revert removing python 3.6, #3605

Author: samuelcolvin

.github/workflows/ci.yml

@@ -100,7 +100,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: ['3.7', '3.8', '3.9', '3.10']
+        python-version: ['3.6', '3.7', '3.8', '3.9', '3.10']
     env:
       PYTHON: ${{ matrix.python-version }}
       OS: ubuntu
@@ -155,9 +155,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu, macos, windows]
-        python-version: ['3.7', '3.8', '3.9', '3.10']
-        include:
-        - os: ubuntu
+        python-version: ['3.6', '3.7', '3.8', '3.9', '3.10']
 
     env:
       PYTHON: ${{ matrix.python-version }}
@@ -310,7 +308,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu , macos , windows]
-        python-version: ['7', '8', '9', '10']
+        python-version: ['6', '7', '8', '9', '10']
         include:
         - os: ubuntu
           platform: linux

README.md

@@ -11,7 +11,7 @@
 Data validation and settings management using Python type hints.
 
 Fast and extensible, *pydantic* plays nicely with your linters/IDE/brain.
-Define how data should be in pure, canonical Python 3.7+; validate it with *pydantic*.
+Define how data should be in pure, canonical Python 3.6+; validate it with *pydantic*.
 
 ## Help
 

changes/3605-samuelcolvin.md

@@ -33,7 +33,7 @@ To make contributing as easy and fast as possible, you'll want to run tests and
 *pydantic* has few dependencies, doesn't require compiling and tests don't need access to databases, etc.
 Because of this, setting up and running the tests should be very simple.
 
-You'll need to have a version between **python 3.7 and 3.10**, **virtualenv**, **git**, and **make** installed.
+You'll need to have a version between **python 3.6 and 3.10**, **virtualenv**, **git**, and **make** installed.
 
 ```bash
 # 1. clone your fork and cd into the repo directory
@@ -44,7 +44,7 @@ cd pydantic
 virtualenv -p `which python3.8` env
 source env/bin/activate
 # Building docs requires 3.8. If you don't need to build docs you can use
-# whichever version; 3.7 will work too.
+# whichever version; 3.6 will work too.
 
 # 3. Install pydantic, dependencies, test dependencies and doc dependencies
 make install

docs/contributing.md

@@ -4,9 +4,10 @@ Installation is as simple as:
 pip install pydantic
 ```
 
-*pydantic* has no required dependencies except python 3.7, 3.8, 3.9 or 3.10 and
-[`typing-extensions`](https://pypi.org/project/typing-extensions/).
-If you've got python 3.7+ and `pip` installed, you're good to go.
+*pydantic* has no required dependencies except python 3.6, 3.7, 3.8, 3.9 or 3.10,
+[`typing-extensions`](https://pypi.org/project/typing-extensions/), and the
+[`dataclasses`](https://pypi.org/project/dataclasses/) backport package for python 3.6.
+If you've got python 3.6+ and `pip` installed, you're good to go.
 
 Pydantic is also available on [conda](https://www.anaconda.com) under the [conda-forge](https://conda-forge.org)
 channel:

docs/install.md

@@ -1,6 +1,8 @@
 If you don't want to use _pydantic_'s `BaseModel` you can instead get the same data validation on standard
 [dataclasses](https://docs.python.org/3/library/dataclasses.html) (introduced in python 3.7).
 
+Dataclasses work in python 3.6 using the [dataclasses backport package](https://github.com/ericvsmith/dataclasses).
+
 ```py
 {!.tmp_examples/dataclasses_main.py!}
 ```

docs/usage/dataclasses.md

@@ -296,6 +296,10 @@ For example, in the example above, if `_fields_set` was not provided,
 
 Pydantic supports the creation of generic models to make it easier to reuse a common model structure.
 
+!!! warning
+    Generic models are only supported with python `>=3.7`, this is because of numerous subtle changes in how
+    generics are implemented between python 3.6 and python 3.7.
+
 In order to declare a generic model, you perform the following steps:
 
 * Declare one or more `typing.TypeVar` instances to use to parameterize your model.

docs/usage/models.md

@@ -1,5 +1,5 @@
 site_name: pydantic
-site_description: Data validation and settings management using python type hints
+site_description: Data validation and settings management using python 3.6 type hinting
 strict: true
 site_url: https://pydantic-docs.helpmanual.io/
 

mkdocs.yml

@@ -1,6 +1,6 @@
 import json
 from enum import Enum
-from typing import TYPE_CHECKING, Any, Callable, Dict, ForwardRef, Optional, Tuple, Type, Union
+from typing import TYPE_CHECKING, Any, Callable, Dict, Optional, Tuple, Type, Union
 
 from .typing import AnyCallable
 from .utils import GetterDict
@@ -59,7 +59,8 @@ class BaseConfig:
     schema_extra: Union[Dict[str, Any], 'SchemaExtraCallable'] = {}
     json_loads: Callable[[str], Any] = json.loads
     json_dumps: Callable[..., str] = json.dumps
-    json_encoders: Dict[Union[Type[Any], str, ForwardRef], AnyCallable] = {}
+    # key type should include ForwardRef, but that breaks with python3.6
+    json_encoders: Dict[Union[Type[Any], str], AnyCallable] = {}
     underscore_attrs_are_private: bool = False
 
     # whether inherited models as fields should be reconstructed as base model

pydantic/config.py

@@ -1138,6 +1138,7 @@ def is_complex(self) -> bool:
     def _type_display(self) -> PyObjectStr:
         t = display_as_type(self.type_)
 
+        # have to do this since display_as_type(self.outer_type_) is different (and wrong) on python 3.6
         if self.shape in MAPPING_LIKE_SHAPES:
             t = f'Mapping[{display_as_type(self.key_field.type_)}, {t}]'  # type: ignore
         elif self.shape == SHAPE_TUPLE: