ENH Refactor Element and graphql hoc

[emteknetnz]

Issue https://github.com/silverstripe/silverstripe-elemental/issues/1164

Notes:

• I haven't implemented the AC "(refactor) Triggering actions on state change" - I've made a comment about it https://github.com/silverstripe/silverstripe-elemental/issues/1164#issuecomment-2068481613 - I don't intend to implement it though if someone is able to demonstrate a non-convoluted way to implement an alternative I'm happy to refactor
• I haven't implemented the AC "Passing down callbacks in context (note that context is required to pass props to the buttons because they're added via js injector in registerTransforms.js)" - I've made a comment about it https://github.com/silverstripe/silverstripe-elemental/issues/1164#issuecomment-2068381327. I do not intend to implement this AC
• I haven't implmented the AC "Use of setTimeout in Element.js" as things don't work when setTimeout is removed, presumably due to timing issues. I have made it to the original timeout happens inside a .then() block so it always happens after the HTTP request has returned.
• I haven't implemented the AC "Moving context to its own file ... or remove the need for context." as I was asked to revert it in this comment in peer review
• I've removed some jest tests on SaveAction-test.js and PublishAction-test.js as there is simply now less logic and props passed via context in SaveAction and PublishAction. I choose not to add in equivalent tests in Element-test.js adding in "child Actions" would require doing stuff with injector to replicate what registerTransforms.js does now, and also because these functionality is already covered by behat tests in edit-block-element.feature (save button) and publish-block-element.feature (publish button)
• I've commented out a failing behat test and raised a follow card to address https://github.com/silverstripe/silverstripe-elemental/issues/1177. It's a very weird and convoluted issue which I spent a great deal of time on though was ultimately unable to resolve

[GuySartorelli]

Seems to work fine in the browser as-is so we're just dealing with code quality at this point. If the refetch queries change I've suggested above doesn't remove the need for the timeout, it may not be worth doing.