minimal workaround, this change fixes the order for our usecase:

@@ -278,7 +278,9 @@
             # list in the end and get all attributes in the order they have
             # been defined.
             for a in reversed(sub_attrs):
-                if a.name not in taken_attr_names:
+                if a.name in super_cls.__super_names__:
+                    continue
+                elif a.name not in taken_attr_names:
                     super_attrs.append(a)
                     taken_attr_names.add(a.name)
 
@@ -354,6 +356,7 @@
         self._has_post_init = bool(getattr(cls, "__attrs_post_init__", False))
 
         self._cls_dict["__attrs_attrs__"] = self._attrs
+        self._cls_dict["__super_names__"] = self._super_names
 
         if frozen:
             self._cls_dict["__setattr__"] = _frozen_setattrs

The downside is obviously cluttering the generated classes with the __super_names__ attribute. Perhaps each Attribute could keep a reference to the class where it was defined.
That would allow _transform_attrs to filter out those defined on other super classes and only handle them later where they were actually defined.

hynek added a commit that referenced this issue Nov 9, 2017

@hynek

This comment has been minimized.

Member

hynek commented Nov 9, 2017

I think it’s simpler. The problem is that it’s possible to overwrite attribute definitions now and that’s exactly what’s happening here. The bug is that I didn’t take into account such hierarchies that introduce the same attribute multiple times.

#287 should fix that, would you mind checking it out?

@jluebbe

This comment has been minimized.

jluebbe commented Nov 9, 2017

Yes, #287 fixes our test suite. Thanks!

@hynek hynek closed this in #287 Nov 11, 2017

bors-fusion bot added a commit to fusionapp/documint that referenced this issue Jan 1, 2018

Merge #124
124: Scheduled weekly dependency update for week 00 r=mithrandi a=pyup-bot




## Updates
Here's a list of all the updates bundled in this pull request. I've added some links to make it easier for you to find all the information you need.
<table align="center">

<tr>
<td><b>asn1crypto</b></td>
<td align="center">0.23.0</td>
<td align="center">&raquo;</td>
<td align="center">0.24.0</td>
<td>
     <a href="https://pypi.python.org/pypi/asn1crypto">PyPI</a> | <a href="https://pyup.io/changelogs/asn1crypto/">Changelog</a> | <a href="https://github.com/wbond/asn1crypto/issues">Repo</a> 

</td>

<tr>
<td><b>attrs</b></td>
<td align="center">17.3.0</td>
<td align="center">&raquo;</td>
<td align="center">17.4.0</td>
<td>
     <a href="https://pypi.python.org/pypi/attrs">PyPI</a> | <a href="https://pyup.io/changelogs/attrs/">Changelog</a> | <a href="http://www.attrs.org/">Homepage</a> 

</td>

<tr>
<td><b>ipaddress</b></td>
<td align="center">1.0.18</td>
<td align="center">&raquo;</td>
<td align="center">1.0.19</td>
<td>
     <a href="https://pypi.python.org/pypi/ipaddress">PyPI</a> | <a href="https://github.com/phihag/ipaddress">Repo</a> 

</td>

</tr>
</table>



## Changelogs


### asn1crypto 0.23.0 -> 0.24.0

>### 0.24.0


> - `x509.Certificate().self_signed` will no longer return `&quot;yes&quot;` under any
>   circumstances. This helps prevent confusion since the library does not
>   verify the signature. Instead a library like oscrypto should be used
>   to confirm if a certificate is self-signed.
> - Added various OIDs to `x509.KeyPurposeId()`
> - Added `x509.Certificate().private_key_usage_period_value`
> - Added structures for parsing common subject directory attributes for
>   X.509 certificates, including `x509.SubjectDirectoryAttribute()`
> - Added `algos.AnyAlgorithmIdentifier()` for situations where an
>   algorithm identifier may contain a digest, signed digest or encryption
>   algorithm OID
> - Fixed a bug with `x509.Certificate().subject_directory_attributes_value`
>   not returning the correct value
> - Fixed a bug where explicitly-tagged fields in a `core.Sequence()` would
>   not function properly when the field had a default value
> - Fixed a bug with type checking in `pem.armor()`







### attrs 17.3.0 -> 17.4.0

>### 17.4.0

>-------------------

>Backward-incompatible Changes
>^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

>- The traversal of MROs when using multiple inheritance was backward:
>  If you defined a class ``C`` that subclasses ``A`` and ``B`` like ``C(A, B)``, ``attrs`` would have collected the attributes from ``B`` *before* those of ``A``.

>  This is now fixed and means that in classes that employ multiple inheritance, the output of ``__repr__`` and the order of positional arguments in ``__init__`` changes.
>  Due to the nature of this bug, a proper deprecation cycle was unfortunately impossible.

>  Generally speaking, it&#39;s advisable to prefer ``kwargs``-based initialization anyways – *especially* if you employ multiple inheritance and diamond-shaped hierarchies.

>  `298 &lt;https://github.com/python-attrs/attrs/issues/298&gt;`_,
>  `299 &lt;https://github.com/python-attrs/attrs/issues/299&gt;`_,
>  `304 &lt;https://github.com/python-attrs/attrs/issues/304&gt;`_
>- The ``__repr__`` set by ``attrs``
>  no longer produces an ``AttributeError``
>  when the instance is missing some of the specified attributes
>  (either through deleting
>  or after using ``init=False`` on some attributes).

>  This can break code
>  that relied on ``repr(attr_cls_instance)`` raising ``AttributeError``
>  to check if any attr-specified members were unset.

>  If you were using this,
>  you can implement a custom method for checking this::

>      def has_unset_members(self):
>          for field in attr.fields(type(self)):
>              try:
>                  getattr(self, field.name)
>              except AttributeError:
>                  return True
>          return False

>  `308 &lt;https://github.com/python-attrs/attrs/issues/308&gt;`_


>Deprecations
>^^^^^^^^^^^^

>- The ``attr.ib(convert=callable)`` option is now deprecated in favor of ``attr.ib(converter=callable)``.

>  This is done to achieve consistency with other noun-based arguments like *validator*.

>  *convert* will keep working until at least January 2019 while raising a ``DeprecationWarning``.

>  `307 &lt;https://github.com/python-attrs/attrs/issues/307&gt;`_


>Changes
>^^^^^^^

>- Generated ``__hash__`` methods now hash the class type along with the attribute values.
>  Until now the hashes of two classes with the same values were identical which was a bug.

>  The generated method is also *much* faster now.

>  `261 &lt;https://github.com/python-attrs/attrs/issues/261&gt;`_,
>  `295 &lt;https://github.com/python-attrs/attrs/issues/295&gt;`_,
>  `296 &lt;https://github.com/python-attrs/attrs/issues/296&gt;`_
>- ``attr.ib``\ ’s ``metadata`` argument now defaults to a unique empty ``dict`` instance instead of sharing a common empty ``dict`` for all.
>  The singleton empty ``dict`` is still enforced.

>  `280 &lt;https://github.com/python-attrs/attrs/issues/280&gt;`_
>- ``ctypes`` is optional now however if it&#39;s missing, a bare ``super()`` will not work in slots classes.
>  This should only happen in special environments like Google App Engine.

>  `284 &lt;https://github.com/python-attrs/attrs/issues/284&gt;`_,
>  `286 &lt;https://github.com/python-attrs/attrs/issues/286&gt;`_
>- The attribute redefinition feature introduced in 17.3.0 now takes into account if an attribute is redefined via multiple inheritance.
>  In that case, the definition that is closer to the base of the class hierarchy wins.

>  `285 &lt;https://github.com/python-attrs/attrs/issues/285&gt;`_,
>  `287 &lt;https://github.com/python-attrs/attrs/issues/287&gt;`_
>- Subclasses of ``auto_attribs=True`` can be empty now.

>  `291 &lt;https://github.com/python-attrs/attrs/issues/291&gt;`_,
>  `292 &lt;https://github.com/python-attrs/attrs/issues/292&gt;`_
>- Equality tests are *much* faster now.

>  `306 &lt;https://github.com/python-attrs/attrs/issues/306&gt;`_
>- All generated methods now have correct ``__module__``, ``__name__``, and (on Python 3) ``__qualname__`` attributes.

>  `309 &lt;https://github.com/python-attrs/attrs/issues/309&gt;`_


>----











That's it for now!

Happy merging! 🤖

bors-fusion bot added a commit to fusionapp/entropy that referenced this issue Jan 1, 2018

Merge #167
167: Scheduled weekly dependency update for week 00 r=mithrandi a=pyup-bot




## Updates
Here's a list of all the updates bundled in this pull request. I've added some links to make it easier for you to find all the information you need.
<table align="center">

<tr>
<td><b>asn1crypto</b></td>
<td align="center">0.23.0</td>
<td align="center">&raquo;</td>
<td align="center">0.24.0</td>
<td>
     <a href="https://pypi.python.org/pypi/asn1crypto">PyPI</a> | <a href="https://pyup.io/changelogs/asn1crypto/">Changelog</a> | <a href="https://github.com/wbond/asn1crypto/issues">Repo</a> 

</td>

<tr>
<td><b>attrs</b></td>
<td align="center">17.3.0</td>
<td align="center">&raquo;</td>
<td align="center">17.4.0</td>
<td>
     <a href="https://pypi.python.org/pypi/attrs">PyPI</a> | <a href="https://pyup.io/changelogs/attrs/">Changelog</a> | <a href="http://www.attrs.org/">Homepage</a> 

</td>

<tr>
<td><b>ipaddress</b></td>
<td align="center">1.0.18</td>
<td align="center">&raquo;</td>
<td align="center">1.0.19</td>
<td>
     <a href="https://pypi.python.org/pypi/ipaddress">PyPI</a> | <a href="https://github.com/phihag/ipaddress">Repo</a> 

</td>

<tr>
<td><b>txaws</b></td>
<td align="center">0.4.0</td>
<td align="center">&raquo;</td>
<td align="center">0.5.0</td>
<td>
     <a href="https://pypi.python.org/pypi/txaws">PyPI</a> | <a href="https://pyup.io/changelogs/txaws/">Changelog</a> | <a href="https://github.com/twisted/txaws">Repo</a> 

</td>

</tr>
</table>



## Changelogs


### asn1crypto 0.23.0 -> 0.24.0

>### 0.24.0


> - `x509.Certificate().self_signed` will no longer return `&quot;yes&quot;` under any
>   circumstances. This helps prevent confusion since the library does not
>   verify the signature. Instead a library like oscrypto should be used
>   to confirm if a certificate is self-signed.
> - Added various OIDs to `x509.KeyPurposeId()`
> - Added `x509.Certificate().private_key_usage_period_value`
> - Added structures for parsing common subject directory attributes for
>   X.509 certificates, including `x509.SubjectDirectoryAttribute()`
> - Added `algos.AnyAlgorithmIdentifier()` for situations where an
>   algorithm identifier may contain a digest, signed digest or encryption
>   algorithm OID
> - Fixed a bug with `x509.Certificate().subject_directory_attributes_value`
>   not returning the correct value
> - Fixed a bug where explicitly-tagged fields in a `core.Sequence()` would
>   not function properly when the field had a default value
> - Fixed a bug with type checking in `pem.armor()`







### attrs 17.3.0 -> 17.4.0

>### 17.4.0

>-------------------

>Backward-incompatible Changes
>^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

>- The traversal of MROs when using multiple inheritance was backward:
>  If you defined a class ``C`` that subclasses ``A`` and ``B`` like ``C(A, B)``, ``attrs`` would have collected the attributes from ``B`` *before* those of ``A``.

>  This is now fixed and means that in classes that employ multiple inheritance, the output of ``__repr__`` and the order of positional arguments in ``__init__`` changes.
>  Due to the nature of this bug, a proper deprecation cycle was unfortunately impossible.

>  Generally speaking, it&#39;s advisable to prefer ``kwargs``-based initialization anyways – *especially* if you employ multiple inheritance and diamond-shaped hierarchies.

>  `298 &lt;https://github.com/python-attrs/attrs/issues/298&gt;`_,
>  `299 &lt;https://github.com/python-attrs/attrs/issues/299&gt;`_,
>  `304 &lt;https://github.com/python-attrs/attrs/issues/304&gt;`_
>- The ``__repr__`` set by ``attrs``
>  no longer produces an ``AttributeError``
>  when the instance is missing some of the specified attributes
>  (either through deleting
>  or after using ``init=False`` on some attributes).

>  This can break code
>  that relied on ``repr(attr_cls_instance)`` raising ``AttributeError``
>  to check if any attr-specified members were unset.

>  If you were using this,
>  you can implement a custom method for checking this::

>      def has_unset_members(self):
>          for field in attr.fields(type(self)):
>              try:
>                  getattr(self, field.name)
>              except AttributeError:
>                  return True
>          return False

>  `308 &lt;https://github.com/python-attrs/attrs/issues/308&gt;`_


>Deprecations
>^^^^^^^^^^^^

>- The ``attr.ib(convert=callable)`` option is now deprecated in favor of ``attr.ib(converter=callable)``.

>  This is done to achieve consistency with other noun-based arguments like *validator*.

>  *convert* will keep working until at least January 2019 while raising a ``DeprecationWarning``.

>  `307 &lt;https://github.com/python-attrs/attrs/issues/307&gt;`_


>Changes
>^^^^^^^

>- Generated ``__hash__`` methods now hash the class type along with the attribute values.
>  Until now the hashes of two classes with the same values were identical which was a bug.

>  The generated method is also *much* faster now.

>  `261 &lt;https://github.com/python-attrs/attrs/issues/261&gt;`_,
>  `295 &lt;https://github.com/python-attrs/attrs/issues/295&gt;`_,
>  `296 &lt;https://github.com/python-attrs/attrs/issues/296&gt;`_
>- ``attr.ib``\ ’s ``metadata`` argument now defaults to a unique empty ``dict`` instance instead of sharing a common empty ``dict`` for all.
>  The singleton empty ``dict`` is still enforced.

>  `280 &lt;https://github.com/python-attrs/attrs/issues/280&gt;`_
>- ``ctypes`` is optional now however if it&#39;s missing, a bare ``super()`` will not work in slots classes.
>  This should only happen in special environments like Google App Engine.

>  `284 &lt;https://github.com/python-attrs/attrs/issues/284&gt;`_,
>  `286 &lt;https://github.com/python-attrs/attrs/issues/286&gt;`_
>- The attribute redefinition feature introduced in 17.3.0 now takes into account if an attribute is redefined via multiple inheritance.
>  In that case, the definition that is closer to the base of the class hierarchy wins.

>  `285 &lt;https://github.com/python-attrs/attrs/issues/285&gt;`_,
>  `287 &lt;https://github.com/python-attrs/attrs/issues/287&gt;`_
>- Subclasses of ``auto_attribs=True`` can be empty now.

>  `291 &lt;https://github.com/python-attrs/attrs/issues/291&gt;`_,
>  `292 &lt;https://github.com/python-attrs/attrs/issues/292&gt;`_
>- Equality tests are *much* faster now.

>  `306 &lt;https://github.com/python-attrs/attrs/issues/306&gt;`_
>- All generated methods now have correct ``__module__``, ``__name__``, and (on Python 3) ``__qualname__`` attributes.

>  `309 &lt;https://github.com/python-attrs/attrs/issues/309&gt;`_


>----











That's it for now!

Happy merging! 🤖

bors-fusion bot added a commit to fusionapp/fusion-index that referenced this issue Jan 1, 2018

Merge #174
174: Scheduled weekly dependency update for week 00 r=mithrandi a=pyup-bot




## Updates
Here's a list of all the updates bundled in this pull request. I've added some links to make it easier for you to find all the information you need.
<table align="center">

<tr>
<td><b>asn1crypto</b></td>
<td align="center">0.23.0</td>
<td align="center">&raquo;</td>
<td align="center">0.24.0</td>
<td>
     <a href="https://pypi.python.org/pypi/asn1crypto">PyPI</a> | <a href="https://pyup.io/changelogs/asn1crypto/">Changelog</a> | <a href="https://github.com/wbond/asn1crypto/issues">Repo</a> 

</td>

<tr>
<td><b>attrs</b></td>
<td align="center">17.3.0</td>
<td align="center">&raquo;</td>
<td align="center">17.4.0</td>
<td>
     <a href="https://pypi.python.org/pypi/attrs">PyPI</a> | <a href="https://pyup.io/changelogs/attrs/">Changelog</a> | <a href="http://www.attrs.org/">Homepage</a> 

</td>

<tr>
<td><b>hypothesis</b></td>
<td align="center">3.40.1</td>
<td align="center">&raquo;</td>
<td align="center">3.44.4</td>
<td>
     <a href="https://pypi.python.org/pypi/hypothesis">PyPI</a> | <a href="https://pyup.io/changelogs/hypothesis/">Changelog</a> | <a href="https://github.com/HypothesisWorks/hypothesis/issues">Repo</a> 

</td>

<tr>
<td><b>ipaddress</b></td>
<td align="center">1.0.18</td>
<td align="center">&raquo;</td>
<td align="center">1.0.19</td>
<td>
     <a href="https://pypi.python.org/pypi/ipaddress">PyPI</a> | <a href="https://github.com/phihag/ipaddress">Repo</a> 

</td>

<tr>
<td><b>pyrsistent</b></td>
<td align="center">0.14.1</td>
<td align="center">&raquo;</td>
<td align="center">0.14.2</td>
<td>
     <a href="https://pypi.python.org/pypi/pyrsistent">PyPI</a> | <a href="https://pyup.io/changelogs/pyrsistent/">Changelog</a> | <a href="http://github.com/tobgu/pyrsistent/">Repo</a> 

</td>

<tr>
<td><b>toolz</b></td>
<td align="center">0.8.2</td>
<td align="center">&raquo;</td>
<td align="center">0.9.0</td>
<td>
     <a href="https://pypi.python.org/pypi/toolz">PyPI</a> | <a href="https://github.com/pytoolz/toolz">Repo</a> 

</td>

</tr>
</table>



## Changelogs


### asn1crypto 0.23.0 -> 0.24.0

>### 0.24.0


> - `x509.Certificate().self_signed` will no longer return `&quot;yes&quot;` under any
>   circumstances. This helps prevent confusion since the library does not
>   verify the signature. Instead a library like oscrypto should be used
>   to confirm if a certificate is self-signed.
> - Added various OIDs to `x509.KeyPurposeId()`
> - Added `x509.Certificate().private_key_usage_period_value`
> - Added structures for parsing common subject directory attributes for
>   X.509 certificates, including `x509.SubjectDirectoryAttribute()`
> - Added `algos.AnyAlgorithmIdentifier()` for situations where an
>   algorithm identifier may contain a digest, signed digest or encryption
>   algorithm OID
> - Fixed a bug with `x509.Certificate().subject_directory_attributes_value`
>   not returning the correct value
> - Fixed a bug where explicitly-tagged fields in a `core.Sequence()` would
>   not function properly when the field had a default value
> - Fixed a bug with type checking in `pem.armor()`







### attrs 17.3.0 -> 17.4.0

>### 17.4.0

>-------------------

>Backward-incompatible Changes
>^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

>- The traversal of MROs when using multiple inheritance was backward:
>  If you defined a class ``C`` that subclasses ``A`` and ``B`` like ``C(A, B)``, ``attrs`` would have collected the attributes from ``B`` *before* those of ``A``.

>  This is now fixed and means that in classes that employ multiple inheritance, the output of ``__repr__`` and the order of positional arguments in ``__init__`` changes.
>  Due to the nature of this bug, a proper deprecation cycle was unfortunately impossible.

>  Generally speaking, it&#39;s advisable to prefer ``kwargs``-based initialization anyways – *especially* if you employ multiple inheritance and diamond-shaped hierarchies.

>  `298 &lt;https://github.com/python-attrs/attrs/issues/298&gt;`_,
>  `299 &lt;https://github.com/python-attrs/attrs/issues/299&gt;`_,
>  `304 &lt;https://github.com/python-attrs/attrs/issues/304&gt;`_
>- The ``__repr__`` set by ``attrs``
>  no longer produces an ``AttributeError``
>  when the instance is missing some of the specified attributes
>  (either through deleting
>  or after using ``init=False`` on some attributes).

>  This can break code
>  that relied on ``repr(attr_cls_instance)`` raising ``AttributeError``
>  to check if any attr-specified members were unset.

>  If you were using this,
>  you can implement a custom method for checking this::

>      def has_unset_members(self):
>          for field in attr.fields(type(self)):
>              try:
>                  getattr(self, field.name)
>              except AttributeError:
>                  return True
>          return False

>  `308 &lt;https://github.com/python-attrs/attrs/issues/308&gt;`_


>Deprecations
>^^^^^^^^^^^^

>- The ``attr.ib(convert=callable)`` option is now deprecated in favor of ``attr.ib(converter=callable)``.

>  This is done to achieve consistency with other noun-based arguments like *validator*.

>  *convert* will keep working until at least January 2019 while raising a ``DeprecationWarning``.

>  `307 &lt;https://github.com/python-attrs/attrs/issues/307&gt;`_


>Changes
>^^^^^^^

>- Generated ``__hash__`` methods now hash the class type along with the attribute values.
>  Until now the hashes of two classes with the same values were identical which was a bug.

>  The generated method is also *much* faster now.

>  `261 &lt;https://github.com/python-attrs/attrs/issues/261&gt;`_,
>  `295 &lt;https://github.com/python-attrs/attrs/issues/295&gt;`_,
>  `296 &lt;https://github.com/python-attrs/attrs/issues/296&gt;`_
>- ``attr.ib``\ ’s ``metadata`` argument now defaults to a unique empty ``dict`` instance instead of sharing a common empty ``dict`` for all.
>  The singleton empty ``dict`` is still enforced.

>  `280 &lt;https://github.com/python-attrs/attrs/issues/280&gt;`_
>- ``ctypes`` is optional now however if it&#39;s missing, a bare ``super()`` will not work in slots classes.
>  This should only happen in special environments like Google App Engine.

>  `284 &lt;https://github.com/python-attrs/attrs/issues/284&gt;`_,
>  `286 &lt;https://github.com/python-attrs/attrs/issues/286&gt;`_
>- The attribute redefinition feature introduced in 17.3.0 now takes into account if an attribute is redefined via multiple inheritance.
>  In that case, the definition that is closer to the base of the class hierarchy wins.

>  `285 &lt;https://github.com/python-attrs/attrs/issues/285&gt;`_,
>  `287 &lt;https://github.com/python-attrs/attrs/issues/287&gt;`_
>- Subclasses of ``auto_attribs=True`` can be empty now.

>  `291 &lt;https://github.com/python-attrs/attrs/issues/291&gt;`_,
>  `292 &lt;https://github.com/python-attrs/attrs/issues/292&gt;`_
>- Equality tests are *much* faster now.

>  `306 &lt;https://github.com/python-attrs/attrs/issues/306&gt;`_
>- All generated methods now have correct ``__module__``, ``__name__``, and (on Python 3) ``__qualname__`` attributes.

>  `309 &lt;https://github.com/python-attrs/attrs/issues/309&gt;`_


>----








### hypothesis 3.40.1 -> 3.44.4

>### 3.44.4

>-------------------

>This release fixes :issue:`1044`, which slowed tests by up to 6%
>due to broken caching.

>-------------------


>### 3.44.3

>-------------------

>This release improves the shrinker in cases where examples drawn earlier can
>affect how much data is drawn later (e.g. when you draw a length parameter in
>a composite and then draw that many elements). Examples found in cases like
>this should now be much closer to minimal.

>-------------------


>### 3.44.2

>-------------------

>This is a pure refactoring release which changes how Hypothesis manages its
>set of examples internally. It should have no externally visible effects.

>-------------------


>### 3.44.1

>-------------------

>This release fixes :issue:`997`, in which under some circumstances the body of
>tests run under Hypothesis would not show up when run under coverage even
>though the tests were run and the code they called outside of the test file
>would show up normally.

>-------------------


>### 3.44.0

>-------------------

>This release adds a new feature: The :ref:`reproduce_failure &lt;reproduce_failure&gt;`,
>designed to make it easy to use Hypothesis&#39;s binary format for examples to
>reproduce a problem locally without having to share your example database
>between machines.

>This also changes when seeds are printed:

>* They will no longer be printed for
>  normal falsifying examples, as there are now adequate ways of reproducing those
>  for all cases, so it just contributes noise.
>* They will once again be printed when reusing examples from the database, as
>  health check failures should now be more reliable in this scenario so it will
>  almost always work in this case.

>This work was funded by `Smarkets &lt;https://smarkets.com/&gt;`_.

>-------------------


>### 3.43.1

>-------------------

>This release fixes a bug with Hypothesis&#39;s database management - examples that
>were found in the course of shrinking were saved in a way that indicated that
>they had distinct causes, and so they would all be retried on the start of the
>next test. The intended behaviour, which is now what is implemented, is that
>only a bounded subset of these examples would be retried.

>-------------------


>### 3.43.0

>-------------------

>:exc:`~hypothesis.errors.HypothesisDeprecationWarning` now inherits from
>:exc:`python:FutureWarning` instead of :exc:`python:DeprecationWarning`,
>as recommended by :pep:`565` for user-facing warnings (:issue:`618`).
>If you have not changed the default warnings settings, you will now see
>each distinct :exc:`~hypothesis.errors.HypothesisDeprecationWarning`
>instead of only the first.

>-------------------


>### 3.42.2

>-------------------

>This patch fixes :issue:`1017`, where instances of a list or tuple subtype
>used as an argument to a strategy would be coerced to tuple.

>-------------------


>### 3.42.1

>-------------------

>This release has some internal cleanup, which makes reading the code
>more pleasant and may shrink large examples slightly faster.

>-------------------


>### 3.42.0

>-------------------

>This release deprecates :ref:`faker-extra`, which was designed as a transition
>strategy but does not support example shrinking or coverage-guided discovery.

>-------------------


>### 3.41.0

>-------------------

>:func:`~hypothesis.strategies.sampled_from` can now sample from
>one-dimensional numpy ndarrays. Sampling from multi-dimensional
>ndarrays still results in a deprecation warning. Thanks to Charlie
>Tanksley for this patch.

>-------------------






### pyrsistent 0.14.1 -> 0.14.2

>### 0.14.2

> * Fix 121, regression in PClass.set() introduced in 0.14.1.










That's it for now!

Happy merging! 🤖
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment