add admin web api methods

[manuelmeurer]

I updated the API methods via rake slack:api:update, which added lots of new admin methods.
Once those are merged, we can remove the undocumented Slack::Web::Api::Endpoints::UsersAdmin.users_admin_invite

[manuelmeurer]

I changed a few small things in the changelog:

• the last entry referenced a wrong PR number (302 instead of 303) and misspelled the added endpoint (views.published instead of views.publish)
• most of the entries used past tense ("Added") and only a few present tense ("Add"). I changed the latter ones to past tense for consistency.

[dblock]

This looks good, but there seem to be legit spec failures? I haven't looked any deeper than that.

[manuelmeurer]

Ok, I'll investigate!

Why does Ruby 2.4.1 on Travis only run bundle exec danger, not the actual specs?
https://travis-ci.org/slack-ruby/slack-ruby-client/jobs/620311621

[manuelmeurer]

I get the same failures locally when I run the specs on master, so I don't think they are related to the changes in this PR.
I'll try to find out what's going on anyways. :)

[dblock]

Ok, I'll investigate!

Why does Ruby 2.4.1 on Travis only run bundle exec danger, not the actual specs?
https://travis-ci.org/slack-ruby/slack-ruby-client/jobs/620311621

Danger is what gives us changelog checking and stuff like that.

[dblock]

I get the same failures locally when I run the specs on master, so I don't think they are related to the changes in this PR.
I'll try to find out what's going on anyways. :)

Yes thank you, we want that fixed first.

[manuelmeurer]

@dblock The culprit is Faraday. They are preparing for a v1.0 release and (unintentionally?) introduced some breaking changes from 0.17.0 to 0.17.1.

Basically, in 0.17.0, Faraday::Error inherits from StandardError, so we have to add the possibility to initialize a SlackError with a response.

But in 0.17.1, Faraday::Error already does this itself already.

What to do? I added a check for the arity of Faraday::Error#initialize but I think it's ugly and brittle.
Alternatively, we could limit Faraday to either < 0.17.0 or 0.17.1.

[dblock]

Are they releasing 0.17.2 fixing the regression? Otherwise make our code compatible with 0.17.1 and add a >= 0.17.1 rather than limit below. We want to be compatible with the latest moving forward.

[manuelmeurer]

I opened an issue in the Faraday repo and will add a limit to >= 0.17.1 for now.

[manuelmeurer]

@dblock After the Faraday issue was fixed in #309, this should be good to go!