## The basics
- [x] I [validated my changes](https://developers.google.com/blockly/guides/contribute/core#making_and_verifying_a_change)
## The details
### Resolves
Fixes#8952Fixes#8950Fixes#8971
Fixes a bunch of other keyboard + mouse synchronization issues found during the testing discussed in https://github.com/google/blockly-keyboard-experimentation/pull/482#issuecomment-2846341307.
### Proposed Changes
This introduces a number of changes to:
- Ensure that gestures which change selection state also change focus.
- Ensure that ephemeral focus is more robust against certain classes of failures.
### Reason for Changes
There are some ephemeral focus issues that can come up with certain actions (like clicking or dragging) don't properly change focus. Beyond that, some users will likely mix clicking and keyboard navigation, so it's essential that focus and selection state stay in sync when switching between these two types of navigation modalities.
Other changes:
- Drop-down div was actually incorrectly releasing ephemeral focus before animated closes finish which could reset focus afterwards, breaking focus state.
- Both drop-down and widget divs have been updated to only return focus _after_ marking the workspace as focused since the last focused node should always be the thing returned.
- In a number of gesture cases selection has been removed. This is due to `BlockSvg` self-managing its selection state based on focus, so focusing is sufficient. I've also centralized some of the focusing calls (such as putting one in `bringBlockToFront` to ensure focusing happens after potential DOM changes).
- Similarly, `BlockSvg`'s own `bringToFront` has been updated to automatically restore focus after the operation completes. Since `bringToFront` can always result in focus loss, this provides robustness to ensure focus is restored.
- Block pasting now ensures that focus is properly set, however a delay is needed due to additional rendering changes that need to complete (I didn't dig deeply into the _why_ of this).
- Dragging has been updated to always focus the moved block if it's not in the process of being deleted.
- There was some selection resetting logic removed from gesture's `doWorkspaceClick` function. As far as I can tell, this won't be needed anymore since blocks self-regulate their selection state now.
- `FocusManager` has a new extra check for ephemeral focus to catch a specific class of failure where the browser takes away focus immediately after it's returned. This can happen if there are delay timing situations (like animations) which result in a focused node being deleted (which then results in the document body receiving focus). The robustness check is possibly not needed, but it help discover the animation issue with drop-down div and shows some promise for helping to fix the variables-closing-flyout problem.
Some caveats:
- Some undo/redo steps can still result in focus being lost. This may introduce some regressions for selection state, and definitely introduces some annoyances with keyboard navigation. More work will be needed to understand how to better redirect focus (and to what) in cases when blocks disappear.
- There are many other places where focus is being forced or selection state being overwritten that could, in theory, cause issues with focus state. These may need to be fixed in the future.
- There's a lot of redundancy with `focusNode` being called more than it needs to be. `FocusManager` does avoid calling `focus()` more than once for the same node, but it's possible for focus state to switch between multiple nodes or elements even for a single gesture (for example, due to bringing the block to the front causing a DOM refresh). While the eventual consistency nature of the manager means this isn't a real problem, it may cause problems with screen reader output in the future and warrant another pass at reducing `focusNode` calls (particularly for gestures and the click event pipeline).
### Test Coverage
This PR is largely relying on existing tests for regression verification, though no new tests have been added for the specific regression cases.
#8910 is tracking improving `FocusManager` tests which could include some cases for the new ephemeral focus improvements.
#8915 is tracking general accessibility testing which could include adding tests for these specific regression cases.
### Documentation
No new documentation is expected to be needed here.
### Additional Information
These changes originate from both #8875 and from a branch @rachel-fenichel created to investigate some of the failures this PR addresses. These changes have also been verified against both Core's playground and the keyboard navigation plugin's test environment.
## The basics
- [x] I [validated my changes](https://developers.google.com/blockly/guides/contribute/core#making_and_verifying_a_change)
## The details
### Resolves
Fixes#8922Fixes#8929
Fixes part of #8771
### Proposed Changes
This PR introduces support for fields to be focusable (and thus navigable with keyboard navigation when paired with downstream changes to `LineCursor` and the keyboard navigation plugin). This is a largely isolated change in how it fundamentally works:
- `Field` was updated to become an `IFocusableNode`. Note that it uses a specific string-based ID schema in order to ensure that it can be properly linked back to its unique block (which helps make the search for the field in `WorkspaceSvg` a bit more efficient). This could be done with a globally unique ID, instead, but all fields would need to be searched vs. just those for the field's parent block.
- The drop-down and widget divs have been updated to manage ephemeral focus with `FocusManager` when they're open for non-system dialogs (ephemeral focus isn't needed for system dialogs/prompts since those already take/restore focus in a native way that `FocusManager` will respond to--this may require future work, however, if the restoration causes unexpected behavior for users). This approach was done due to a suggestion from @maribethb as the alternative would be a more complicated breaking change (forcing `Field` subclasses to properly manage ephemeral focus). It may still be the case that certain cases will need to do so, but widget and drop-down divs seem to address the majority of possibilities.
**Important**: `Input`s are not explicitly being supported here. As far as I can tell, we can't run into a case where `LineCursor` tries to set an input node, though perhaps I simply haven't come across this case. Supporting `Fields` and `Connections` (per #8928) seems to cover the main needed cases, though making `Input`s focusable may be a future requirement.
### Reason for Changes
This is part of an ongoing effort to ensure key components of Blockly are focusable so that they can be keyboard-navigable (with other needed changes yet both in Core Blockly and the keyboard navigation plugin).
Note that #8929 isn't broadly addressed since making widget & drop down divs manage ephemeral focus directly addresses a large class of cases. Additional cases may arise where a plugin or Blockly integration may require additional effort to make keyboard navigation work for their field--this may be best addressed with documentation and guidance.
### Test Coverage
No new tests have been added. It's certainly possible to add unit tests for the focusable configurations being introduced in this PR, but it may not be highly beneficial. It's largely assumed that the individual implementations should work due to a highly tested FocusManager, and it may be the case that the interactions of the components working together is far more important to verify (that is, the end user flows). The latter is planned to be tackled as part of #8915.
### Documentation
No new documentation is planned, however it may be prudent to update the field documentation in the future to explain how to utilize ephemeral focus when specifically building compatibility for keyboard navigation.
### Additional Information
This includes changes that have been pulled from #8875.
* fix: Fix bug that prevented dismissing the widgetdiv in a mutator workspace.
* fix: Check if the correct workspace is null.
* fix: Remove errant this.
* chore(deps): Add pretter-plugin-organize-imports
* chore: Remove insignificant blank lines in import sections
Since prettier-plugin-organize-imports sorts imports within
sections separated by blank lines, but preserves the section
divisions, remove any blank lines that are not dividing imports
into meaningful sections.
Do not remove blank lines separating side-effect-only imports
from main imports.
* chore: Remove unneded eslint-disable directives
* chore: Organise imports
* fix: assign to containerDiv if it already exists
* Delete unused return
* Fix if statement
* format file using Prettier
* Deduplicate lines and improve efficiency
* fix: recreate dropdown div
* fix: recreate tooltip div
* fix: recreate widget div
* fix: Stop using global variables and use getDiv()
* Revert "fix: Stop using global variables and use getDiv()"
This reverts commit 322c4d884b.
* fix: reset widget div
* fix(build): Restore erroneously-deleted filter function
This was deleted in PR #7406 as it was mainly being used to
filter core/ vs. test/mocha/ deps into separate deps files -
but it turns out also to be used for filtering error
messages too. Oops.
* refactor(tests): Migrate advanced compilation test to ES Modules
* refactor(build): Migrate main.js to TypeScript
This turns out to be pretty straight forward, even if it would
cause crashing if one actually tried to import this module
instead of just feeding it to Closure Compiler.
* chore(build): Remove goog.declareModuleId calls
Replace goog.declareModuleId calls with a comment recording the
former module ID for posterity (or at least until we decide
how to reformat the renamings file.
* chore(tests): Delete closure/goog/*
For the moment we still need something to serve as base.js for
the benefit of closure-make-deps, so we keep a vestigial
base.js around, containing only the @provideGoog declaration.
* refactor(build): Remove vestigial base.js
By changing slightly the command line arguments to
closure-make-deps and closure-calculate-chunks the need to have
any base.js is eliminated.
* chore: Typo fix for PR #7415
* fix: stop using dom.addClass in most cases
* chore: format
* fix: remove use of dom.addClass in toolbox
* chore: lint and format
* fix: add checks around non-constant class names
* fix: switch back to quoted access
* chore: format
* fix: remove calls to removeClass
* chore: format
* chore: remove unused deps
* fix: remove uses of hasClass
* chore: format and lint
* chore: format
* chore: add linting for tsdoc
* chore: don't require types on return
* chore: remove redundant fileoverview from ts
* chore: change return to returns and add some newlines
* chore: remove license tag
* chore: don't require params/return docs
* chore: remove spurious struct tags
* Revert "chore: change return to returns and add some newlines"
This reverts commit d6d8656a45.
* chore: don't auto-add param names
* chore: disable require-param bc it breaks on this
* return to returns and add line breaks
* chore: configure additional jsdoc rules
* chore: run format
* Revert "chore: remove license tag"
This reverts commit 173455588a.
* chore: allow license tag format
* chore: only require jsdoc on exported items
* chore: add missing jsdoc or silence where needed
* chore: run format
* chore: lint fixes
* fix: convert files to typescript
* fix: add alias for AnyDuringMigration so that tsc will run
* chore: format
* chore: enable ts for the clang-format workflow (#6233)
* chore: Restore @fileoverview comment locations (#6237)
* chore: add declareModuleId (#6238)
* fix: Revert comment change to app_controller.js (#6241)
* fix: Add missing import goog statements (#6240)
I've added the import statement immediately before the
goog.declareModuleId calls that depend on it.
There is an argument to be made that we should put the import
statement in their normal place amongst any other imports, and
move the declareModuleId statement to below the double blank
line below the imports, but as these are so tightly coupled,
replace the previous goog.module calls, and will both be deleted
at the same time once the transition to TypeScript is fully complete
I think it's fine (and certainly much easier) to do it this way.
* chore: Fix whitespace (#6243)
* fix: Remove spurious blank lines
Remove extraneous blank lines introduced by deletion of
'use strict'; pragmas.
Also fix the location of the goog.declareModuleId call in
core/utils/array.ts.
* fix: Add missing double-blank-line before body of modules
Our convention is to have two blank lines between the imports (or
module ID, if there are no imports) and the beginning of the body
of the module. Enforce this.
* fix: one addition format error for PR #6243
* fix(build): Skip npm prepare when running in CI (#6244)
Have npm prepare do nothing when running in CI.
We don't need to do any building, because npm test will build
everything needed in the workflows in which it is run, and we
don't want to build anything in other workflows because a tsc
error would prevent those workflows from completing.
* fix: re-add `@package` annotations as `@internal` annotations (#6232)
* fix: add ~70% of internal attributes
* fix: work on manually adding more @internal annotations
* fix: add more manual internal annotations
* fix: rename package typos to internal
* fix: final manual fixes for internal annotations
* chore: format
* chore: make unnecessary multiline jsdoc a single line
* fix: fix internal tags in serialization exceptions
* fix: tsc errors picked up from develop (#6224)
* fix: relative path for deprecation utils
* fix: checking if properties exist in svg_math
* fix: set all timeout PIDs to AnyDuringMigration
* fix: make nullability errors explicity in block drag surface
* fix: make null check in events_block_change explicit
* fix: make getEventWorkspace_ internal so we can access it from CommentCreateDeleteHelper
* fix: rename DIV -> containerDiv in tooltip
* fix: ignore backwards compat check in category
* fix: set block styles to AnyDuringMigration
* fix: type typo in KeyboardShortcut
* fix: constants name in row measurables
* fix: typecast in mutator
* fix: populateProcedures type of flattened array
* fix: ignore errors related to workspace comment deserialization
* chore: format files
* fix: renaming imports missing file extensions
* fix: remove check for sound.play
* fix: temporarily remove bad requireType.
All `export type` statements are stripped when tsc is run. This means
that when we attempt to require BlockDefinition from the block files, we
get an error because it does not exist.
We decided to temporarily remove the require, because this will no
longer be a problem when we conver the blocks to typescript, and
everything gets compiled together.
* fix: bad jsdoc in array
* fix: silence missing property errors
Closure was complaining about inexistant properties, but they actually
do exist, they're just not being transpiled by tsc in a way that closure
understands.
I.E. if things are initialized in a function called by the constructor,
rather than in a class field or in the custructor itself, closure would
error.
It would also error on enums, because they are transpiled to a weird
IIFE.
* fix: context menu action handler not knowing the type of this.
this: TypeX information gets stripped when tsc is run, so closure could
not know that this was not global. Fixed this by reorganizing to use the
option object directly instead of passing it to onAction to be bound to
this.
* fix: readd getDeveloperVars checks (should not be part of migration)
This was found because ALL_DEVELOPER_VARS_WARNINGS_BY_BLOCK_TYPE was no
longer being accessed.
* fix: silence closure errors about overriding supertype props
We propertly define the overrides in typescript, but these get removed
from the compiled output, so closure doesn't know they exist.
* fix: silence globalThis errors
this: TypeX annotations get stripped from the compiled output, so
closure can't know that we're accessing the correct things. However,
typescript makes sure that this always has the correct properties, so
silencing this should be fine.
* fix: bad jsdoc name
* chore: attempt compiling with blockly.js
* fix: attempt moving the import statement above the namespace line
* chore: add todo comments to block def files
* chore: remove todo from context menu
* chore: add comments abotu disabled errors
* chore: move comments back to their correct positions (#6249)
* fix: work on fixing comments
* chore: finish moving all comments
* chore: format
* chore: move some other messed up comments
* chore: format
* fix: Correct enum formatting, use merged `namespace`s for types that are class static members (#6246)
* fix: formatting of enum KeyCodes
* fix: Use merged namespace for ContextMenuRegistry static types
- Create a namespace to be merged with the ContextMenuRegistry
class containing the types that were formerly declared as static
properties on that class.
- Use type aliases to export them individually as well, for
compatibility with the changes made by MigranTS (and/or
@gonfunko) to how other modules in core/ now import these
types.
- Update renamings.json5 to reflect the availability of the
direct exports for modules that import this module directly
(though they are not available to, and will not be used by,
code that imports only via blockly.js/blockly.ts.)
* fix: Use merged namespace for Input.Align
- Create a merged namespace for the Input.Align enum.
- Use type/const aliases to export it as Input too.
- Update renamings.json5 to reflect the availability of the
direct export.
* fix: Use merged namespace for Names.NameType
- Create a merged namespace for the Names.NameType enum.
- Use type/const aliases to export it as NameType too.
- Update renamings.json5 to reflect the availability of the
direct export. (This ought to have happened in an earlier
version as it was already available by both routes.)
* chore: Fix minor issues for PR #6246
- Use `Align` instead of `Input.Align` where possible.
* fix(build): Suppress irrelevant JSC_UNUSED_LOCAL_ASSIGNMENT errors
tsc generates code for merged namespaces that looks like:
(function (ClassName) {
let EnumName;
(function (EnumName) {
EnumName[EnumNameAlign["v1"] = 0] = "v1";
// etc.
})(EnumName = ClassName.EnumName || (ClassName.EnumName = {}));
})(ClassName || (ClassName = {}));
and Closure Compiler complains about the fact that the EnumName let
binding is initialised but never used. (It exists so that any other
code that was in the namespace could see the enum.)
Suppress this message, since it is not actionable and lint and/or tsc
should tell us if we have actual unused variables in our .ts files.
* chore(build): Suppress spurious warnings from closure-make-deps (#6253)
A little bit of an ugly hack, but it works: pipe stderr through
grep -v to suppress error output starting with "WARNING in".
* fix: remaining enums that weren't properly exported (#6251)
* fix: remaining enums that weren't properly exported
* chore: format
* fix: add enum value exports
* chore: format
* fix: properly export interfaces that were typedefs (#6250)
* fix: properly export interfaces that were typedefs
* fix: allowCollsion -> allowCollision
* fix: convert unconverted enums
* fix: enums that were/are instance properties
* fix: revert changes to property enums
* fix: renamed protected parameter properties (#6252)
* fix: bad protected parameter properties
* chore:format
* fix: gesture constructor
* fix: overridden properties that were renamed
* refactor: Migrate `blockly.js` to TypeScript (#6261)
* chore: Apply changes to blockly.js to blockly.ts
* fix: Build using core/blockly.ts instead of .js
Compiles and runs in compressed mode correctly!
* fix(build): Don't depend on execSync running bash (#6262)
For some reason on Github CI servers execSync uses /bin/sh, which
is (on Ubuntu) dash rather than bash, and does not understand
the pipefail option.
So remove the grep pipe on stderr and just discard all error output
at all.
This is not ideal as errors in test deps will go unreported AND
not even cause test failure, but it's not clear that it's worth
investing more time to fix this at the moment.
* chore: use `import type` where possible (#6279)
* chore: automatically change imports to import types
* chore: revert changes that actually need to be imports
* chore: format
* chore: add more import type statements based on importsNotUsedAsValues
* chore: fix tsconfig
* chore: add link to compiler issue
* fix: add type information to blockly options (#6283)
* fix: add type information to blockly options
* chore: format
* chore: remove erroneous comment
* fix: bugs revealed by getting the built output working (#6282)
* fix: types of compose and decompose in block
* fix: workspace naming in toolbox
* chore: add jsdoc
* chore: restore registry comments to better positions
* chore: pr comments'
* fix(variables): Revert inadvertent change to allDeveloperVariables (#6290)
It appears that a function call got modified incorrectly (probably
in an effort to fix a typing issue). This fix trivially reverts
the line in question to match the original JS version from develop.
This causes the generator tests to pass.
* fix: circular dependencies (#6281)
* chore: fix circular dependencies w/ static workspace funcs
* remove preserved imports that aren't currently necessary (probably)
* fix circular dependency with workspaces and block using stub
* fix dependency between variables and xml by moving function to utils
* add stub for trashcan as well
* fix line endings from rebase
* fix goog/base order
* add trashcan patch
* fix: types of compose and decompose in block
* fix: workspace naming in toolbox
* chore: add jsdoc
* chore: restore registry comments to better positions
* chore: remove implementations in goog.js
* chore: fix types of stubs
* chore: remove added AnyDuringMigration casts
* chore: remove modifications to xml and variables
* chore: format
* chore: remove event requirements in workspace comments
* chore: fix circular dependency with xml and workspace comments
* fixup remove ContextMenu import
* chore: fix dependency between mutator and workspace
* chore: break circular dependency between names and procedures
* chore: get tests to run?
* chore: pr comments'
* chore: fix stubbing field registry fromJson
* chore: fix spying on fire
* chore: fix stubbing parts of connection checker
* chore: fix stubbing dialog
* chore: fix stubbing style
* chore: fix spying on duplicate
* chore: fix stubbing variables
* chore: fix stubbing copy
* chore: fix stubbing in workspace
* chore: remove unnecessary stubs
* chore: fix formatting
* chore: fix other formatting
* chore: add backwards compatible static properties to workspace
* chore: move static type properties
* chore: move and comment stubs
* chore: add newlines at EOF
* chore: improve errors for monkey patched functions
* chore: update comment with a pointer to the doc
* chore: update comment with a pointer to the doc
* chore: format
* chore: revert changes to playground used for testing (#6292)
* chore: get mocha tests to pass. (#6291)
* chore: fix undo and empty code blocks
* chore: skip IE test
* chore: fix gesture test
* chore: fix replace message references test
* chore: fix string table interpolation
* chore: skip getById tests
* chore: fix field tests
* chore: fix console errors by making workspace nullable
* chore: format
* chore: fix definition overwrite warning
* chore: update metadata
* chore: temporarily modify the the advanced compile test
* chore: fix gestures by fixing test instead
Co-authored-by: Neil Fraser <fraser@google.com>
Co-authored-by: Christopher Allen <cpcallen+git@google.com>
