Switch back http provider after requests removes LGPL dependency (#16974)

potiuk · web-flow · commit c46e841519ef · 2021-07-13T22:13:30.000+02:00
Following merging the psf/requests#5797 and requests 2.26.0 release without LGPL chardet dependency, we can now bring back http as pre-installed provider as it does not bring chardet automatically any more.
diff --git a/docs/apache-airflow/extra-packages-ref.rst b/docs/apache-airflow/extra-packages-ref.rst
@@ -260,7 +260,7 @@ Those are extras that provide support for integration with external systems via
 +---------------------+-----------------------------------------------------+--------------------------------------+--------------+
 | grpc                | ``pip install 'apache-airflow[grpc]'``              | Grpc hooks and operators             |              |
 +---------------------+-----------------------------------------------------+--------------------------------------+--------------+
-| http                | ``pip install 'apache-airflow[http]'``              | HTTP hooks, operators and sensors    |              |
+| http                | ``pip install 'apache-airflow[http]'``              | HTTP hooks, operators and sensors    |      *       |
 +---------------------+-----------------------------------------------------+--------------------------------------+--------------+
 | imap                | ``pip install 'apache-airflow[imap]'``              | IMAP hooks and sensors               |      *       |
 +---------------------+-----------------------------------------------------+--------------------------------------+--------------+
diff --git a/setup.py b/setup.py
@@ -231,13 +231,13 @@ def write_version(filename: str = os.path.join(*[my_dir, "airflow", "git_version
     'distributed>=2.11.1, <2.20',
 ]
 databricks = [
-    'requests>=2.20.0, <3',
+    'requests>=2.26.0, <3',
 ]
 datadog = [
     'datadog>=0.14.0',
 ]
 deprecated_api = [
-    'requests>=2.20.0',
+    'requests>=2.26.0',
 ]
 doc = [
     # Sphinx is limited to < 3.5.0 because of https://github.com/sphinx-doc/sphinx/issues/8880
@@ -333,7 +333,9 @@ def write_version(filename: str = os.path.join(*[my_dir, "airflow", "git_version
     'thrift>=0.9.2',
 ]
 http = [
-    'requests>=2.20.0',
+    # The 2.26.0 release of requests got rid of the chardet LGPL mandatory dependency, allowing us to
+    # release it as a requirement for airflow
+    'requests>=2.26.0',
 ]
 http_provider = [
     # NOTE ! The HTTP provider is NOT preinstalled by default when Airflow is installed - because it
@@ -815,12 +817,9 @@ def sort_extras_requirements() -> Dict[str, List[str]]:
 # Those providers are pre-installed always when airflow is installed.
 # Those providers do not have dependency on airflow2.0 because that would lead to circular dependencies.
 # This is not a problem for PIP but some tools (pipdeptree) show those as a warning.
-# NOTE ! The HTTP provider is NOT preinstalled by default when Airflow is installed - because it
-#        depends on `requests` library and until `chardet` is mandatory dependency of `requests`
-#        we cannot make it mandatory dependency. See https://github.com/psf/requests/pull/5797
 PREINSTALLED_PROVIDERS = [
     'ftp',
-    # 'http',
+    'http',
     'imap',
     'sqlite',
 ]

Original file line number	Diff line number	Diff line change
`@@ -231,13 +231,13 @@ def write_version(filename: str = os.path.join(*[my_dir, "airflow", "git_version`
`231`	`231`	`'distributed>=2.11.1, <2.20',`
`232`	`232`	`]`
`233`	`233`	`databricks = [`
`234`		`- 'requests>=2.20.0, <3',`
	`234`	`+ 'requests>=2.26.0, <3',`
`235`	`235`	`]`
`236`	`236`	`datadog = [`
`237`	`237`	`'datadog>=0.14.0',`
`238`	`238`	`]`
`239`	`239`	`deprecated_api = [`
`240`		`- 'requests>=2.20.0',`
	`240`	`+ 'requests>=2.26.0',`
`241`	`241`	`]`
`242`	`242`	`doc = [`
`243`	`243`	`# Sphinx is limited to < 3.5.0 because of https://github.com/sphinx-doc/sphinx/issues/8880`
`@@ -333,7 +333,9 @@ def write_version(filename: str = os.path.join(*[my_dir, "airflow", "git_version`
`333`	`333`	`'thrift>=0.9.2',`
`334`	`334`	`]`
`335`	`335`	`http = [`
`336`		`- 'requests>=2.20.0',`
	`336`	`+ # The 2.26.0 release of requests got rid of the chardet LGPL mandatory dependency, allowing us to`
	`337`	`+ # release it as a requirement for airflow`
	`338`	`+ 'requests>=2.26.0',`
`337`	`339`	`]`
`338`	`340`	`http_provider = [`
`339`	`341`	`# NOTE ! The HTTP provider is NOT preinstalled by default when Airflow is installed - because it`
`@@ -815,12 +817,9 @@ def sort_extras_requirements() -> Dict[str, List[str]]:`
`815`	`817`	`# Those providers are pre-installed always when airflow is installed.`
`816`	`818`	`# Those providers do not have dependency on airflow2.0 because that would lead to circular dependencies.`
`817`	`819`	`# This is not a problem for PIP but some tools (pipdeptree) show those as a warning.`
`818`		`-# NOTE ! The HTTP provider is NOT preinstalled by default when Airflow is installed - because it`
`819`		-# depends on `requests` library and until `chardet` is mandatory dependency of `requests`
`820`		`-# we cannot make it mandatory dependency. See https://github.com/psf/requests/pull/5797`
`821`	`820`	`PREINSTALLED_PROVIDERS = [`
`822`	`821`	`'ftp',`
`823`		`- # 'http',`
	`822`	`+ 'http',`
`824`	`823`	`'imap',`
`825`	`824`	`'sqlite',`
`826`	`825`	`]`