Reusable blocks e2e tests fail if starting from new post

[ellatrix]

Description

Sparked by #9962. These tests should still pass and demonstrate some issues starting from a new post.

How has this been tested?

Screenshots

Types of changes

Checklist:

• My code is tested.
• My code follows the WordPress code style.
• My code follows the accessibility standards.
• My code has proper inline documentation.

[ellatrix]

See #9962 (comment).

[aduth]

cc @noisysocks

[noisysocks]

I rebased this and pushed up some changes which should hopefully fix the failing tests here.

There were three problems:

1) A reusable block's Name field would get selected after a fetchReusableBlocks()

After a successful fetch request, RECEIVE_REUSABLE_BLOCKS was dispatched which would then in turn dispatch RECEIVE_BLOCKS. This meant that there was a brief moment (after the first dispatch but before the second dispatch) where getReusableBlock( ref ) would return a reusable block that referenced a block that didn't exist in the store. This caused the block to render its Block has been deleted or is unavailable message, which in turn caused ReusableBlockEditPanel to remount, which in turn caused the Name field to be select()ed.

I fixed this by removing the RECEIVE_REUSABLE_BLOCKS effect handler and instead making the reducer handle RECEIVE_REUSABLE_BLOCKS directly.

2) Race condition between deleting a reusable block and fetching reusable blocks

After clicking Remove from Reusable Blocks on a new post with a reusable block in it, both deleteReusableBlock() and fetchReusableBlocks() would be dispatched at the same time. This was because the cursor would be placed into a Paragraph block which then activates blocksAutocompleter. These two HTTP requests would race and, often, the response to fetchReusableBlocks() would come back before the block had been deleted, causing Gutenberg to mistakenly show the deleted block in the inserter.

I fixed this by moving fetchReusableBlocks() from blocksAutocompleter and into SETUP_EDITOR.

3) E2E test not able to click on a reusable block in the inserter

Since the tests are now always starting from a new post, we now need to wait for fetchReusableBlocks() to finish before we can click on the reusable block in the inserter.

I fixed this by using waitForSelector().

[noisysocks]

Hm, damn, still some failures on Travis CI. I'm not sure if they're related.

[gziolo]

@iseulde can you rebase with master and see if e2e failures were unrelated to this PR?

[catehstn]

Moving this to 4.6, no rush to get it in today.

[gziolo]

@noisysocks and @iseulde - this one needs to be refreshed. Do you plan to get back to this PR? I'm marking it as Stale to make triaging and reviewing PR easier.

[ellatrix]

I don't. Looked at this at a meetup quick.

[noisysocks]

I don't have the bandwidth to do a lot here, but I can rebase it and maybe that will fix the test failures I was getting in #10267 (comment).