The Branch protection API has been radically changed by GitHub, add new
methods to Branch reflecting it. Branch protection methods are now
called on the Branch object itself, rather than awkwardly hanging off
the Repository object and requiring branch names to be passed in.
This adds an over-arching methods to get, edit and remove protection
entirely, as well as fine-grained methods for getting, setting and
dropping certain aspects of branch protection, as well as three new
objects to encompass querying those aspects.
This also destroys Repository.protect_branch(), the endpoint has been
removed. The old-style protection attributes on Branch have also been
removed to force users onto the new API, since they are still sent, but
no longer populated.
Fixes#586
Add create_review_request() and delete_review_request() methods to
PullRequest, as well as renaming get_reviewer_requests() to
get_review_requests() and cleaning up its return value to firstly respect
teams, and secondly to cut out the middleman of PullRequestReviewerRequests,
which has been removed.
Fixes#597
* List reactions of issue
* Create issue reaction
* List reactions of an issue comment
* Create a reaction on an issue comment
* List reactions of a commit comment
* Create a reaction on a commit comment
* List reactions of a pull request review comment
* Create a reaction on a pull request review comment
* Delete reaction
* Renames IssueReaction to Reaction
* Fixes docstring in Reaction class
* Removes issue_url on Reaction class
* Fixes string concatenation that does not work on python 2.6
This is a minor thing. It’s a convention, not a rule (obviously)—but it’s also a best practice.
Many text editors look for this kind of comment as the first line—or second line, if there is a shebang (the `#!/usr/bin/env python`, or whatever else)—and make use of this to determine the file’s encoding.