Restore user credentials in HttpBackgroundJob

[hyzx86]

use httpContextAccessor.HttpContext.User.Clone(); to set new HttpContext.User

This PR is submitted in order to start the workflow asynchronously in the script, because this will create a new http context, which will make it impossible to retrieve user information

[hyzx86]

Hi @jtkech ,I use these changes to work well in my code.
What do you think of this implementation?

[Piedone]

Can you explain scenarios where you used this?

[Piedone]

Just to confirm: are you done?

[hyzx86]

Just to confirm: are you done?

Yes， please review it ,thanks!

[sebastienros]

Should it be done directly in shellContext.CreateHttpContext() maybe optionally as other callers of this method could have the same problem?

[Piedone]

Not really, because you don't always have an actual HTTP request. This is about executing something after a real HTTP request, but with a different HttpContext, and thus a user can exist. In e.g. background tasks, there's no HTTP request and thus can be no user. Out of the four callers of this method, only HttpBackgroundJob needs this: