Pydantic 1.10.3 incompatible with typing-extensions 4.1.1

[hsheth2]

Initial Checks

• I have searched GitHub for a duplicate issue and I'm sure this is something new
• I have searched Google & StackOverflow for a solution and couldn't find anything
• I have read and followed the docs and still think this is a bug
• I am confident that the issue is with pydantic (not my code, or another library in the ecosystem like FastAPI or mypy)

Description

• Installed latest pydantic and typing-extensions 4.1.1
• Code crashes on import pydantic

I suspect we just need to bump the min version requirement or fix a cross-version incompatibility, since it works with typing-extensions 4.4.0.

Example Code

# Set up a clean venv
# pip install pydantic==1.10.3 typing-extensions==4.1.1

>>> import pydantic
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "pydantic/__init__.py", line 2, in init pydantic.__init__
  File "pydantic/dataclasses.py", line 46, in init pydantic.dataclasses
    # | None    | Attribute is set to None.               |
  File "pydantic/main.py", line 121, in init pydantic.main
TypeError: dataclass_transform() got an unexpected keyword argument 'field_specifiers'

Python, Pydantic & OS Version

Crashes so I can't run that

Tested with python 3.7.9

Affected Components

• Compatibility between releases
• Data validation/parsing
• Data serialization - .dict() and .json()
• JSON Schema
• Dataclasses
• Model Config
• Field Types - adding or changing a particular data type
• Function validation decorator
• Generic Models
• Other Model behaviour - construct(), pickling, private attributes, ORM mode
• Plugins and integration with other tools - mypy, FastAPI, python-devtools, Hypothesis, VS Code, PyCharm, etc.

[samuelcolvin]

Grrr, good catch, sorry I didn't see this before.

PR welcome to fix this.

[samuelcolvin]

See #4886

[adrianeboyd]

On behalf of spacy-related packages: would it be possible for you to temporarily yank v1.10.3?

To address this and be compatible with v1.10.4, we'd have to release new versions of a whole series of packages and nearly everyone (including me) is currently on vacation. Even if v1.10.4 is released with a fix, pip would still back off to v1.10.3 for spacy, etc. because of its current pins for typing_extensions. If it could instead back off to v1.10.2, we'd have a bit more breathing room to make the updates on our end.

[samuelcolvin]

I'll make a new release of v1.10.4 today, I think that's better.

Sorry again for the inconvenience.

[samuelcolvin]

Humm, just read your message fully 🙈.

Okay, unless @hramezani or @PrettyWood have a problem, I'll yank v1.10.3.

Once that's done, do you have a problem with me releasing v1.10.4 with just this fixed?

[adrianeboyd]

Thanks for the quick response! I don't think that a v1.10.4 release with the updated requirements should cause any problems on our end (and would be used for python 3.8+).

[samuelcolvin]

v1.10.3 has been yanked.

[samuelcolvin]

v1.10.4 is building now.

[kjozef-work]

we do see the same error with 1.10.4 still
Downloading pydantic-1.10.4-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (3.1 MB)

[samuelcolvin]

Have you checked the content of that wheel (it's just a zip file)? It should be correct.

[kjozef-work]

apologize for not providing details - pydantic 1.10.4 fails still on python 3.8 with typing-extensions 4.1.1 , it passes though on python 3.7 with te-4.1.1

[samuelcolvin]

Yes, that's why we changed the minimum allowed version in dependencies

pydantic/setup.py

Line 133 in 63c683d

'typing-extensions>=4.2.0'

Please review the release details.

[Hunterlige]

Hello,

The pydantic release on conda does not enforce typing-extensions>=4.2.0 and cause the same error:
TypeError: dataclass_transform() got an unexpected keyword argument 'field_specifiers'
https://github.com/conda-forge/pydantic-feedstock/blob/main/recipe/meta.yaml#L30

[samuelcolvin]

Thanks for reporting, I thought the bot auto updated that, apparently not.

I'll check tomorrow if I can update it without having to create another patch release.

[samuelcolvin]

Hello,

The pydantic release on conda does not enforce typing-extensions>=4.2.0 and cause the same error: TypeError: dataclass_transform() got an unexpected keyword argument 'field_specifiers' https://github.com/conda-forge/pydantic-feedstock/blob/main/recipe/meta.yaml#L30

See conda-forge/pydantic-feedstock#73