Windows - Add common util for web requests

[jborean93]

SUMMARY

Adds a common URL handler for PowerShell modules. It tries to combine common elements like shared module arg spec, reduce code duplication for the somewhat convoluted WebRequest class in .NET and add the ability to easily add common module options for URL based modules in the future.

Things I know that still need to be fixed/solved before this can move away from a WIP PR;

• Fix validate-modules to load other PS and C# module utils so the module arg spec can be from a util. Currently it fails because we don't load them Edit: Done with Load Ansible module_utils for ps_argspec validator #58571
• Figure out how we want to document the shared values and when they were added. Right now it just omits the version_added component in the shared doc fragment because some options were present/added in some modules but not others - Doc fragments are more versatile than I expected, have been able to preserve the version_added for specific modules while still sharing the docs as much as possible.
• Whether Invoke-WithWebRequest is too complicated with it's use of a ScriptBlock
• Whether Register-AnsibleWebRequestParams is something we want to use. Unfortunately we cannot take advantage of splatting from $module.Params and this was the next best thing I could find that allowed the module to ignore options that were implemented inside. - Alternative option was chosen where the module (if passed in) would provide a backup/default value for parameters not explicitly passed in

ISSUE TYPE

• Feature Pull Request

COMPONENT NAME

win_get_url
win_uri

[ansibot]

cc @blakfeld @dagwieers @if-meaton @jhawkesworth @nitzmahone @tksarah
click here for bot help

[ansibot]

cc @ShachafGoldstein
click here for bot help

[jhawkesworth]

Another impressive piece of work.
I haven't spotted anything that looks like it needs changing, other than pslint/ignore.txt conflict.

The only thing I can think of is maybe to benchmark against existing modules, although I can't think of a reason why performance might change materially (and network conditions are much more likely to affect actual execution time in use) but more testing before merging a big refactoring allways improves confidence that the change will stick.

[ansibot]

The test ansible-test sanity --test future-import-boilerplate [explain] failed with 1 error:

lib/ansible/plugins/doc_fragments/url_windows.py:0:0: missing: from __future__ import (absolute_import, division, print_function)

The test ansible-test sanity --test metaclass-boilerplate [explain] failed with 1 error:

lib/ansible/plugins/doc_fragments/url_windows.py:0:0: missing: __metaclass__ = type

The test ansible-test sanity --test validate-modules [explain] failed with 2 errors:

lib/ansible/modules/windows/win_get_url.ps1:0:0: E325 Argument 'client_cert' in argument_spec defines type as 'str' but documentation defines type as 'path'
lib/ansible/modules/windows/win_uri.ps1:0:0: E325 Argument 'client_cert' in argument_spec defines type as 'str' but documentation defines type as 'path'

click here for bot help

[nitzmahone]

Looks great, one little thing to think about with the default handling on explicit HttpAgent arg vs headers. Much nicer setup in general. I don't love the pre-existing continuation pattern of Invoke-WithWebRequest, but it seems to work OK for most things right now- something that could probably be revisited later if there's a good reason.

[ansibot]

cc @acozine
click here for bot help

[jborean93]

Merging to get this out into people's installs so we can hopefully detect any bugs before the 2.9 release. Thanks everyone for the reviews.