Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Refactor to remove usage of post related effects in packages/editor. #13716

Merged
merged 32 commits into from
Feb 28, 2019

Conversation

nerrad
Copy link
Contributor

@nerrad nerrad commented Feb 7, 2019

Description

This pull is the first step in moving away from the lingering usage of effects in various data stores among packages. This pull specifically will deal with the @wordpress/editor package (core/editor store).

For the most part the following goals are in mind:

  • organize file structure so things are a bit more clearly sorted.
  • As much as possible I'm preserving the existing logic. This means shifting up an action creator that returns a action.type that had effects into an action generator (i.e. savePost) and then moving the effects into the action-generator. Dispatches and selects in the effects would then be moved into the generator and yield control dispatch and selects (same for usage apiFetch etc).
  • Of course refactoring tests for all the changes.
  • This will also allow me to keep things in the existing files until the last moment when I'm ready to roll over to using the new code and test.

How has this been tested?

  • covered changes in unit tests
  • existing e2e tests coverage
  • Tested various post saves and status changes via browser testing to ensure there's no issues.

Types of changes

There are no breaking changes in this pull. The interfaces remain effectively the same. However:

  • See the note below regarding the trashPost action and some consistency introduction there. The signature has changed for trashPost but ultimately that does not affect any behaviour.
  • Some new actions are exposed to retain existing behaviour from the effects. These actions are not really something that should be publicly exposed so they are prefixed with __experimental so that provides room for later iterations potentially changing/removing them. I didn't do any refactoring there because the goal of this pull is to just convert from effects to controls/action-generators.

Checklist:

  • My code is tested.
  • My code follows the WordPress code style.
  • My code follows the accessibility standards.
  • My code has proper inline documentation.
  • I've included developer documentation if appropriate.

Additional Notes

trashPost action.

Currently trashPost accepts a postId and a postTypeSlug but they are inherently ignored by the effect callback handling the actual trash in favor of trashing the currentPost. I brought this up in slack and @aduth and I resolved that:

  • core/editor can keep trashPost but drop the args. The interface then assumes that trashPost means trash the current post in the state for the editor.
  • at some point in the future we could introduce a more explicit trashPost( postId, postType ) for the core store that would trash the given post for the given id.

Related #11124

@nerrad nerrad self-assigned this Feb 7, 2019
@nerrad nerrad added Framework Issues related to broader framework topics, especially as it relates to javascript [Status] In Progress Tracking issues with work in progress [Package] Editor /packages/editor labels Feb 7, 2019
@nerrad
Copy link
Contributor Author

nerrad commented Feb 7, 2019

Need to implement createRegistryControl from #13722 when it lands in master

@nerrad
Copy link
Contributor Author

nerrad commented Feb 8, 2019

Effects to be refactored:

  • REQUEST_POST_UPDATE
  • REQUEST_POST_UPDATE_SUCCESS
  • REQUEST_POST_UPDATE_FAILURE
  • TRASH_POST
  • TRASH_POST_FAILURE
  • REFRESH_POST

Test Coverage completed

  • REQUEST_POST_UPDATE
  • REQUEST_POST_UPDATE_SUCCESS
  • REQUEST_POST_UPDATE_FAILURE
  • TRASH_POST
  • TRASH_POST_FAILURE
  • REFRESH_POST

Following will get done in a separate pull:

  • MERGE_BLOCKS
  • SETUP_EDITOR
  • RESET_BLOCKS
  • SYNCHRONIZE_TEMPLATE
  • FETCH_REUSABLE_BLOCKS
  • SAVE_REUSABLE_BLOCK
  • DELETE_REUSABLE_BLOCK
  • RECEIVE_REUSABLE_BLOCKS
  • CONVERT_BLOCK_TO_STATIC
  • CONVERT_BLOCK_TO_REUSABLE
  • REMOVE_BLOCKS
  • REPLACE_BLOCKS
  • MULTI_SELECT

@youknowriad
Copy link
Contributor

Do you think it's possible to ship PRs for these effects one after the other?

@nerrad
Copy link
Contributor Author

nerrad commented Feb 8, 2019

I'm wary about doing multiple pulls because I'm not just refactoring I'm also re-organizing. That's a lot of potential for merge conflicts. With that said, what I could do is ship all the post related effect refactors first and then do the block related affects in a different pull. Would something like that work better?

@youknowriad
Copy link
Contributor

Would something like that work better?

Yes, If it's a big PR, it's very hard to review properly. splitting is better.

@nerrad nerrad force-pushed the refactor/switch-from-effects-to-controls branch 2 times, most recently from 0e89246 to 0fd3194 Compare February 8, 2019 20:06
@nerrad nerrad changed the title WIP: Refactor to remove usage of effects in packages/editor. Refactor to remove usage of post related effects in packages/editor. Feb 11, 2019
@nerrad
Copy link
Contributor Author

nerrad commented Feb 11, 2019

Getting close to done. However there are some failing e2e tests I need to look at (initial look shows it appears to be something involving autopost saves).

Browser testing has gone okay for me so far. If anyone wants to start reviewing this that would be fine. I'll keep looking into why the tests are failing.

@nerrad
Copy link
Contributor Author

nerrad commented Feb 11, 2019

So browser testing is working great for me. I think the issue with the e2e tests is caused by post.id not set in the savePost action triggered by autosave in the e2e tests. It's possible there's some special autosave handling in the e2e tests that is affected by this refactor (haven't dug in yet). Just making these notes for myself so I don't forget.

@nerrad
Copy link
Contributor Author

nerrad commented Feb 11, 2019

So browser testing is working great for me.

Sigh, I didn't have the GB plugin active :/ I'm reproducing issues while user testing (after actually activating the plugin with the changes) so clearly something is borked.

@nerrad
Copy link
Contributor Author

nerrad commented Feb 11, 2019

whoah, found the cause. The issue is redux-routine determines if something "quacks like an action" by the presence of a type property. The selector for getCurrentPost is return a post object which has a type property. So redux-routine considers that another action to dispatch :/

@aduth
Copy link
Member

aduth commented Feb 11, 2019

whoah, found the cause. The issue is redux-routine determines if something "quacks like an action" by the presence of a type property. The selector for getCurrentPost is return a post object which has a type property. So redux-routine considers that another action to dispatch :/

Related Slack conversation (link requires registration):

https://wordpress.slack.com/archives/C02QB2JS7/p1549904073023400

@nerrad nerrad mentioned this pull request Feb 11, 2019
9 tasks
@nerrad
Copy link
Contributor Author

nerrad commented Feb 11, 2019

I've tested the change in #13818 in this branch and it resolves the issues I was seeing here. So once that pull lands and we get it in this branch this should be good to go to (after review of course)

@nerrad nerrad force-pushed the refactor/switch-from-effects-to-controls branch from 05c7970 to 9c22491 Compare February 12, 2019 12:44
@nerrad
Copy link
Contributor Author

nerrad commented Feb 12, 2019

rats. still some e2e fails (and linting). I'm currently debugging the e2e fails locally.

@nerrad
Copy link
Contributor Author

nerrad commented Feb 12, 2019

Okay there is only one remaining failing test and it is:

The command "npm run check-local-changes" exited with 1.

I'm not sure how to resolve this. I imagine it has to do with a package-lock change when travis runs npm build. So does a package-lock change need to be generated in a commit? Should it be executed in a specific environment?

@nerrad nerrad removed the [Status] In Progress Tracking issues with work in progress label Feb 13, 2019
@@ -0,0 +1,2 @@
export * from './post-actions';
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is there a compelling reason to segregate the two?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you mean among the post actions? or do you mean specifically having "post" type actions and eventually "block" type actions?

If the former, I find it easier to know where to go for general actions as opposed to action generators. It also aids with splitting up tests to cover the specific things.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I find it easier to know where to go for general actions as opposed to action generators.

It doesn't seem to me that it's something someone should need to distinguish. I think a current reality is that we have spaghetti generators, but they're all just "actions (dispatchers)", ideally returning or yielding simple action objects.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm probably ignorant of potential reasons why we wouldn't want to split up things in this case. Are there technical/performance/build reasons why its good to avoid splitting up files like this?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I usually try to logically split files up into smaller components when I can because:

a. it helps with finding things quicker when I'm just looking for a "class" of logic
b. it makes for smaller number of lines in a file, so less scrolling.
c. My IDE built-in linting etc doesn't get bogged down as much due to larger files.

But as I already mentioned, I'm aware I may be ignorant of good reasons for not splitting things up as I did here.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nope, no directives. Just surfacing my own opinions stemming from the initial concern / probe for rationale of the split. It seems there's at least a second opinion in your favor from @youknowriad , but I'd go on record to disagree with the "domain"-based split (it wasn't but an hour after my previous complaint about the effects structuring that I wrongly navigated to effects/posts.js as my first destination in pursuit of this snippet of code).

To your question, yes, I'd see it all just being in a single actions.js file.

There's an interesting overlap between this discussion and some older ones like in #3030, where I think ultimately it was the introduction of @wordpress/data and its store-namespacing that served as the alternative to this problem of files becoming large and unwieldy, where the separation was by stores (modules), not by separate files within a store.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

actions.js is better if we had one file but if the file grows indefinitely, it becomes hard to navigate the file. maybe there's a middleground to find at some point.

I also said to @nerrad that in general I'm not very opinionated on these matters :)

Copy link
Contributor Author

@nerrad nerrad Feb 14, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

but if the file grows indefinitely, it becomes hard to navigate the file.

Ya this is one criteria I use when deciding whether to split up a file. For example just with the post actions here, combined, they would be 487 lines. That's before adding in all the block and editor related actions.

as my first destination in pursuit of this snippet of code).

@aduth my plan was to have editor-actions and block-actions files that would accompany the post-actions. That snippet is found within the SETUP_EDITOR editor effect that would have landed in editor-actions (under the original plan).

Would a middle ground here be to not split between action generators and action functions and just have actions/post.js, actions/editor.js, and actions/blocks.js? That might also be useful for aiding future refactors where some (or all) of the block specific actions are moved elsewhere.

Copy link
Contributor Author

@nerrad nerrad Feb 14, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Keep in mind, I'm not only considering the line count for a combined actions file, but also the line count that would be exponentially larger for the corresponding tests file. Currently combining actions/test/post.js and actions/test/post-generators.js into one file results in 759 lines. As soon as a file gets over roughly 500 lines I find it starts to get unwieldy. I imagine the line count for a combined file of editor, post, and block actions tests will start to go over 2000 easily.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There's an interesting overlap between this discussion and some older ones like in #3030, where I think ultimately it was the introduction of @wordpress/data and its store-namespacing that served as the alternative to this problem of files becoming large and unwieldy, where the separation was by stores (modules), not by separate files within a store.

To explore this further then, are you suggesting that we might want to consider applying some more granular namespacing here and split these up into more stores? Of course, that could introduce some complications with deprecating the old actions etc but might be preferable over the file splitting. So we could end up with something like core/editor/posts, core/editor (for general editor actions), and core/editor/blocks for block actions?

@nerrad nerrad force-pushed the refactor/switch-from-effects-to-controls branch from 87990b5 to 4cc4015 Compare February 28, 2019 13:46
@nerrad
Copy link
Contributor Author

nerrad commented Feb 28, 2019

The latest commits

  • fix spacing issues in jsdoc blocks
  • switch to use STORE_KEY instead of MODULE_KEY for the store name reference to contribute to the general renaming schema as initiated by Data: Use consistent, accurate naming for store keys #14150
  • switched autosave() generator action to use the dispatch control for dispatching the savePost action. Thus I can remove the generatorActions default export and test that the autosave() is dispatching with the right args.

@aduth
Copy link
Member

aduth commented Feb 28, 2019

  • switched autosave() generator action to use the dispatch control for dispatching the savePost action. Thus I can remove the generatorActions default export and test that the autosave() is dispatching with the right args.

While normally I'd be reluctant to alter the implementation for the sake of tests (in part the reason for the original review comment), it doesn't really strike me as any better or worse than the original implementation of the action creator to use the dispatch control this way. 👍

@nerrad
Copy link
Contributor Author

nerrad commented Feb 28, 2019

Ya I agree wrt reluctance about altering implementation for the sake of tests. But the benefit of this approach is that there isn't anything additional added for the sake of easier testing (i.e. the previous behaviour). This is a straight implementation that just happens to be more testable.

@aduth
Copy link
Member

aduth commented Feb 28, 2019

The notice generated for a failed trash when offline is a bit unexpected. But the same exists on master, so ought be addressed separately. There might already be an issue.

image

Copy link
Member

@aduth aduth left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nice work here 👍

packages/editor/src/store/actions.js Outdated Show resolved Hide resolved
*/
export const POST_UPDATE_TRANSACTION_ID = 'post-update';
const PERMALINK_POSTNAME_REGEX = /%(?:postname|pagename)%/;
export const INSERTER_UTILITY_HIGH = 3;
Copy link
Member

@aduth aduth Feb 28, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I guess these exported constants were overlooked in the split for block editor? I verified they're not used in editor, but are used (and defined within) the new block editor selectors file.

Copy link
Contributor Author

@nerrad nerrad Feb 28, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Either that or I resolved the merge incorrectly when bringing that work into this branch. I noticed it when I was pulling out constants from this file to put them in the new constants.js file.

Co-Authored-By: nerrad <[email protected]>
@nerrad nerrad merged commit c3cacd3 into master Feb 28, 2019
@nerrad nerrad deleted the refactor/switch-from-effects-to-controls branch February 28, 2019 19:26
youknowriad pushed a commit that referenced this pull request Mar 6, 2019
…13716)

This pull is the first step in moving away from the lingering usage of effects in various data stores among packages. This pull specifically deals with post related effects in the @wordpress/editor package (`core/editor` store).
youknowriad pushed a commit that referenced this pull request Mar 6, 2019
…13716)

This pull is the first step in moving away from the lingering usage of effects in various data stores among packages. This pull specifically deals with post related effects in the @wordpress/editor package (`core/editor` store).
Tug added a commit that referenced this pull request Mar 14, 2019
* Add ESNext syntax to meta block tutorial (#13954)

* Add ESNext syntax to meta block tutorial

* Applied WordPress code styles to the examples

* Apply suggestions from code review

Co-Authored-By: mkaz <[email protected]>

* Editor: RichText: Check for presence of inputType (#13986)

* Bump plugin version to 5.1.1 (#13990)

* Added a snippet for observing the browser when running e2e tests (#13993)

* Added a snippet for observing the browser when running e2e tests

* Extract reusable part of Webpack config and put in @wordpress/scripts (#13814)

* New build-config package with webpack config.

Pull the Gutenberg webpack config into a package so it can be re-used for
block/extension development.

* Require new build-config package.

* Dynamically handle WP externals with a function.

Use code from WP Calypso for handling WP externals so we don't have to have the
actual list of packages accessible in our webpack configuration.

* Use webpack config from build-config package.

* Require build-config package.

* Adjust file refs for WP packages.

* Move main gutenberg entry definition and webpack copy plugin out of build-config.

* Add react-dev-utils for formatting webpack compiler messages.

* Implement build script using webpack config from build-config.

* Adjust output path so build goes to working directory.

* Update package name to webpack-config

* Apply more tweaks to the way webpack config package is structured

* Update the way externals are handled

* Add default values for entry and output

* Move shared webpack config under @wordpress/scripts package

* Improve the way how loaders are handled

* Replace GUTENBERG with WP in webpack config env variables

Co-Authored-By: gziolo <[email protected]>

* Bring back feature flag to webpack config accidentally removed during merge

* Add missing dev dependencies for the packages used in webpack config

* Fix the list of excluded folders for babel-loader

* Use globals instead of imports in tutorials (#13995)

* Use globals instead of imports

* Update docs/designers-developers/developers/tutorials/metabox/meta-block-3-add.md

Co-Authored-By: nosolosw <[email protected]>

* Update docs/designers-developers/developers/tutorials/metabox/meta-block-3-add.md

Co-Authored-By: nosolosw <[email protected]>

* URL input popover visual cleanup (#13973)

* Use chevron instead of ellipsis in url input field options.

* Mimic toolbar icon styles for the icons in the URL popover.

* Add a left divider before the URL settings toggle

* Even up the spacing in the settings panel.

* Add periods to code comments.

* Update snapshot

* Reinstate "underline" in default formats list (#14008)

* Upgrade React to 16.8.2: Welcome React Hooks (#13992)

* Upgrade React to 16.8.2

* Update package-lock.json file

* Expose React Hooks

* Revert "Upgrade React to 16.8.2: Welcome React Hooks (#13992)" (#14017)

This reverts commit 6f88bec.

* Cleanup URL Popover stylesheet. (#14015)

As noted in #13973, this stylesheet uses a relatively non-standard SCSS method of nesting some classnames. For instance:

`.editor-url-popover { &__settings-toggle { ... } }`.

... instead of:

`.editor-url-popover__settings-toggle { ... }`

This is different from the conventions used elsewhere in Gutenberg, and is a bit more difficult to follow for that reason.

This commit un-nests those styles, and should have no effect on the compiled CSS.

* Add a generic block editor module (#13088)

* RichText: warn when using inline container (#13921)

* RichText: warn when using inline container

* Add env check

* Update documentation

* Try updating the WP database (#14048)

* Merge native mobile release v1.0 to master (#14061)

* Bump plugin version to 5.1.0-rc.1

* RichText: only ignore input types that insert HTML (#13914)

* RichText: only ignore input types that insert HTML

* Mark RichTextInputEvent as unstable

* Use Set

* Bump plugin version to 5.1.0

* Deprecate RichTextInputEvent on mobile too (#13975)

* The undelying RichText component implementation has changed the parameters returned onChange, and we forgot to update the PostTitle component (#13967)

* Fixes wrong state comparison (#13987)

Upload media progress bar is missing while media is uploading new

* Re-add rootTagsToEliminate prop (#14006)

* [Mobile]Update PostTitle to apply borders when it is focused (#13970)

* Trigger onFocusStatusChange from PostTitle

* Fix lint issue

* Update post title shadow mechanism

Also open inner ref so that focus state can be updated when focus is made programmatically

* Update props

* Update onRef as ref

* Update title padding&margin

* Mobile: Rename ref to innerRef on PostTitle (#14024)

* Fixes a red screen in mobile. (#14011)

* Change background color on image placeholder block (#14033)

* Changed upload media icon color

* Changed media placeholder background color

* Fix post title native syntax (#14041)

* Fix unexpected token in native code

* Dummy commit to trigger Travis

* Include the rnmobile release branch to Travis builds

* Mobile: Links UI using BottomSheet component (#13972)

* Mobile: Replaced Links UI with bottom-sheet component

* Mobile links UI: Removed commented code.

* Mobile: Fix lint issues

* Mobile Links UI: Remove autofocus on Android.
This hides an issue where the modal sometimes will be under the keyboard on Android.

* Fixes pasting links. (#14038)

* Update post title vertical paddings (#14040)

* Add try/catch fallback to plain text for pasteHandler (#14044)

* Fix link interface. (#14052)

* [Mobile]Fix title padding on Android (#14057)

* Remove title vertical paddings for Android

* Revert "Remove title vertical paddings for Android"

This reverts commit 09f0d35.

* Import padding variables

* Revert wrong format image color (#14058)

* Stop building the mobile release branch on Travis (#14060)

* Avoid mutating imported default config in webpack config (#14039)

* Make Babel import JSX pragma plugin aware of `wp.element.createElement` (#13809)

* Test: Add test which verifies wheter JSX pragma detects WP global

* Update packages/babel-plugin-import-jsx-pragma/test/index.js

Co-Authored-By: gziolo <[email protected]>

* Skip import when the scope variable is already defined

* Add failing tests for inner scope variable defined verification

* Add import statement when there is any undefined scope variable

* Docs: Add details about changes introduced to Babel plugin

* Upgrade Jest to version 24 (breaking changes) (#13922)

* Upgrade Jest to version 24 (breaking changes)

* Update changelog files to list all dependencies upgraded

* Downgrade puppeteer to the previous version 1.6.1

* Try to fix failing e2e tests setup

* Added clarification in the changelog

* Testing: Remove expect-puppeteer import reference

* Address issues raised during code review

* Codeowners changes for @chrisvanpatten (#14062)

* Codeowners changes for @chrisvanpatten

* Remove obsolete ghost  placeholder for no longer onowned styles

* Add repository.directory fields (fixes #13946) (#14059)

* Improve default Babel preset to include JSX pragma (#13540)

* Include custome JSX pragma support in Babel preset

* Stop using Babel tranpiliation for two packages
babel-plugin-import-jsx-pragma and postcss-themes

* Add engines field to node based packages

* Use the block editor store instead of the editor one (#13105)

* Docs: release.md: miscellaneous fixes (#14083)

* Docs: release.md: misc. fixes

* Disable block navigation and document outline items in text mode (#14081)

* Use aria-disabled instead of disabled.

* Fix unrecognized prop warning.

* Disable Document Outline items in text mode.

* Improve buttons alignment.

* Pass isTextModeEnabled as prop from parent.

* Fix selector in document outline. (#14094)

* Fix selector in document outline.

* Import getBlocks from its new location.

* Plugin: Require includes for deprecated `use_block_editor_for_` functions (#14096)

* Plugin: Fix 5.1.0 deprecated functions to correct plugin version

* Plugin: Require includes for deprecated `use_block_editor_for_` functions

* Plugin: Update server blocks script to use core equivalent function (#14097)

`gutenberg_prepare_blocks_for_js` was deprecated in Gutenberg 5.0

* Correct visual error in the quote block icon (#14091)

Fixes #13659.

The current quote block icon appears to have some over-simplified edges. This replaces it with a crisper version.

* Try: Add a subtle animation to the is-active indicator for sidebar tabs (#13956)

* Add subtle animation to the is-active indicator for sidebar tabs

* Re-instate the empty border, to prevent browser defaults from kicking in.

* Remove extra 1px of empty space

* Focus state cleanup for tab buttons

Allow them to inherit the box shadow (used for the active tab)

Use the usual standard dotted outline instead of a solid one.

* Add a pseudoclass to ensure the active border appears in Windows High Contrast Mode.

* Fix typo.

This should match the value from `/packages/edit-post/src/components/sidebar/settings-header/style.scss`

* Plugin: Remove vendor script registration (#13573)

* chore: Fix: FormToggle docs documents props the component does not uses (#14099)

* Babel Plugin Import JSX Pragma: Remove import visitor (#14106)

Props to @aduth for coding this optimization.

* Add: Block specific toolbar button sample to the format api tutorial (#14113)

## Description
This PR updates the format API tutorial (toolbar button section) to include a sample of a button that only renders on a certain block.
Answers a question posted on #14104.
Closes: #14104

## How has this been tested?
I pasted the sample code on the browser console and verified it works as expected.

* Use the editor settings to pass a mediaUpload handler (#14115)

* Use the editor settings to pass a mediaUpload handler

* Update media block snapshots

* Block Editor: Consider RECEIVE_BLOCKS as non-persistent change (#14108)

* Block Editor: Consider RECEIVE_BLOCKS as non-persistent change

* Block Editor: Compare last action in reducer enhancer only if non-ignored

* Plugin: Remove PHP functions slated for removal in 5.2 (#14090)

* Plugin: Remove PHP functions slated for removal in 5.2

* Documentation: Update FAQ to avoid reference to deprecated function

* Paste: ignore Google Docs UID tag (#14138)

* Fix  typos in copy-guide.md, readme.md and scripts.md. (#14089)

* Separate the block editor shortcuts from the post editor shortcuts (#14116)

* Fix: FocalPointPicker renders unlabelled input fields (#14152)

* Testing: Remove unnecessary Enzyme React 16 workarounds (#14156)

* Testing: Bump `enzyme-adapter-react-16` to 1.10.0

* Testing: Avoid unforwarded Button mock

No longer necessary with native support for forwardRef in Enzyme

* Testing: Un-skip BlockControls snapshot test

* Plugin: Preserve inline scripts in Gutenberg override (#13581)

* Plugin: Preserve inline scripts in Gutenberg override

* Plugin: Restore storageKey assignment for persistence migration

* Refactor to remove usage of post related effects in packages/editor. (#13716)

This pull is the first step in moving away from the lingering usage of effects in various data stores among packages. This pull specifically deals with post related effects in the @wordpress/editor package (`core/editor` store).

* Add array-callback-return rule; Fix current code breaking the rule. (#14154)

## Description
Adds a rule to make sure Array functions that iterate on the array and should return something contain the return statement, otherwise, a forEach should probably be used instead.
A case like this was fixed at #13953.

In PR #13953 @aduth suggested the implementation of a lint rule to catch these cases. While trying to implement the rule and researching the best ways to do it, I noticed a rule like that already existed in the community and this PR is enabling it.

We are also changing the code to respect the new rule no observable changes should be expected.

## How has this been tested?
Observe the tests pass.
Do some smoke testing, adding blocks, uploading files, and verify everything still works as before.

* Update text displayed when an embed can't be previewed (#13715)

* Update text displayed when an embed can't be previewed

* Add translator note for update to embedded content np preview message

* Fix: Image that is uploaded to an existing gallery does not appear in the edit gallery view (#12435)

* RichText: fix wordwise selection on Windows (#14184)

* RichText: fix wordwise selection on Windows

* Fix crtl ctrl typo.

* Add docs

* Chore: Update: Code Quality:  Remove some editor store references from block-editor (#14161)

We missed to update the editor references from selectPreviousBlock, and selectNextBlock.

## Tests

I added multiple blocks, I verified that when I remove a block the previous block still gets selected.

* Framework: Update package-lock.json for new Enzyme adapter (#14192)

* Notices: Remove inaccurate createNotice sole argument feature (#14177)

* New package to auto-generate public API documentation (#13329)

* Update nosolosw notifs (#14196)

* Make IconButton able to be referenced. (#14163)

* Make IconButton able to get referenced.

* Components: Forward IconButton ref as stateless function

* Components: Restore IconButton aria-pressed prop pass-through

* Paste: preserve empty table cells (#14137)

* Update Codeowners, add mkaz to docgen (#14198)

* Plugin: Remove wp-editor-font stylesheet override (#14176)

* Bring i18n functionality up to date (#12559)

* Plugin: Add Text Domain to plugin headers

* Bring i18n functionality up to date

* Plugin: Set Gutenberg script translations as default domain

Filter loading behavior to load from plugin translation files

* Testing: Update gutenberg_override_script per localization changes

* Plugin: Provide second argument for set script translations

Technically optional, only after WP5.1+, which is newer than the current minimum version supported by Gutenberg

* Fix: deleting the last block triggers a focus loss. (#14189)

## Description
This PR fixes a problem: if the post contains one block and it is removed the focus is lost.

This is a regression that happened when removeBlocks action was refactored to be a generator. We had an effect that inserts the default block during remove blocks action when certain conditions are met, this effect stopped working.
This PR removes the effect and makes sure everything is handled by the removeBlocks action creator.

End to end test available at #14191.

## How has this been tested?
I created a new post.
I wrote something in a paragraph I removed the paragraph using the remove button in the side menu and I verified the default block was added.

* Clarify the block editor settings from the post editor settings (#14082)

* Add repository.directory linting rule (fixes #13947) (#14200)

* Add repository.directory rule to npm-package-json-lint-config (fix #13947)

* Fix lock file (props @aduth)

* Fix rule, update changelogs

* Add PR reference

* Add directory field to packages/docgen

* Apply changes suggested by @ntwb

* Update CHANGELOG.md

* Plugin: Remove deprecated `_wpLoadGutenbergEditor`, `gutenberg` theme supports (#14144)

* Plugin: Remove deprecated `gutenberg` theme supports
* Plugin: Remove deprecated `_wpLoadGutenbergEditor`

* Bump plugin version to 5.2.0-rc.1 (#14210)

* Packages: Add missing expect-pupeteer dependency to e2e-tests package (#14212)

* Lowercase Block Editor to block editor as per the core spelling Best Practices (#14205)

## Description
I've updated all references to Block Editor to be lowercase as they aren't proper nouns. This conforms to the core spelling Best Practices;
“block editor” or “block-based editor” | “Block Editor” or “Gutenberg” | When referring to the new editor.
Reference - https://make.wordpress.org/core/handbook/best-practices/spelling/

Also ties back to the original discussion around the capitalization convention here;
#12856

Previously opened #14203 to handle the Classic Editor changes.

## How has this been tested?
Only tested that it didn't break anything.

## Types of changes
Mostly comment changes, there was one string change and a couple translator comment changes.

## Checklist:
- [x] My code is tested.
- [x] My code follows the WordPress code style. <!-- Check code: `npm run lint`, Guidelines: https://make.wordpress.org/core/handbook/best-practices/coding-standards/javascript/ -->
- [x] My code follows the accessibility standards. <!-- Guidelines: https://make.wordpress.org/core/handbook/best-practices/coding-standards/accessibility-coding-standards/ -->
- [x] My code has proper inline documentation. <!-- Guidelines: https://make.wordpress.org/core/handbook/best-practices/inline-documentation-standards/javascript/ -->
- [x] I've included developer documentation if appropriate. <!-- Handbook: https://wordpress.org/gutenberg/handbook/designers-developers/ -->

* Babel Preset Default: Avoid disabling regenerator option (#14130)

* Babel Preset Default: Remove redundant (defaulted) corejs option

* Scripts: Assign Babel runtime regenerator as externals in Webpack config

* Babel Preset Default: Avoid disabling regenerator option

* Testing: Await E2E promise interactions (#14219)

* Testing: Await E2E promise interactions

* Testing: Disable animations on every admin screen

* Testing: Refactor reusable block management to compare before/after entry count

* Lowercase classic editor when not in reference to the plugin. (#14203)

* Lowercase classic editor when not in reference to the plugin.

* Reverted heading to use capitalized as the heading choice throughout uses full capitalization style on headings so went back to make consistent as @torres126 recommended. Also updated the snap file to resolve the Travis CI error.

* Missed saving README.md where the revert on headding capitalization was in my editor.

* Fix: Calendar block: Always show current month for non post types on the editor (#13873)

* Fix allowed_block_types regression (#14229)

* Fix: Latest posts: Title is clickable across the full width of the row (#14109)

## Description
The title in latest posts block was clickable across the full width of the row, making it easy to click on the title by mistake.
This problem only affects the editor on the front end things worked as expected.

## How has this been tested?
I added the latest posts block.
I verified that the title is only clickable if the mouse is above the title and not above any part of its row.


## Screenshots <!-- if applicable -->
Before:
![feb-25-2019 21-49-15](https://user-images.githubusercontent.com/11271197/53371643-b3e3f380-3948-11e9-95f5-a8f235fabab2.gif)

After:
![feb-25-2019 21-53-30](https://user-images.githubusercontent.com/11271197/53371657-c2320f80-3948-11e9-8dfd-ab7ae4640955.gif)

* Blocks: Regenerate RSS, Search block fixtures (#14122)

* Component: date-time: Remove comparison to now when testing getMomentDate (#14230)

* RichText: Fix undo after pattern (#13917)

* Fix undo after pattern

* Update e2e test

* RichText: don't use DOM to add line padding (#13850)

* RichText: don't use DOM to add line padding

* Clean up

* Keep track of user inserted line breaks

* Mark isEditableTree param unstable

* Use a central script instead of one per package to generate docs (#14216)

* Update: Use escape key press instead of mouse movement to show block toolbar (#14247)

This PR updates our end to end tests to use escape key press instead of mouse movement to show the block toolbar. The PR follows a suggestion by @aduth in #14191 (comment)

## How has this been tested?
We only need to make sure end 2 end tests pass

* Switch the Travis badge from travis-ci.org to travis-ci.com. (#14250)

* Update the npm packages release process (#14136)

* Update plugin version to 5.2 (#14255)

* chore(release): publish

 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]

* Update the package changelogs after the npm release

* Add clickBlockToolbarButton end to end test util (#14254)

This PR adds an end 2 end test util that allows clicking in a block toolbar button and refactors existing code to use it.

## How has this been tested?
We only need to verify that the end 2 end tests pass.

* element: set up autogenerated API docs (#14269)

* Testing: Trash existing posts as admin user (#14244)

* escape-html: set up autogenerated API docs (#14268)

* html-entities: set up auto-generated API docs (#14267)

* keycodes: set up auto-generated API docs (#14265)

* Use currentColor for fill of placeholder icon—ensures icon contrasts with background color of block (#14257)

* a11y: set up auto-generated API docs (#14288)

* blob: set up auto-generated API docs (#14286)

* block-library: set up auto-generated API docs (#14282)

* compose: set up auto-generated API docs (#14278)

* dom: set up auto-generated API docs (#14273)

* i18n: set up auto-generated API docs (#14266)

* Clarify that we should rebase the branch when preparing the npm release branches (#14260)

* autop: set up auto-generated API docs (#14287)

* dom-ready: set up autogenerated API docs (#14272)

* block-editor: set up auto-generated API docs (#14285)

* Adding an e2e test verifying simple keyboard navigation through blocks (#13455)

* Adding an e2e test verifying simple keyboard navigation through blocks (Issue #12392) using bug resolved in #11773 as the basis for the steps

* Moving `navigateToContentEditorTop`, `tabThroughParagraphBlock`, `tabThroughBlockMoverControl` and `tabThroughBlockToolbar` to the parent scope. Using pressKeyWithModifier within navigateToContentEditorTop.

* Add nested blocks inside cover block (#13822)

* rich-text: set up autogenerated API docs (#14220)

* Add link to WordPress Support documentation (#14316)

* Remove users documentation (#14318)

* Fix the double dash issue (#14321)

The double dash was being converted to a single dash (&#8211;)

* docgen: Generate package docs in parallel (#14295)

* Add new actions for invalidating resolution caches (#14225)

* add new actions for invalidating resolution caches

* use path array for omit

* clarify logic

* CC-BY-3.0 is not GPLv2 compatible. (#14329)

* blocks: set up auto-generated API docs (#14279)

* deprecated: set up auto-generated API docs (#14275)

* priority-queue: set-up auto-generated API docs (#14262)

* shortcode: set up autogenerated API docs (#14218)

* viewport: set up autogenerated API docs in README (#14214)

* url: set up autogenerated API docs (#14217)

* redux-routine: set up autogenerated API docs (#14228)

* date: set up auto-generated API docs (#14276)

* block-serialization-default-parser: set up auto-generated API docs (#14280)

* RichText: collapse toolbar (#14233)

* RichText: collapse toolbar

* Move inline image

* Add balanced margin

* Add tooltip

* Update e2e test

* Update e2e test

* Remove RichTextInserterItem

* Ensure there are fills before rendering

* Make toggle button bigger

* Use clickBlockToolbarButton

* Fix e2e test

*  RichText: fix br padding in multiline and nesting (#14315)

* RichText: fix br padding in multiline and nesting

* Add e2e test

* Docs: Release: Suggest lighter SVN checkout (#14259)

* plugins: set up auto-generated API docs (#14263)

* Move the block components to the block editor module (#14112)

* wordcount: set up autogenerated API docs (#14213)

* Update packages used by eslint plugin to remove warnings (#14077)

* Update packages used by eslint plugin to remove warnings

* Manually modify package-lock.json

Props @gziolo.

* Specify React version in eslint-plugin-react settings

* Fix react/jsx-no-target lint errors

* Try Legacy widget block (#13511)

## Description
Implements: #4770

This PR is a **proof of concept** of a legacy widget block. A block that allows existing WordPress widgets to be added as Gutenberg blocks.


The design is similar to the one proposed by @melchoyce in #4770 (comment) (option 1). Although it seems option two was preferred, it would require to expose each widget as a block similar to what embeds do, and it would increase the technical complexity and make testing/debugging harder, so I preferred to use option 1 for now. I will gladly iterate on the UX after this proof of concept gets more stable.


## Some technical details

The available widgets are preloaded to Gutenberg similar to what happens with page templates.

### REST-API user story
I want to able to pass a widget identifier to an endpoint, pass the existing widget attributes and the changes the user is making if any, and receive from the rest API, the sanitized new widget attributes ready to save and an HTML form that allows the user to edit the widget in the new state.

###  REST-API endpoint
A very simple REST-API endpoint was implemented.  The endpoint receives the previous instance of a widget (previous attributes) the changed instance of a widget (changed attributes if any) and returns the new instance of the widget and the HTML form that allows editing this widget.
There are two ajax-admin endpoints save-widget and update-widget. It looks like each one has specificities that make using it here complex. The ajax admin code would probably require some changes to be used here. Our use case is straightforward from the backend perspective as the widget does not need to be saved anywhere, and the widget is not associated with any widget area. The most straightforward approach seemed to be using a very simple endpoint. If I missed something and adapting existing endpoints is simpler feel free to comment, and I will have a look.


### Block Architecture

The edit component of the block handles the start placeholder that allows selecting a widget, and the tab mechanism that allows switching between edit and preview.
The preview is done using the ServerSideRender component.
The edit is done using two components:
**WidgetEditHandler:** Is responsible for server communication using the endpoint we created, and for keeping the required local state for the edition. Renders an update button, when pressed we retrieve from the dom the changed fields (using a method provided by WidgetEditDomManager) issues a request to the server and updates the legacy widget instance attribute with the server answer.
**WidgetEditDomManager:** Component responsible for rendering the starting dom for a widget. After the first render React never rerenders the component again, the content rendered by this component is then managed by the scripts the widgets may implement. When a new instance of the form HTML is received we manually update the dom changing .widget-content (like the customizer and the widget screen) do. This component provides a method that returns an object with the widget changed attributes. When this component is mounted it triggers a widget-added event when a new update happens and the dom is changed by the component widget-update jQuery event is triggered.

On front end widget are rendered with a simple call to [the_widget](https://codex.wordpress.org/Function_Reference/the_widget).

 

## Screenshots <!-- if applicable -->

![jan-25-2019 19-37-44](https://user-images.githubusercontent.com/11271197/51768637-c3b5b100-20d8-11e9-941f-00adb4c7b0a1.gif)
![jan-25-2019 18-58-49](https://user-images.githubusercontent.com/11271197/51768645-cb755580-20d8-11e9-89e7-1aa9ba7256c3.gif)

## Known problems
- The block is not aware of any change inside the widget until the update button is pressed. This replicates the save button on the widgets screen. But it is annoying if we change something on the widget and go to the widget preview right away our changes are not reflected in the preview. Having an explicit update, makes testing and debugging easier, we may than explore other approaches e.g.: also save when preview happens, save on blur events, etc.

- The text widget that contains TinyMCE crashes and fails to init. It calls wp.editor.initialize to reference TinyMCE and on Gutenberg, wp.editor is our editor module. This problem may have happened with meta boxes if it was solved probably the same approach may be applied.

- The widget design may be affected by CSS that exists in Gutenberg, so the design of the widgets does not look the same. Ideally, Gutenberg CSS would not affect the widgets but as they are on the same page that's not the case.

- Some third-party widgets using don't initialize correctly. That happens because the dom of the editor is not equal to the dom of the customizer and/or the widget screen. Some JS widgets use click events on the widgets screen to initialize while on Gutenberg these events don't happen, some check if they are on the customizer page (body contains customizer classes) before handling widget-updated events. Normally adapting a widget that does not initialize correctly is a matter of changing a very simple condition on the plugin.

* Update changelog for esling-plugin (#14339)

* Add ability to transform [video] shortcodes to video block (#14042)

* CSS to remove the clipping of the toolbar in RTL languages (#14088)

* added the more tag via the old editor's image (#14173)

* Seimplify hierarchical term selector strings. (#13938)

* Add: End to end test to make sure when all  blocks get removed the default block appender gets inserter and selected (#14191)

* Fix the Deprecated Blocks link (#14355)

The *Deprecated Blocks link* was pointing to the wrong URL.

* Added testcases for isKeyboardEvent in keycodes module (#14073)

* Added testcases for isKeyboardEvent in keycodes module

* Fixed assertion checks and updated attachEventListeners function

* Added testcases for isKeyboardEvent in keycodes module

* Fixed assertion checks and updated attachEventListeners function

* Fix: Quote to heading transform (#14348)

* Latest Comments block: use align supports flag + code cleanup (#11411)

* ToggleControl allows setting custom classes (#13804)

* ToggleControl allows setting custom classes

Related Issue: #11349 
This pull request fixes the "ToggleControl does not allow setting custom classes" problem. This lets the user to add a custom classname on the element, just like in TextControl.

* Update README of toggle-control

Add property className and its description

* Update packages/components/src/toggle-control/README.md

Formatted style of README

Co-Authored-By: AmartyaU <[email protected]>

* Edited text of button to be actionable (#14347)

* Block library: Remove all test snapshots for blocks (#14349)

* Fix: Inserter impossible to collapse panels while searching. (#13884)

* Make taxonomies test relieable. (#14340)

* edit-post: set up autogenerated API docs (#14271)

* Plugin: Remove postinstall step (#14353)

* Plugin: Remove postinstall step

* Framework: Run build for JS unit tests

Assumed previously relied on postinstall

* Add check to the merge function in headings and paragraph blocks (#13981)

If trying to merge a blank heading or paragraph, if this check isn't performed,
the result is the string 'Null' appearing in the text.

* Update internationalization process with complete updated example. (#13909)

* Update internationalization process with complete updated example.

* Minor edits

* Update i18n package documentation

Updates with links to expanded section in Gutenberg Handbook

Fixes instructions to use wp-cli which is the recommended tool for
creating pot files, and po2json to convert the format.

* Apply suggestions from code review

Co-Authored-By: mkaz <[email protected]>

* Remove duplicate documentation, just link to Handbook

* Batch of changes from reviews props @swisspiddy

* :shakes-fist-at-whitespace:

* Updates to include full .pot and .po files per @nosolosw review

* Add JSON translation example

* Fix syntax to make gutenberg build in this branch

* E2E Test Utils: Add missing babel/runtime dependency (#14374)

* Build: Only prompt clean if unclean (#14352)

* Remove unused mobile stylesheets

* Port editor refactoring for mobile

* fix names (#14382)

* Add withRegistry HigherOrderComponent (#14370)

* Fix block validation error message (#13499)

* Rename variable to make code more clear

* Fix display of expected and actual block HTML

The variable were in the wrong order, making the message confusing. It
would show the actual HTML as the expected HTML and visa versa.

* Change wording to be more clear

Expected & Actual were confusing words to use in this content. This
change makes the message actually reflect what the values are.

* Regenerate docs

* Fix: Global inserter does not validates block insert restrictions (#14020)

* Teach build and start commands to use Webpack default if none is provided (#13877)

* Use a default webpack config if none is provided

* Extract webpack utils from build command

* Update start command

* Add docs

* Add Webpack documentation

* Add example of how to overwrite the default plugins

* Do not export hasWebpackConfig

as it is not used anywhere else.

* Always pass webpack CLI args to command

* Update README

* Remove section on extending the default webpack config file

* Simplify array passing

Co-Authored-By: nosolosw <[email protected]>

* Simplify pushing to array

Co-Authored-By: nosolosw <[email protected]>

* Fix externals docs

* Use webpack instead of Webpack

* Add Changelog entry

* RichText: Fix prepareEditableTree (#14284)

* Fix use of __experimentalCreatePrepareEditableTree without __experimentalCreateOnChangeEditableValue

* Add unit tests

* Add module entry point to notices package.json file (#14388)

* Plugin: Remove PHP functions slated for 5.3 removal (#14380)

* Make @wordpress/edit-post compatible with native

* Unify styles for UnsupportedBlock after sass build default imports is fixed on native

* Remove scss variables and colors import after fixing sass build in gb-mobile

* RichText: change value to have separate keys for line and object formats (#13948)

* Add objects and lineFormats

* Update RichText

* Fix image toolbar

* Update format placeholder

* lineFormat => lines

* concatPair => mergePair

* Update selectedFormat checks

* Add some extra info to create docs

* Move create docs inline

* Merge lines and objects

* Fix typos

* Add getActiveObject unit tests

* Update docs

* Rebase

* Adjust unstableToDom arguments

* Remove normaliseFormats from list functions

* Update native files

* Update native file

* Remove code formatting for mobile

* RichText: try alternative list shortcuts (to tab) (#14343)

* RichText: try alternative list shortcuts

* Try tooltips

* Change tooltips to use text

* Add inline comments

* Add e2e test

* Rebase

* Remove unused styles classes. (#14338)

* [Mobile]Fix placeholder position of block appender (#14386)

* Fix placeholder position of block appender

* Remove unused import

* Change import position

* Remove unused import in css file

* Remove extra container view

* Update travis.yml from master branch to fix CI issues

* [Mobile]Update caret position on insert link (#14317)

* Update selection on text change if needed

* Update caret position when editing the link

* Revert putting the caret at end when we transform the selected text into link

* Fix source map paths published to npm (#14409)

* Components: update Button readme to add design guidelines (#14194)

These changes add design documentation. Co-Authored-By: kjellr <[email protected]>

* Remove export from syntax still a proposal

* Update README.md
Tug added a commit that referenced this pull request Mar 21, 2019
* Bump plugin version to 5.1.0-rc.1

* RichText: only ignore input types that insert HTML (#13914)

* RichText: only ignore input types that insert HTML

* Mark RichTextInputEvent as unstable

* Use Set

* Bump plugin version to 5.1.0

* Deprecate RichTextInputEvent on mobile too (#13975)

* The undelying RichText component implementation has changed the parameters returned onChange, and we forgot to update the PostTitle component (#13967)

* Fixes wrong state comparison (#13987)

Upload media progress bar is missing while media is uploading new

* Re-add rootTagsToEliminate prop (#14006)

* [Mobile]Update PostTitle to apply borders when it is focused (#13970)

* Trigger onFocusStatusChange from PostTitle

* Fix lint issue

* Update post title shadow mechanism

Also open inner ref so that focus state can be updated when focus is made programmatically

* Update props

* Update onRef as ref

* Update title padding&margin

* Mobile: Rename ref to innerRef on PostTitle (#14024)

* Fixes a red screen in mobile. (#14011)

* Change background color on image placeholder block (#14033)

* Changed upload media icon color

* Changed media placeholder background color

* Fix post title native syntax (#14041)

* Fix unexpected token in native code

* Dummy commit to trigger Travis

* Include the rnmobile release branch to Travis builds

* Mobile: Links UI using BottomSheet component (#13972)

* Mobile: Replaced Links UI with bottom-sheet component

* Mobile links UI: Removed commented code.

* Mobile: Fix lint issues

* Mobile Links UI: Remove autofocus on Android.
This hides an issue where the modal sometimes will be under the keyboard on Android.

* Fixes pasting links. (#14038)

* Update post title vertical paddings (#14040)

* Add try/catch fallback to plain text for pasteHandler (#14044)

* Fix link interface. (#14052)

* [Mobile]Fix title padding on Android (#14057)

* Remove title vertical paddings for Android

* Revert "Remove title vertical paddings for Android"

This reverts commit 09f0d35.

* Import padding variables

* Revert wrong format image color (#14058)

* Stop building the mobile release branch on Travis (#14060)

* [Native mobile] Bring release v1.0.1 back to "mobile develop" (#14075)

* Don't allow placeholder to moves once we tap on it (#14066)

* Use platform based SCSS vars for block minHeight (#14070)

* Rnmobile/refactor rich text sizing code (#14164)

* Make richText height changes contained to the rich text block.

* Remove commented out code.

* Remove aztec height from state.

* Allow minHeight to be optional.

* Remove minHeight from postTitle.

* Remove minHeight on heading block.

* Mobile: Fix pasting in header (#14118)

* Fixes pasting in heading blocks.

* Fixed an issue with a missing constant.

* Removes some code that was added back as part of a merge conflict.

* Re-adds some code removed by mistake.

* Mobile: Fix links ui on landscape iOS v2 (#14240)

* RNMobile: Adding original versin of keyboard-avoiding-view (to be modified)

* Modified keyboard-avoiding-view to work well with bottom-sheets.

* Mobile: Avoid adding empty link to text. (#14270)

* Mobile Links UI: Avoid creating link with empty url field

* Mobile Links UI: Using URL Keyboard type for url field

* Fix lint issues

* Avoid to reset html to empty string if block is heading and platform is android (#14301)

* Avoid to reset html to empty string if block is heading on android platform

* Send empty tag flag from heading block to RichText

* Set minHeight based on fontSize or style. (#14344)

* Set minHeight based on fontSize or style.

* Only use the styles to set the min-height.

* Remove unused styles classes. (#14338)

* [Mobile]Fix placeholder position of block appender (#14386)

* Fix placeholder position of block appender

* Remove unused import

* Change import position

* Remove unused import in css file

* Remove extra container view

* Update travis.yml from master branch to fix CI issues

* [Mobile]Update caret position on insert link (#14317)

* Update selection on text change if needed

* Update caret position when editing the link

* Revert putting the caret at end when we transform the selected text into link

* Get the last master changes into the mobile develop branch (#14375)

* Add ESNext syntax to meta block tutorial (#13954)

* Add ESNext syntax to meta block tutorial

* Applied WordPress code styles to the examples

* Apply suggestions from code review

Co-Authored-By: mkaz <[email protected]>

* Editor: RichText: Check for presence of inputType (#13986)

* Bump plugin version to 5.1.1 (#13990)

* Added a snippet for observing the browser when running e2e tests (#13993)

* Added a snippet for observing the browser when running e2e tests

* Extract reusable part of Webpack config and put in @wordpress/scripts (#13814)

* New build-config package with webpack config.

Pull the Gutenberg webpack config into a package so it can be re-used for
block/extension development.

* Require new build-config package.

* Dynamically handle WP externals with a function.

Use code from WP Calypso for handling WP externals so we don't have to have the
actual list of packages accessible in our webpack configuration.

* Use webpack config from build-config package.

* Require build-config package.

* Adjust file refs for WP packages.

* Move main gutenberg entry definition and webpack copy plugin out of build-config.

* Add react-dev-utils for formatting webpack compiler messages.

* Implement build script using webpack config from build-config.

* Adjust output path so build goes to working directory.

* Update package name to webpack-config

* Apply more tweaks to the way webpack config package is structured

* Update the way externals are handled

* Add default values for entry and output

* Move shared webpack config under @wordpress/scripts package

* Improve the way how loaders are handled

* Replace GUTENBERG with WP in webpack config env variables

Co-Authored-By: gziolo <[email protected]>

* Bring back feature flag to webpack config accidentally removed during merge

* Add missing dev dependencies for the packages used in webpack config

* Fix the list of excluded folders for babel-loader

* Use globals instead of imports in tutorials (#13995)

* Use globals instead of imports

* Update docs/designers-developers/developers/tutorials/metabox/meta-block-3-add.md

Co-Authored-By: nosolosw <[email protected]>

* Update docs/designers-developers/developers/tutorials/metabox/meta-block-3-add.md

Co-Authored-By: nosolosw <[email protected]>

* URL input popover visual cleanup (#13973)

* Use chevron instead of ellipsis in url input field options.

* Mimic toolbar icon styles for the icons in the URL popover.

* Add a left divider before the URL settings toggle

* Even up the spacing in the settings panel.

* Add periods to code comments.

* Update snapshot

* Reinstate "underline" in default formats list (#14008)

* Upgrade React to 16.8.2: Welcome React Hooks (#13992)

* Upgrade React to 16.8.2

* Update package-lock.json file

* Expose React Hooks

* Revert "Upgrade React to 16.8.2: Welcome React Hooks (#13992)" (#14017)

This reverts commit 6f88bec.

* Cleanup URL Popover stylesheet. (#14015)

As noted in #13973, this stylesheet uses a relatively non-standard SCSS method of nesting some classnames. For instance:

`.editor-url-popover { &__settings-toggle { ... } }`.

... instead of:

`.editor-url-popover__settings-toggle { ... }`

This is different from the conventions used elsewhere in Gutenberg, and is a bit more difficult to follow for that reason.

This commit un-nests those styles, and should have no effect on the compiled CSS.

* Add a generic block editor module (#13088)

* RichText: warn when using inline container (#13921)

* RichText: warn when using inline container

* Add env check

* Update documentation

* Try updating the WP database (#14048)

* Merge native mobile release v1.0 to master (#14061)

* Bump plugin version to 5.1.0-rc.1

* RichText: only ignore input types that insert HTML (#13914)

* RichText: only ignore input types that insert HTML

* Mark RichTextInputEvent as unstable

* Use Set

* Bump plugin version to 5.1.0

* Deprecate RichTextInputEvent on mobile too (#13975)

* The undelying RichText component implementation has changed the parameters returned onChange, and we forgot to update the PostTitle component (#13967)

* Fixes wrong state comparison (#13987)

Upload media progress bar is missing while media is uploading new

* Re-add rootTagsToEliminate prop (#14006)

* [Mobile]Update PostTitle to apply borders when it is focused (#13970)

* Trigger onFocusStatusChange from PostTitle

* Fix lint issue

* Update post title shadow mechanism

Also open inner ref so that focus state can be updated when focus is made programmatically

* Update props

* Update onRef as ref

* Update title padding&margin

* Mobile: Rename ref to innerRef on PostTitle (#14024)

* Fixes a red screen in mobile. (#14011)

* Change background color on image placeholder block (#14033)

* Changed upload media icon color

* Changed media placeholder background color

* Fix post title native syntax (#14041)

* Fix unexpected token in native code

* Dummy commit to trigger Travis

* Include the rnmobile release branch to Travis builds

* Mobile: Links UI using BottomSheet component (#13972)

* Mobile: Replaced Links UI with bottom-sheet component

* Mobile links UI: Removed commented code.

* Mobile: Fix lint issues

* Mobile Links UI: Remove autofocus on Android.
This hides an issue where the modal sometimes will be under the keyboard on Android.

* Fixes pasting links. (#14038)

* Update post title vertical paddings (#14040)

* Add try/catch fallback to plain text for pasteHandler (#14044)

* Fix link interface. (#14052)

* [Mobile]Fix title padding on Android (#14057)

* Remove title vertical paddings for Android

* Revert "Remove title vertical paddings for Android"

This reverts commit 09f0d35.

* Import padding variables

* Revert wrong format image color (#14058)

* Stop building the mobile release branch on Travis (#14060)

* Avoid mutating imported default config in webpack config (#14039)

* Make Babel import JSX pragma plugin aware of `wp.element.createElement` (#13809)

* Test: Add test which verifies wheter JSX pragma detects WP global

* Update packages/babel-plugin-import-jsx-pragma/test/index.js

Co-Authored-By: gziolo <[email protected]>

* Skip import when the scope variable is already defined

* Add failing tests for inner scope variable defined verification

* Add import statement when there is any undefined scope variable

* Docs: Add details about changes introduced to Babel plugin

* Upgrade Jest to version 24 (breaking changes) (#13922)

* Upgrade Jest to version 24 (breaking changes)

* Update changelog files to list all dependencies upgraded

* Downgrade puppeteer to the previous version 1.6.1

* Try to fix failing e2e tests setup

* Added clarification in the changelog

* Testing: Remove expect-puppeteer import reference

* Address issues raised during code review

* Codeowners changes for @chrisvanpatten (#14062)

* Codeowners changes for @chrisvanpatten

* Remove obsolete ghost  placeholder for no longer onowned styles

* Add repository.directory fields (fixes #13946) (#14059)

* Improve default Babel preset to include JSX pragma (#13540)

* Include custome JSX pragma support in Babel preset

* Stop using Babel tranpiliation for two packages
babel-plugin-import-jsx-pragma and postcss-themes

* Add engines field to node based packages

* Use the block editor store instead of the editor one (#13105)

* Docs: release.md: miscellaneous fixes (#14083)

* Docs: release.md: misc. fixes

* Disable block navigation and document outline items in text mode (#14081)

* Use aria-disabled instead of disabled.

* Fix unrecognized prop warning.

* Disable Document Outline items in text mode.

* Improve buttons alignment.

* Pass isTextModeEnabled as prop from parent.

* Fix selector in document outline. (#14094)

* Fix selector in document outline.

* Import getBlocks from its new location.

* Plugin: Require includes for deprecated `use_block_editor_for_` functions (#14096)

* Plugin: Fix 5.1.0 deprecated functions to correct plugin version

* Plugin: Require includes for deprecated `use_block_editor_for_` functions

* Plugin: Update server blocks script to use core equivalent function (#14097)

`gutenberg_prepare_blocks_for_js` was deprecated in Gutenberg 5.0

* Correct visual error in the quote block icon (#14091)

Fixes #13659.

The current quote block icon appears to have some over-simplified edges. This replaces it with a crisper version.

* Try: Add a subtle animation to the is-active indicator for sidebar tabs (#13956)

* Add subtle animation to the is-active indicator for sidebar tabs

* Re-instate the empty border, to prevent browser defaults from kicking in.

* Remove extra 1px of empty space

* Focus state cleanup for tab buttons

Allow them to inherit the box shadow (used for the active tab)

Use the usual standard dotted outline instead of a solid one.

* Add a pseudoclass to ensure the active border appears in Windows High Contrast Mode.

* Fix typo.

This should match the value from `/packages/edit-post/src/components/sidebar/settings-header/style.scss`

* Plugin: Remove vendor script registration (#13573)

* chore: Fix: FormToggle docs documents props the component does not uses (#14099)

* Babel Plugin Import JSX Pragma: Remove import visitor (#14106)

Props to @aduth for coding this optimization.

* Add: Block specific toolbar button sample to the format api tutorial (#14113)

## Description
This PR updates the format API tutorial (toolbar button section) to include a sample of a button that only renders on a certain block.
Answers a question posted on #14104.
Closes: #14104

## How has this been tested?
I pasted the sample code on the browser console and verified it works as expected.

* Use the editor settings to pass a mediaUpload handler (#14115)

* Use the editor settings to pass a mediaUpload handler

* Update media block snapshots

* Block Editor: Consider RECEIVE_BLOCKS as non-persistent change (#14108)

* Block Editor: Consider RECEIVE_BLOCKS as non-persistent change

* Block Editor: Compare last action in reducer enhancer only if non-ignored

* Plugin: Remove PHP functions slated for removal in 5.2 (#14090)

* Plugin: Remove PHP functions slated for removal in 5.2

* Documentation: Update FAQ to avoid reference to deprecated function

* Paste: ignore Google Docs UID tag (#14138)

* Fix  typos in copy-guide.md, readme.md and scripts.md. (#14089)

* Separate the block editor shortcuts from the post editor shortcuts (#14116)

* Fix: FocalPointPicker renders unlabelled input fields (#14152)

* Testing: Remove unnecessary Enzyme React 16 workarounds (#14156)

* Testing: Bump `enzyme-adapter-react-16` to 1.10.0

* Testing: Avoid unforwarded Button mock

No longer necessary with native support for forwardRef in Enzyme

* Testing: Un-skip BlockControls snapshot test

* Plugin: Preserve inline scripts in Gutenberg override (#13581)

* Plugin: Preserve inline scripts in Gutenberg override

* Plugin: Restore storageKey assignment for persistence migration

* Refactor to remove usage of post related effects in packages/editor. (#13716)

This pull is the first step in moving away from the lingering usage of effects in various data stores among packages. This pull specifically deals with post related effects in the @wordpress/editor package (`core/editor` store).

* Add array-callback-return rule; Fix current code breaking the rule. (#14154)

## Description
Adds a rule to make sure Array functions that iterate on the array and should return something contain the return statement, otherwise, a forEach should probably be used instead.
A case like this was fixed at #13953.

In PR #13953 @aduth suggested the implementation of a lint rule to catch these cases. While trying to implement the rule and researching the best ways to do it, I noticed a rule like that already existed in the community and this PR is enabling it.

We are also changing the code to respect the new rule no observable changes should be expected.

## How has this been tested?
Observe the tests pass.
Do some smoke testing, adding blocks, uploading files, and verify everything still works as before.

* Update text displayed when an embed can't be previewed (#13715)

* Update text displayed when an embed can't be previewed

* Add translator note for update to embedded content np preview message

* Fix: Image that is uploaded to an existing gallery does not appear in the edit gallery view (#12435)

* RichText: fix wordwise selection on Windows (#14184)

* RichText: fix wordwise selection on Windows

* Fix crtl ctrl typo.

* Add docs

* Chore: Update: Code Quality:  Remove some editor store references from block-editor (#14161)

We missed to update the editor references from selectPreviousBlock, and selectNextBlock.

## Tests

I added multiple blocks, I verified that when I remove a block the previous block still gets selected.

* Framework: Update package-lock.json for new Enzyme adapter (#14192)

* Notices: Remove inaccurate createNotice sole argument feature (#14177)

* New package to auto-generate public API documentation (#13329)

* Update nosolosw notifs (#14196)

* Make IconButton able to be referenced. (#14163)

* Make IconButton able to get referenced.

* Components: Forward IconButton ref as stateless function

* Components: Restore IconButton aria-pressed prop pass-through

* Paste: preserve empty table cells (#14137)

* Update Codeowners, add mkaz to docgen (#14198)

* Plugin: Remove wp-editor-font stylesheet override (#14176)

* Bring i18n functionality up to date (#12559)

* Plugin: Add Text Domain to plugin headers

* Bring i18n functionality up to date

* Plugin: Set Gutenberg script translations as default domain

Filter loading behavior to load from plugin translation files

* Testing: Update gutenberg_override_script per localization changes

* Plugin: Provide second argument for set script translations

Technically optional, only after WP5.1+, which is newer than the current minimum version supported by Gutenberg

* Fix: deleting the last block triggers a focus loss. (#14189)

## Description
This PR fixes a problem: if the post contains one block and it is removed the focus is lost.

This is a regression that happened when removeBlocks action was refactored to be a generator. We had an effect that inserts the default block during remove blocks action when certain conditions are met, this effect stopped working.
This PR removes the effect and makes sure everything is handled by the removeBlocks action creator.

End to end test available at #14191.

## How has this been tested?
I created a new post.
I wrote something in a paragraph I removed the paragraph using the remove button in the side menu and I verified the default block was added.

* Clarify the block editor settings from the post editor settings (#14082)

* Add repository.directory linting rule (fixes #13947) (#14200)

* Add repository.directory rule to npm-package-json-lint-config (fix #13947)

* Fix lock file (props @aduth)

* Fix rule, update changelogs

* Add PR reference

* Add directory field to packages/docgen

* Apply changes suggested by @ntwb

* Update CHANGELOG.md

* Plugin: Remove deprecated `_wpLoadGutenbergEditor`, `gutenberg` theme supports (#14144)

* Plugin: Remove deprecated `gutenberg` theme supports
* Plugin: Remove deprecated `_wpLoadGutenbergEditor`

* Bump plugin version to 5.2.0-rc.1 (#14210)

* Packages: Add missing expect-pupeteer dependency to e2e-tests package (#14212)

* Lowercase Block Editor to block editor as per the core spelling Best Practices (#14205)

## Description
I've updated all references to Block Editor to be lowercase as they aren't proper nouns. This conforms to the core spelling Best Practices;
“block editor” or “block-based editor” | “Block Editor” or “Gutenberg” | When referring to the new editor.
Reference - https://make.wordpress.org/core/handbook/best-practices/spelling/

Also ties back to the original discussion around the capitalization convention here;
#12856

Previously opened #14203 to handle the Classic Editor changes.

## How has this been tested?
Only tested that it didn't break anything.

## Types of changes
Mostly comment changes, there was one string change and a couple translator comment changes.

## Checklist:
- [x] My code is tested.
- [x] My code follows the WordPress code style. <!-- Check code: `npm run lint`, Guidelines: https://make.wordpress.org/core/handbook/best-practices/coding-standards/javascript/ -->
- [x] My code follows the accessibility standards. <!-- Guidelines: https://make.wordpress.org/core/handbook/best-practices/coding-standards/accessibility-coding-standards/ -->
- [x] My code has proper inline documentation. <!-- Guidelines: https://make.wordpress.org/core/handbook/best-practices/inline-documentation-standards/javascript/ -->
- [x] I've included developer documentation if appropriate. <!-- Handbook: https://wordpress.org/gutenberg/handbook/designers-developers/ -->

* Babel Preset Default: Avoid disabling regenerator option (#14130)

* Babel Preset Default: Remove redundant (defaulted) corejs option

* Scripts: Assign Babel runtime regenerator as externals in Webpack config

* Babel Preset Default: Avoid disabling regenerator option

* Testing: Await E2E promise interactions (#14219)

* Testing: Await E2E promise interactions

* Testing: Disable animations on every admin screen

* Testing: Refactor reusable block management to compare before/after entry count

* Lowercase classic editor when not in reference to the plugin. (#14203)

* Lowercase classic editor when not in reference to the plugin.

* Reverted heading to use capitalized as the heading choice throughout uses full capitalization style on headings so went back to make consistent as @torres126 recommended. Also updated the snap file to resolve the Travis CI error.

* Missed saving README.md where the revert on headding capitalization was in my editor.

* Fix: Calendar block: Always show current month for non post types on the editor (#13873)

* Fix allowed_block_types regression (#14229)

* Fix: Latest posts: Title is clickable across the full width of the row (#14109)

## Description
The title in latest posts block was clickable across the full width of the row, making it easy to click on the title by mistake.
This problem only affects the editor on the front end things worked as expected.

## How has this been tested?
I added the latest posts block.
I verified that the title is only clickable if the mouse is above the title and not above any part of its row.


## Screenshots <!-- if applicable -->
Before:
![feb-25-2019 21-49-15](https://user-images.githubusercontent.com/11271197/53371643-b3e3f380-3948-11e9-95f5-a8f235fabab2.gif)

After:
![feb-25-2019 21-53-30](https://user-images.githubusercontent.com/11271197/53371657-c2320f80-3948-11e9-8dfd-ab7ae4640955.gif)

* Blocks: Regenerate RSS, Search block fixtures (#14122)

* Component: date-time: Remove comparison to now when testing getMomentDate (#14230)

* RichText: Fix undo after pattern (#13917)

* Fix undo after pattern

* Update e2e test

* RichText: don't use DOM to add line padding (#13850)

* RichText: don't use DOM to add line padding

* Clean up

* Keep track of user inserted line breaks

* Mark isEditableTree param unstable

* Use a central script instead of one per package to generate docs (#14216)

* Update: Use escape key press instead of mouse movement to show block toolbar (#14247)

This PR updates our end to end tests to use escape key press instead of mouse movement to show the block toolbar. The PR follows a suggestion by @aduth in #14191 (comment)

## How has this been tested?
We only need to make sure end 2 end tests pass

* Switch the Travis badge from travis-ci.org to travis-ci.com. (#14250)

* Update the npm packages release process (#14136)

* Update plugin version to 5.2 (#14255)

* chore(release): publish

 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]
 - @wordpress/[email protected]

* Update the package changelogs after the npm release

* Add clickBlockToolbarButton end to end test util (#14254)

This PR adds an end 2 end test util that allows clicking in a block toolbar button and refactors existing code to use it.

## How has this been tested?
We only need to verify that the end 2 end tests pass.

* element: set up autogenerated API docs (#14269)

* Testing: Trash existing posts as admin user (#14244)

* escape-html: set up autogenerated API docs (#14268)

* html-entities: set up auto-generated API docs (#14267)

* keycodes: set up auto-generated API docs (#14265)

* Use currentColor for fill of placeholder icon—ensures icon contrasts with background color of block (#14257)

* a11y: set up auto-generated API docs (#14288)

* blob: set up auto-generated API docs (#14286)

* block-library: set up auto-generated API docs (#14282)

* compose: set up auto-generated API docs (#14278)

* dom: set up auto-generated API docs (#14273)

* i18n: set up auto-generated API docs (#14266)

* Clarify that we should rebase the branch when preparing the npm release branches (#14260)

* autop: set up auto-generated API docs (#14287)

* dom-ready: set up autogenerated API docs (#14272)

* block-editor: set up auto-generated API docs (#14285)

* Adding an e2e test verifying simple keyboard navigation through blocks (#13455)

* Adding an e2e test verifying simple keyboard navigation through blocks (Issue #12392) using bug resolved in #11773 as the basis for the steps

* Moving `navigateToContentEditorTop`, `tabThroughParagraphBlock`, `tabThroughBlockMoverControl` and `tabThroughBlockToolbar` to the parent scope. Using pressKeyWithModifier within navigateToContentEditorTop.

* Add nested blocks inside cover block (#13822)

* rich-text: set up autogenerated API docs (#14220)

* Add link to WordPress Support documentation (#14316)

* Remove users documentation (#14318)

* Fix the double dash issue (#14321)

The double dash was being converted to a single dash (&#8211;)

* docgen: Generate package docs in parallel (#14295)

* Add new actions for invalidating resolution caches (#14225)

* add new actions for invalidating resolution caches

* use path array for omit

* clarify logic

* CC-BY-3.0 is not GPLv2 compatible. (#14329)

* blocks: set up auto-generated API docs (#14279)

* deprecated: set up auto-generated API docs (#14275)

* priority-queue: set-up auto-generated API docs (#14262)

* shortcode: set up autogenerated API docs (#14218)

* viewport: set up autogenerated API docs in README (#14214)

* url: set up autogenerated API docs (#14217)

* redux-routine: set up autogenerated API docs (#14228)

* date: set up auto-generated API docs (#14276)

* block-serialization-default-parser: set up auto-generated API docs (#14280)

* RichText: collapse toolbar (#14233)

* RichText: collapse toolbar

* Move inline image

* Add balanced margin

* Add tooltip

* Update e2e test

* Update e2e test

* Remove RichTextInserterItem

* Ensure there are fills before rendering

* Make toggle button bigger

* Use clickBlockToolbarButton

* Fix e2e test

*  RichText: fix br padding in multiline and nesting (#14315)

* RichText: fix br padding in multiline and nesting

* Add e2e test

* Docs: Release: Suggest lighter SVN checkout (#14259)

* plugins: set up auto-generated API docs (#14263)

* Move the block components to the block editor module (#14112)

* wordcount: set up autogenerated API docs (#14213)

* Update packages used by eslint plugin to remove warnings (#14077)

* Update packages used by eslint plugin to remove warnings

* Manually modify package-lock.json

Props @gziolo.

* Specify React version in eslint-plugin-react settings

* Fix react/jsx-no-target lint errors

* Try Legacy widget block (#13511)

## Description
Implements: #4770

This PR is a **proof of concept** of a legacy widget block. A block that allows existing WordPress widgets to be added as Gutenberg blocks.


The design is similar to the one proposed by @melchoyce in #4770 (comment) (option 1). Although it seems option two was preferred, it would require to expose each widget as a block similar to what embeds do, and it would increase the technical complexity and make testing/debugging harder, so I preferred to use option 1 for now. I will gladly iterate on the UX after this proof of concept gets more stable.


## Some technical details

The available widgets are preloaded to Gutenberg similar to what happens with page templates.

### REST-API user story
I want to able to pass a widget identifier to an endpoint, pass the existing widget attributes and the changes the user is making if any, and receive from the rest API, the sanitized new widget attributes ready to save and an HTML form that allows the user to edit the widget in the new state.

###  REST-API endpoint
A very simple REST-API endpoint was implemented.  The endpoint receives the previous instance of a widget (previous attributes) the changed instance of a widget (changed attributes if any) and returns the new instance of the widget and the HTML form that allows editing this widget.
There are two ajax-admin endpoints save-widget and update-widget. It looks like each one has specificities that make using it here complex. The ajax admin code would probably require some changes to be used here. Our use case is straightforward from the backend perspective as the widget does not need to be saved anywhere, and the widget is not associated with any widget area. The most straightforward approach seemed to be using a very simple endpoint. If I missed something and adapting existing endpoints is simpler feel free to comment, and I will have a look.


### Block Architecture

The edit component of the block handles the start placeholder that allows selecting a widget, and the tab mechanism that allows switching between edit and preview.
The preview is done using the ServerSideRender component.
The edit is done using two components:
**WidgetEditHandler:** Is responsible for server communication using the endpoint we created, and for keeping the required local state for the edition. Renders an update button, when pressed we retrieve from the dom the changed fields (using a method provided by WidgetEditDomManager) issues a request to the server and updates the legacy widget instance attribute with the server answer.
**WidgetEditDomManager:** Component responsible for rendering the starting dom for a widget. After the first render React never rerenders the component again, the content rendered by this component is then managed by the scripts the widgets may implement. When a new instance of the form HTML is received we manually update the dom changing .widget-content (like the customizer and the widget screen) do. This component provides a method that returns an object with the widget changed attributes. When this component is mounted it triggers a widget-added event when a new update happens and the dom is changed by the component widget-update jQuery event is triggered.

On front end widget are rendered with a simple call to [the_widget](https://codex.wordpress.org/Function_Reference/the_widget).

 

## Screenshots <!-- if applicable -->

![jan-25-2019 19-37-44](https://user-images.githubusercontent.com/11271197/51768637-c3b5b100-20d8-11e9-941f-00adb4c7b0a1.gif)
![jan-25-2019 18-58-49](https://user-images.githubusercontent.com/11271197/51768645-cb755580-20d8-11e9-89e7-1aa9ba7256c3.gif)

## Known problems
- The block is not aware of any change inside the widget until the update button is pressed. This replicates the save button on the widgets screen. But it is annoying if we change something on the widget and go to the widget preview right away our changes are not reflected in the preview. Having an explicit update, makes testing and debugging easier, we may than explore other approaches e.g.: also save when preview happens, save on blur events, etc.

- The text widget that contains TinyMCE crashes and fails to init. It calls wp.editor.initialize to reference TinyMCE and on Gutenberg, wp.editor is our editor module. This problem may have happened with meta boxes if it was solved probably the same approach may be applied.

- The widget design may be affected by CSS that exists in Gutenberg, so the design of the widgets does not look the same. Ideally, Gutenberg CSS would not affect the widgets but as they are on the same page that's not the case.

- Some third-party widgets using don't initialize correctly. That happens because the dom of the editor is not equal to the dom of the customizer and/or the widget screen. Some JS widgets use click events on the widgets screen to initialize while on Gutenberg these events don't happen, some check if they are on the customizer page (body contains customizer classes) before handling widget-updated events. Normally adapting a widget that does not initialize correctly is a matter of changing a very simple condition on the plugin.

* Update changelog for esling-plugin (#14339)

* Add ability to transform [video] shortcodes to video block (#14042)

* CSS to remove the clipping of the toolbar in RTL languages (#14088)

* added the more tag via the old editor's image (#14173)

* Seimplify hierarchical term selector strings. (#13938)

* Add: End to end test to make sure when all  blocks get removed the default block appender gets inserter and selected (#14191)

* Fix the Deprecated Blocks link (#14355)

The *Deprecated Blocks link* was pointing to the wrong URL.

* Added testcases for isKeyboardEvent in keycodes module (#14073)

* Added testcases for isKeyboardEvent in keycodes module

* Fixed assertion checks and updated attachEventListeners function

* Added testcases for isKeyboardEvent in keycodes module

* Fixed assertion checks and updated attachEventListeners function

* Fix: Quote to heading transform (#14348)

* Latest Comments block: use align supports flag + code cleanup (#11411)

* ToggleControl allows setting custom classes (#13804)

* ToggleControl allows setting custom classes

Related Issue: #11349 
This pull request fixes the "ToggleControl does not allow setting custom classes" problem. This lets the user to add a custom classname on the element, just like in TextControl.

* Update README of toggle-control

Add property className and its description

* Update packages/components/src/toggle-control/README.md

Formatted style of README

Co-Authored-By: AmartyaU <[email protected]>

* Edited text of button to be actionable (#14347)

* Block library: Remove all test snapshots for blocks (#14349)

* Fix: Inserter impossible to collapse panels while searching. (#13884)

* Make taxonomies test relieable. (#14340)

* edit-post: set up autogenerated API docs (#14271)

* Plugin: Remove postinstall step (#14353)

* Plugin: Remove postinstall step

* Framework: Run build for JS unit tests

Assumed previously relied on postinstall

* Add check to the merge function in headings and paragraph blocks (#13981)

If trying to merge a blank heading or paragraph, if this check isn't performed,
the result is the string 'Null' appearing in the text.

* Update internationalization process with complete updated example. (#13909)

* Update internationalization process with complete updated example.

* Minor edits

* Update i18n package documentation

Updates with links to expanded section in Gutenberg Handbook

Fixes instructions to use wp-cli which is the recommended tool for
creating pot files, and po2json to convert the format.

* Apply suggestions from code review

Co-Authored-By: mkaz <[email protected]>

* Remove duplicate documentation, just link to Handbook

* Batch of changes from reviews props @swisspiddy

* :shakes-fist-at-whitespace:

* Updates to include full .pot and .po files per @nosolosw review

* Add JSON translation example

* Fix syntax to make gutenberg build in this branch

* E2E Test Utils: Add missing babel/runtime dependency (#14374)

* Build: Only prompt clean if unclean (#14352)

* Remove unused mobile stylesheets

* Port editor refactoring for mobile

* fix names (#14382)

* Add withRegistry HigherOrderComponent (#14370)

* Fix block validation error message (#13499)

* Rename variable to make code more clear

* Fix display of expected and actual block HTML

The variable were in the wrong order, making the message confusing. It
would show the actual HTML as the expected HTML and visa versa.

* Change wording to be more clear

Expected & Actual were confusing words to use in this content. This
change makes the message actually reflect what the values are.

* Regenerate docs

* Fix: Global inserter does not validates block insert restrictions (#14020)

* Teach build and start commands to use Webpack default if none is provided (#13877)

* Use a default webpack config if none is provided

* Extract webpack utils from build command

* Update start command

* Add docs

* Add Webpack documentation

* Add example of how to overwrite the default plugins

* Do not export hasWebpackConfig

as it is not used anywhere else.

* Always pass webpack CLI args to command

* Update README

* Remove section on extending the default webpack config file

* Simplify array passing

Co-Authored-By: nosolosw <[email protected]>

* Simplify pushing to array

Co-Authored-By: nosolosw <[email protected]>

* Fix externals docs

* Use webpack instead of Webpack

* Add Changelog entry

* RichText: Fix prepareEditableTree (#14284)

* Fix use of __experimentalCreatePrepareEditableTree without __experimentalCreateOnChangeEditableValue

* Add unit tests

* Add module entry point to notices package.json file (#14388)

* Plugin: Remove PHP functions slated for 5.3 removal (#14380)

* Make @wordpress/edit-post compatible with native

* Unify styles for UnsupportedBlock after sass build default imports is fixed on native

* Remove scss variables and colors import after fixing sass build in gb-mobile

* RichText: change value to have separate keys for line and object formats (#13948)

* Add objects and lineFormats

* Update RichText

* Fix image toolbar

* Update format placeholder

* lineFormat => lines

* concatPair => mergePair

* Update selectedFormat checks

* Add some extra info to create docs

* Move create docs inline

* Merge lines and objects

* Fix typos

* Add getActiveObject unit tests

* Update docs

* Rebase

* Adjust unstableToDom arguments

* Remove normaliseFormats from list functions

* Update native files

* Update native file

* Remove code formatting for mobile

* RichText: try alternative list shortcuts (to tab) (#14343)

* RichText: try alternative list shortcuts

* Try tooltips

* Change tooltips to use text

* Add inline comments

* Add e2e test

* Rebase

* Remove unused styles classes. (#14338)

* [Mobile]Fix placeholder position of block appender (#14386)

* Fix placeholder position of block appender

* Remove unused import

* Change import position

* Remove unused import in css file

* Remove extra container view

* Update travis.yml from master branch to fix CI issues

* [Mobile]Update caret position on insert link (#14317)

* Update selection on text change if needed

* Update caret position when editing the link

* Revert putting the caret at end when we transform the selected text into link

* Fix source map paths published to npm (#14409)

* Components: update Button readme to add design guidelines (#14194)

These changes add design documentation. Co-Authored-By: kjellr <[email protected]>

* Remove export from syntax still a proposal

* Update README.md

* [RNmobile] Bring Placeholder text back working on Heading blocks (#14404)

* Remove the temporary fix for heading issue on Android, and let's pass the `tagName` value down to the Native wrapper

* Do not pass the `tagName` prop to AztecView

* The heading block now uses a minHeight defined in variables.scss, in the outer mobile project. This is a fix for an ugly rendering issue we're seeing with the placeholder when starting wrinting text.

* Add missing file from the prev commit

* Remove import of variaboe.scss since it's not required an explict import anymore after latest changes to GB.

* Add disableEditingMenu prop to manage showing editing menu on iOS (#14435)

* Fix merge errors

* Update README.md

* Fix tab linting error in scss

* [React Native] request cancel image upload (#14391)

* added hook blocks.onRemoveBlockCheckUpload called in componentWillUnmount to make sure to cancel upload if block is deleted

* removed unused import

* moving bridge-specific code to Image component - if the filter exists and returns true, the upload cancellation will be requested

* using hook actions instead of filters

* Fix lint errors

* Investigate travis error

* Line modifications for packages/block-editor/src/index.js

* Fix circle dependency

* Update README.md

* Revert travis logging
@aduth aduth mentioned this pull request Apr 24, 2019
6 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Framework Issues related to broader framework topics, especially as it relates to javascript [Package] Editor /packages/editor
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants