Copying nodes:
- Multiple nodes supported.
- Node comments and user-specified colors included.
- Google Sheets data can be pasted to produce a `Table` node, handled the same way as Excel data.
# Important Notes
- Fix E2E tests on OS X.
- Add E2E and unit tests for clipboard.
- Use the lexer to test text escaping; fix text escaping issues and inconsistencies.
Fixes#8520
If the websocket is closed not by us, we automatically try to reconnect with it, and initialize the protocol again. **Restoring state (execution contexts, attached visualizations) is not part of this PR**.
It's a part of making IDE work after hibernation (or LS crash).
# Important Notes
It required somewhat heavy refactoring:
1. I decided to use an existing implementation of reconnecting websocket. Replaced (later discovered by me) our implementation.
2. The LanguageServer class now handles both reconnecting and re-initializing - that make usage of it simpler (no more `Promise<LanguageServer>` - each method will just wait for (re)connection and initialization.
3. The stuff in `net` src's module was partially moved to shared's counterpart (with tests). Merged `exponentialBackoff` implementations, which also brought me to
4. Rewriting LS client, so it returns Result instead of throwing, what is closer our desired state, and allows us using exponentialBackoff method without any wrappers.
Fixes on of the issues in #9354
Stale method call info for inner sub-application was causing additional argument placeholders on the node for certain expressions. Now it is fixed:
1. We only create function widget for the most top-level expression in the prefix application chain.
2. We reuse method call info from inner expressions, assuming it will be always correct for our purposes.
https://github.com/enso-org/enso/assets/6566674/91d2b4ba-a789-4c7b-b40c-f09ac45da7f0
In GUI E2E `widgets.spec.ts`, be more specific about which tested dropdowns should be visible. Since the tested node actually receives two dropdowns, the test could fail by detecting both and expecting only one. Since we already have a very specific selector for the dropdown we are testing, use it instead of fresh global locator.
Fixes#9492
Implemented generic component for flawless size-based transitions, then used it for hiding arguments and dropdown animation. That replaced the `max-size`-based CSS animation that caused original issue. Refactored dropdown positioning to avoid further issues related to animation overflow clipping. The dropdown also got a bit of a lift to fit closer to styles in current Figma designs.
https://github.com/enso-org/enso/assets/919491/e85fd68c-b2e8-4d58-90e1-4fd7b33f1c9b
Prevent interactions such as an open dropdown from being disrupted by the temporary state of absent type information occurring after an edit.
https://github.com/enso-org/enso/assets/1047859/bcc7fa02-847f-4cd0-b9fc-3186a97c5816
All widget-edit interactions that are active when a component is to be unmounted save state and are suspended. When a new component defining a `WidgetEditHandler` is instantiated, if the component is found to be equivalent to a component that was suspended, and no other interaction has been initiated in the interim, the interaction is restarted using the suspended state.
# Important Notes
- Fix a bug caused by a variable tracking an interaction's active state getting out of sync with the interaction. `WidgetEditHandler` now provides a reactive `active` property; using it is simpler and avoids this type of bug in the future.
- Fix a flickering bug that sometimes made it hard to open dropdowns by clicking the arrow.
Fixed a long-standing annoyance that widgets weren't fully hot-reloadable. Now when the widget definition (e.g. `score` function) is modified, it is hot-reloaded and new version immediately takes effect.
https://github.com/enso-org/enso/assets/919491/8e6d5a67-68ec-4353-8235-32657b32e2ec
# Important Notes
Because of how HMR API works, it needs to be passed from each widget module to the `widgetDefinition` function as an argument. When not provided, the definition will not be hot-reloadable (but the widget will still work as it used to).
Fixes#6250
With this change, I've also slightly refactored the graph editor component by grouping related functionality into neat block and moving already loosely coupled groups to separate files. Further work will be needed to simplify it, but it is a good first step.
https://github.com/enso-org/enso/assets/919491/fedce111-ea79-463f-a543-da3ecce28bf5
When a node has an error/warning/panic that exactly matches one of its input nodes, hide the message until the node is interacted with, showing an icon.
https://github.com/enso-org/enso/assets/1047859/4b1b5e3d-c236-40d7-a3e7-e6ab8182ecd5
# Important Notes
- New icon is used for panics.
- Opening circular menu now shifts any message out of the way, not just warnings.
- Close#8610
# Important Notes
QA notes:
- The GUI2 warning screen should not show up - the arguments that GUI2 do not understand have been removed.
- However, it should be tested that the warnings screen should correctly work when invalid arguments really *are* passed in:
- Via URL query parameters (electron, might need to open the electron app then the browser, *or* do `location.href = ` in DevTools in Electron.)
- By editing `Editor.tsx` to inject invalid args to the big configuration object we pass to the GUI entrypoint.
Fixes the issue with numeric input being unclickable if the WidgetSelection is present. The issue was caused by double handling of the click event both in numeric input (it opened the dropdown) and in dropdown (it closed itself).
Closes#8680
The color picker is triggered by new Circular menu item or by keyboard shortcut (set to `Mod+Shift+C` for **C**olor).
The color picker affects *selected* nodes, so it works nicely both in cases when we use Circular menu on a single node a shortcut with multiple selected nodes.
Color is only changed when the user selects a new one inside the color picker, but there is no specific way to reset default color (you would have to recreate the node).
https://github.com/enso-org/enso/assets/6566674/a4497bea-f432-486d-85f8-f2772ba5694f
Fixes#9314
The node deletion does not remove AST node from the module, only unpin it from its parent; so undoing does not add this node, just modify it, and thus we weren't informed about metadata change.
While blocked FQN substitution when creating/editing node, the drop-downs still used it, and it does not work.
This amends that + added a reference to a task where we want to fix the original issue.
Move the types from `Standard.Table.Data` to `Standard.Table`.
Exceptions:
- `Standard.Table.Data.Report_Unmatched` => `Standard.Table.Constants`.
- `Standard.Table.Data.Join_Kind_Cross` => `Standard.Table.Internal.Join_Kind_Cross`.
Also removed constructor as an atom type.
- `Standard.Table.Extensions.Table_Ref` => `Standard.Table.Internal.Table_Ref`.
- `Standard.Table.Data.Type.Value_Type_Helpers` => `Standard.Table.Internal.Value_Type_Helpers`.
- `Standard.Table.Data.Type.Enso_Types` => `Standard.Table.Internal.Value_Type_Helpers`.
- `Standard.Table.Data.Type.Storage` => `Standard.Table.Internal.Storage`.
Changed all `Standard.Table` imports inside project to be project.
Favoured importing from `Standard.Table.Main` in `Standard.Database`.
Also fixed some linting in Enso_File.
- Implement https://github.com/enso-org/cloud-v2/issues/961
- Allow directory management on Local backend
- Setting a custom root directory is currently *out of scope* of this PR.
- Listing directories
- Deleting directories and files
- Adjust project-related APIs to accept parent directory path (as required by PM when not interacting with root directory)
- QoL improvements related to testing this PR
- New watch script (`npm run watch2`, `npm run watch:linux`) in `app/ide-desktop/lib/client/`) for testing IDE2 on Electron without having to build the entire app
- Adjustments to `gui2`'s `vite.config.ts` to allow React HMR when doing dev in Electron
# Important Notes
- Support for deleting files and folders uses the API introduced by #9359 - so it will not work until that PR is merged in.
- Support for uploading files uses the API specified by #9360 - so it will not work until that issue is closed.
Fixes#9313
[Screencast from 2024-03-22 09-09-07.webm](https://github.com/enso-org/enso/assets/3919101/6ad86145-6882-4bde-993d-b1270f1ec06c)
# Important Notes
* This is PoC, so I didn't spend time on polishing the visuals; the design will likely change.
* I modified the shortcut handler a bit, allowing making multiple actions for same binding - the action's handler will be called in unspecified order, until one of them handle the event (i.e. not return false).
* To make it working regardless of imports, I needed to export AI module in Standard.Visualization. Moreover, needed to remove build_ai_prompt for Any, because it was causing issues - expect a bug report soon.
This PR updates the Rust toolchain to recent nightly.
Most of the changes are related to fixing newly added warnings and adjusting the feature flags. Also the formatter changed its behavior slightly, causing some whitespace changes.
Other points:
* Changed debug level of the `buildscript` profile to `lint-tables-only` — this should improve the build times and space usage somewhat.
* Moved lint configuration to the worksppace `Cargo.toml` definition. Adjusted the formatter appropriately.
* Removed auto-generated IntelliJ run configurations, as they are not useful anymore.
* Added a few trivial stdlib nightly functions that were removed to our codebase.
* Bumped many dependencies but still not all:
* `clap` bump encountered https://github.com/clap-rs/clap/issues/5407 — for now the warnings were silenced by the lint config.
* `octocrab` — our forked diverged to far with the original, needs more refactoring.
* `derivative` — is unmaintained and has no updated version, despite introducing warnings in the generated code. There is no direct replacement.
One of our tests had transient failures, which were very tricky to track - it seems that not every `mouse.move` action actually emits 'pointerleave' action, on which we rely when hiding node menu.
I played with different configurations, and this one looks quite reliable (100 runs of tests in a row passed, at least). But I still don't understand what is happening there, and the investigation will be in scope of #9478
Hide scrollbars until track area is hovered
- Hovering track area now shows a track
- Clicking in track moves center of scrollbar to position
(Small change requested by @wdanilo)
Fixes#9379.
# Important Notes
- The existence-registry could be implemented with a counter, but a set is more debuggable and the performance cost is negligible.
Fixes#9058 - the filtering so far is a bit aggressive, but I tune it up in next PR(s).
[Screencast from 2024-03-13 15-20-17.webm](https://github.com/enso-org/enso/assets/3919101/112ce65a-a8c6-4818-b8b8-9f493caf9c81)
Added new special `WidgetEditHandler,` allowing handling "multi-widget" interactions needed for drop down filtering.
# Important Notes
* Now when clicking on argument name, the edit is accepted (as normal "outside" click), and then the dropdown is opened again (due to handling click event). I didn't figure out how to handle this case properly, left something least confusing.
Closes#9209https://github.com/enso-org/enso/assets/6566674/43984978-004b-4ed2-a55e-de9302dc1fdd
- Added new menu item `Code editor` which toggles visibility of, well, code editor
- Changed the menu implementation:
- Fixed platform-dependent “More” icon. It was displayed awfully on Mac.
- Added simple opacity animation.
- Reduced width slightly.
- Refactored the layout, removed unused classes and CSS…
- Added animated background when hovering “Show all” icon
- Added close button to the upper left corner of the code editor (upper right is occupied by resize controls)
This is a set of split off changes made as a side effect while working on engine reconnection handling.
Cleaned up GUI e2e setup, unified as much of the entrypoint code as possible. Currently the only real difference between the real and testing entrypoint is mocking of all network calls and not loading through dashboard.
I've managed to completely get rid of `MockApp`, and remove tricky mocking of pinia stores.
Some e2e tests (about leaving nodes) were always failing on my machine because of the race condition in the tests. The issue was caused by edges positions lagging behind for a few frames when switching Enso functions, which caused incorrect handling of clicks in the test code.
Now we wait for edges being initialized *and* node sizes being updated.
Thanks to @farmaazon for helping with debugging.
This PR:
* removes much of logic related to building and packaging the gui1;
* made `./run gui` and `./run ide` work with the new gui;
* rename numerous references to the "gui2" or "new gui" in favor of simply "gui", same for "ide".
* Comment UI refinement
- Eliminate edit mode:
- Beginning edit does not change appearance
- Edit text as rendered, not as formatted in code
- Enter finishes edit
- Shift+Enter inserts a newline
- Click begins edit regardless of Ctrl
Fixes#9357
The main issue was the spread operator using at the wrong place in functions overriding spacing of nodes. The bug, to be visible, required copying AST node before, because during copying `whitespace` field was explicitly set to undefined (in opposite to being unset), what in turns make spread overriding the value set by those functions.
# Important Notes
* To enable VSCode debugging, added a workspace for vitest and fix any relative path to be working-dir independent.
Move selections to a new Vue component so that they don't interfere with clicking on edges.
https://github.com/enso-org/enso/assets/1047859/190006df-1402-4a5f-9117-f43788fa4187
# Important Notes
- The new `GraphNodeSelection` components are drawn in a `GraphNodeSelections` container. They are created by the `GraphNode`s and teleported so that the node can set properties directly.
Add `Vector` AST type, corresponding to the `RawAst.Tree.Array` type (name `Array` not used for obvious reasons).
This is the first step of #5138.
### Important Notes
- Switched some string-based vector construction to `Vector.new`, improving type-safety.
- The `Ast` changes are covered by the round-trip tests; the use-site changes have been tested manually.
When zooming with Ctrl+wheel, used fixed-factor steps instead of scroll wheel rate information.
Fixes#9177.
Important Notes
- Wheel events are distinguished from trackpad gestures, so that OS X pinch-zoom still works nicely.
- When zooming with the mouse wheel, scale factor is rounded (geometrically) to the nearest power of √2, and then stepped up or down by the same factor; this ensures that round values like 100% are never skipped over.
- Added directed-clamping logic, so that if the zoom is moved outside the clamping range of a zoom-method, it can be stepped back into range without jumping.
Fixes#9331
Fixed issues with wrong initial size and missing edit caret in text widgets.
<img width="311" alt="image" src="https://github.com/enso-org/enso/assets/919491/44f257cc-18a1-4a9f-9ae0-c1dd9b86674e">
# Important Notes
Automated tests for font loading/initial size will follow shortly. The text caret is not really testable, since it is a hosted object visual issue.
Addressing review suggestions from #9130
- Removing `N` binding
- Removing duplicated `Toggle fullscreen vis` binding
A few important differences from the suggested implementation:
1. There is no easy way to implement `nextType` on GraphEditor – we simply don’t have the required API
2. The keydown handler in `GraphVisualization` must be defined on window level still, otherwise it won’t get keydown events unless visualization is focused, and thus `nextType` won’t work because of (1)
No visual changes to the IDE.
- Close#9164
- Fix appearance of Record/Record Once icon in top menu
- Change icon for overriding execution context to record icon
- Unconditionally show per-node record icon if it is set
- Remove the ability to override the execution context to disabled
- Fix the icon for nodes with an overridden execution context always being the Enso icon
# Important Notes
None
- Close https://github.com/enso-org/cloud-v2/issues/866
- Remove *all* references to client keys and API base URLs from the codebase.
- The app can still be built by external contributors. *However*, the cloud backend (among some other things) will be completely disabled, as the required keys and base URLs will be missing.
- Add entry to `.gitignore` to allow `*.env` files in `app/ide-desktop/lib/dashboard/`
# Important Notes
- Tested (no `.env`; `.env` with prod backend; `.pbuchu.env`) on:
- `npm run dev` in `app/ide-desktop/lib/dashboard/`
- `./run ide build`
- `./run ide2 build`
- `./run gui watch`
Removes the old GUI1 code base and reduces the Rust code footprint by removing unused code.
# Important Notes
Updates build scripts and reformats part of the codebase with the autoformatter.
When the CB is opened, pan to show it.
Large screen:
<video src="https://github.com/enso-org/enso/assets/1047859/1a07c8cc-5818-420a-9fb3-1d1cb308cb87">
Small screen:
<video src="https://github.com/enso-org/enso/assets/1047859/a9f18df5-c0ca-426c-959a-bda5cd077541">
# Important Notes
A prioritized-coordinates approach is used to adjust panning goals based on screen space:
- Fitting the input area is highest-priority.
- If possible, the whole component panel area will be fit.
- If possible, the visualization preview will be fit.
- If there's extra room, margins will be included; the top and left are prioritized because those margins prevent overlap with fixed UI elements.
- Improved performance by batching simulatenous node edits, including metadata updates when dragging many selected nodes together.
- Updated Vue to new version, allowing us to use `defineModel`.
- Fixed#9161
- Unified all handling of auto-blur by making `useAutoBlur` cheap to register - all logic goes through a single window event handler.
- Combined all `ResizeObserver`s into one.
- Fixed the behaviour of repeated toast messages. Now only the latest compilation status is visible at any given time, and the errors disappear once compilation passes.
- Actually fixed broken interaction of node and visualization widths. There no longer is a style feedback loop and the visible node backdrop width no longer jumps or randomly fails to update.
After investigating some errors, I found another two missing awaits in our tests. Because those are so easy to overlook, I added a lint rule which makes failure on unhandled promise (for e2e tests only).
Also, enabled HTML reports again, with traces this time, to enable closer investigation of any failure in the future. @mwu-tow added code for uploading them in GH.
Fixes#7562. Close a dropdown when:
- A click outside the dropdown occurs
- `Esc` is pressed
- Any other `Interaction` is started (i.e. using a shortcut)
# Important Notes
- Simplifies `Interaction` API and uses it for closing/canceling CB as well as dropdowns.
- Adjusted some event handlers so that handled clicks don't also register as GraphEditor background clicks.
- Introduces a CSS approach to prevent unwanted text-selections; we were doing it with JS until my previous PR, and the JS solution was breaking things.
Sometimes our test realizes, that `currentTarget` of `MouseEvent` is a read-only property. I have no idea why it does not do it every time, but this fix should work.
Introduce basic display of node documentation-comments.
Part of #9162.
# Important Notes
- Add synchronization of `documentation` updates to the GraphDB; add a type assertion that ensures `Node` synchronization is updated when new fields are added to `Node`.
- Introduce read-only comment rendering.
Part of #9162.
- Add support for representing and interpreting the full text-literal/documentation syntax (escape codes, platform-independent newlines, string interpolations); build on generalized operations for structured-fields that will simplify future representation of other types like `Vector`.
- Load parsed and interpreted node documentation into `Node`s.
- Fix#9107
- Allow matching initials of aliases
- Allow matching spaces instead of underscores (aliases use spaces, not underscores)
- Fix only the first alias being detected. This is due to incorrect docs parsing keeping the leading space - `foo, bar` turns into `["foo", " bar"]`
# Important Notes
None
Fixes#9182
Now we are careful when to handle events - before or after updating intersected set.
Added general E2E and detailed unit tests for selecting.
Also discovered, that sometimes the test are running while nodes are not yet in right positions. Added an instruction ensuring the position is right.
- Fixes support for vibrancy
- Disable background when vibrancy is enabled
- Minor refactors to pass the relevant config option to dashboard, so that the dashboard knows whether vibrancy is enabled
- Enable vibrancy on Linux if a special environment variable is set
# Important Notes
- MUST be tested on macOS
- Windows *does not support resizing* when vibrancy is enabled. This is a limitation of Windows - see [this issue](https://github.com/electron/electron/issues/6107).
Fixes#8570.
- Visualization type can be set per suggestion entry.
- Default visualization type for each suggestion is determined by suggestion type info.
# Important Notes
Previewing non-default visualizations seems to be broken (#9194), but this PR updates the GUI to select visualization types appropriately.
- Fix#9184
- It turns out the logic for handling macOS was already present elsewhere - and even *in the same file*! It seems like this was an oversight
# Important Notes
None
The replacement for https://github.com/enso-org/enso/pull/9144
We try to match bindings using "buttons" field, but it in fact does not contain the just-released buttons.
# Important Notes
I added tests for editing node. Also, I decided that the cursor position should be at the end of line when starting editing by clicking at "edit" icon.
Fixes#9151
In one of my previous PRs I merged two `escape` implementations having exactly the same documentation. However, one of it missed escaping `\` character.
# Important Notes
One test even expected that `\` are _not_ escaped, but I doubt it should, given the above bug report?
Fixes one failure in #8942 which caught a real issue: clicks at various panels were triggering many handlers at once, often unexpectedly - for example quick clicking at breadcrumbs (and automatic tests always click fast) we could also trigger getting out the current node.
Therefore, I added `stopPropagation` for all mouse events on "panel" level + where I think the event should be considered "handled" and no longer bother anyone. Also, to unify things, most actions are for `click` event.
Additionally I spotted and fixed some issues:
* When "clicking-off" Component Browser, it creates new node only if anything was actually typed in (no more dangling `operatorX.` nodes)
* Filtering now works for operators
* When, after opening CB with source node, user starts typing operator, we replace dot with space
* Fixed our shortcut handler, so it works properly with `click` event.
* Fixed problems with defocusing input in CB when clicking at links.
# Important Notes
I removed `PointerMain` binding for deselectAll, because it was triggered every time did the area selection.
- Fix a bug where longitude was used as both latitude and longitude, due to a typo....
# Important Notes
- As our Mapbox token is no longer valid, the map visualization no longer works, even with this fix. As a result, this cannot be tested properly.
- However, this can still be tested somewhat, by seeing that the error is no longer present in the repro:
- Open the `Colorado COVID` template
- Open the viz on this node (at the bottom right of graph. it may help to zoom out using Ctrl+Shift+A.):
- Compare the errors logged in the console between `develop`, and this branch. This branch should not have the error about `lat` (latitude) being out of range.
* Prevent duplication of external IDs in syncTree
* trim component browser accepted value
* use flush post for visualization data watcher
* do not create empty nodes when closing component browser
* Lint
---------
Co-authored-by: Kaz <kaz@lambdaverse.org>
- Fix the UI problems with our CodeMirror integration (Fixed view stability; Fixed a focus bug; Fixed errors caused by diagnostics range exceptions; Fixed linter invalidation--see https://discuss.codemirror.net/t/problem-trying-to-force-linting/5823; Implemented edit-coalescing for performance).
- Introduce an algorithm for applying text edits to an AST. Compared to the GUI1 approach, the new algorithm supports deeper identity-stability for expressions (which is important for subexpression metadata and Y.Js sync), as well as reordered-subtree identification.
- Enable the code editor.
Fixes#9049
There were still some stubs left in node creation code.
# Important Notes
I haven't added any tests, as unit testing the graph store is very difficult - it has much tangled logic there. I'm going to try to untangle it a bit, but in a separate PR.
Fixes#9008
Now, even if we inherit FunctionCall config from parent widget (e.g. drop down), we still ask for config of the current call and try to merge them (preferring the inherited parameters).
To guard us from flaky tests, CI will run every test three times and fail if _any_ of the run fails.
This way we hope most flakiness will be catch before merging PR.
Configured dashboard in the same way.
Closes#8932
Now we use a bit more robust mechanism for passing dynamic configuration down the widget tree inside dropdowns, no longer relying on the `label`s used for dropdown items.
Curiously, we still need to use a hotfix implemented earlier, as we won’t have info about the currently selected item otherwise. Highlight for the currently selected item is not crucial as proper dynamic config, so we can leave with the current solution in the meantime.
No visual changes to the IDE, apart from fixed highlight for currently selected item.
# Important Notes
Target branch: #8950, for easier testing.
When using `npm run lint`, there are a lot of warnings for the last couple of weeks, and it is really annoying because I keep skipping important warnings for my files.
Introduce `SourceDocument`, a reactive source code representation that can be synced from a `MutableModule`. `SourceDocument` replaces various logic for tracking source code and spans--most importantly, `ReactiveModule`. There is no longer any reactively-tracked `Module`, per-se: Changes to the `MutableModule` attached to the synchronized `ydoc` are pushed as Y.Js events to the `SourceDocument` and `GraphDb`, which are reactively tracked. This avoids a problem in the upcoming text-synchronization (next PR) that was caused by a reactive back channel bypassing the `GraphDb` and resulting in observation of inconsistent states.
Stacked on #8956. Part of #8238.
Changes in preparation for #8238 features.
# Important Notes
Changed edit APIs:
- **`graph.astModule` is deprecated.** It will be removed in my next PR.
- Prefer `graph.edit` to start and commit an edit.
- Use `graph.startEdit` / `graph.commitEdit` if the edit can't be confined to one scope.
Probably fixes the first point in #8942
My guess is that the mockExpressionUpdate may be not set in cases when the test runs before App mounting. To make sure all the setup is done, we wait for `App` widget to being mounted before proceeding.
# Important Notes
This PR also change the way we run test's server. Before it was always a production build, but this makes development iteration long. Now we test production build only on CI or when `PROD=true` env variable is set.
Fixes#8871
The issue was caused by invalid port registration. Because of the existing context switch expression (which is not visible in GUI), the port incorrectly considered itself to belong to another node. This happened because the port was only aware of the visible part of the node’s AST and considered it the whole node.
There are two fixes in this PR. Either of them fixes the issue, and they are both implemented for robustness:
1. We provide `nodeId` information to the widget tree, so it no longer assumes the node ID from AST.
2. The order of checks in `getPortNodeId` is swapped. Now we first search by AST, only then try to look up the port. It makes sense to me because the AST is a single root of truth, and we should only rely on registered ports if AST does not exist (which happens for unconnected ports).
- Synchronize Y.Js clients by AST (implements #8237).
- Before committing an edit, insert any parentheses-nodes needed for the concrete syntax to reflect tree structure (fixes#8884).
- Move `externalId` and all node metadata into a Y.Map owned by each `Ast`. This allows including metadata changes in an edit, enables Y.Js merging of changes to different metadata fields, and will enable the use of Y.Js objects in metadata. (Implements #8804.)
### Important Notes
- Metadata is now set and retrieved through accessors on the `Ast` objects.
- Since some metadata edits need to take effect in real time (e.g. node dragging), new lower-overhead APIs (`commitDirect`, `skipTreeRepair`) are provided for careful use in certain cases.
- The client is now bundled as ESM.
- The build script cleans up git-untracked generated files in an outdated location, which fixes lint errors related to `src/generated` that may occur when switching branches.
This is a prerequisite for adding a CI action for E2E tests.
- Fix E2E tests
- Remove visual regression testing (VRT) and associated fixtures (screenshots) for now
- Switch dashboard almost fully to Vite, from ESBuild, to match GUI2's build tooling.
- Add some new E2E tests:
- Creating assets
- Deleting assets
- Creating assets from the samples on the home page
- Sort assets
- Includes fixes for sorting:
- Group sorted assets by type again (regression) (see https://github.com/enso-org/cloud-v2/issues/554)
- Make sorting by title, case insensitive. This is because it is more intuitive for non-programmers if all items with uppercase names *aren't* separated from those with lowercase names - especially since the Windows FS is case-insensitive.
- Normalization of Unicode letters is *not* currently being done. It can potentially be added later.
- Double-clicking *anywhere* on a directory row now expands it. Previously it was only being expanded when double clicking
- Add recursive label adding/removal to mirror backend
- Note: The current implementation is not exactly the same as the backend's implementation.
- Fix https://github.com/enso-org/cloud-v2/issues/872
- Unset "saved project details" (for opening the last open project) if fetching it produces an error.
# Important Notes
- All tests pass. (run `npm run test:e2e` in `app/ide-desktop/lib/dashboard`)
- All `npm` commands should be run in `app/ide-desktop/lib/dashboard`. `dashboard:*` npm scripts have been removed from `app/ide-desktop` to prevent a mess.
- `npm run dev` confirmed to still work. Note that it has not been changed as it was already using Vite.
- `npm run build` now uses `vite build`. This has been tested using a local HTTP server that supports `404.html`.
- Other cases have been tested:
- `npm run test:e2e` works (all tests pass)
- `./run ide build` works
- `./run ide watch` works
- `./run ide2 build` works
- `./run gui watch` works
Closes#8823https://github.com/enso-org/enso/assets/6566674/966576ec-6507-401c-98d3-bd71c2ffc6b2
Adds a basic text widget for text literals.
### Important Notes
Several known restrictions:
- Separators would always be replaced with single quotation marks. All types of separators in Enso are supported though, and they would be correctly escaped if needed.
- Logic for widget selection probably needs refinement (works for text literals and `Text` types, but does not work for `Text | Integer`, for example)
- **(!)** There is a very annoying issue when the input field suddenly loses focus, closing the editing mode and discarding any changes. Debugging shows that it happens when we receive an engine update (and probably recreate the node component/widget tree (???)). It requires a separate investigation.
Adds a new test scenario that checks that types are correctly set and displayed after an expression update. Includes checking the string set on the node as well as showing the type label when hovering the output port of a node.
Part of #8518Closes#8822
The drop downs were missing, because WidgetFunction had a higher score than them and took precedence. The fix was to adjust scores, so drop down is before both argument name and function. Generally, any widget expecting to "overlap" existing code (and possibly further widgets) should go before WidgetFunction.
# Important Notes
* Extended mockExpressionUpdate to accept a subexpression of a node. Also, I removed it from GraphDb - I want this function to have as little impact on app code as possible.
Introduce new AST APIs. They will be used for synchronization in the next PR.
# Important Notes
- `Module.edit` now clones a YDoc containing all the module's data; the resulting `MutableModule` can be edited with mutation APIs.
- Every AST type now has a mutable subtype. These `MutableAst` types expose all editing operations that were previously in `MutableModule`, such as `replace`; they also provide field setters.
Fixes several points from #8745
* Brought back cursor pointer in full screen visualization
* Changed behavior of edge disconnected from one side: Esc will bring the connection back, while click will remove it (and start creating new node if the source was connected).
* When hovering connection, the "active" part (i.e. the part which will be the hanging connection after click) is in normal color, and the rest is dimmed.
[Screencast from 2024-01-19 09-06-48.webm](https://github.com/enso-org/enso/assets/3919101/6df28a9c-51b3-4f98-be10-a35275eac800)
Closes#8751Closes#8752
- The numeric widget allows the use of the input field after clicking with LMB
- Slider is only visible if the engine provides widget configuration with set limits (see below for testing)
- Setting value outside limits is possible
- For now, to distinguish drag from click, we compare relative mouse movement on the mouse up event. We might benefit from using a timer instead, but let’s see how good it is now.
- Changes after demo
- No more input validation. You can enter literally anything and it would be accepted.
- Updates debouncing – the code is updated on defocus or when slider dragging has finished.
https://github.com/enso-org/enso/assets/6566674/b3580083-c678-4734-881c-97f8ac56176b
Some refactoring separated from #8825 for easier review.
# Important Notes
**ID types**
The new *synchronization IDs* will replace `ExprId` for `Ast` references in frontend logic. `ExprId` (now called `ExternalId`) is now used only for module serialization and engine communication. The graph database will maintain an index that is used to translate at the boundaries. For now, this translation is implemented as a type cast, as the IDs have the same values until the next PR.
- `AstId`: Identifies an `Ast` node.
- `NodeId`: A subtype of `AstId`.
- `ExternalId`: UUID used for serialization and engine communication.
**Other changes**:
- Immediate validation of `Owned` usage.
- Eliminate `Ast.RawCode`.
- Prepare to remove `IdMap` from yjsModel.
A follow-up of [this comment](https://github.com/enso-org/enso/pull/8740#discussion_r1452282468)
`addRequiredImport` now reads the existing import from the edit instead of taking them from the current module. This way, it will consider any imports added so far as part of this edit.
Related to https://github.com/enso-org/enso/issues/8518
These tests already caught one regression in nav breadcrumbs: https://github.com/enso-org/enso/issues/8756
It also provides API for mocking expression updates for arbitrary nodes on the screen. The implementation is a bit convoluted and includes setting a callback on `window`, but it looks like the only possible solution given our architecture and playwright restrictions.
Fixes#8788
- Fixed missing argument lists on constructors, and improved handling for various cases of partially applied functions.
- Extended tests to check for correct `self` argument placeholders.
- Additionally reworked some questionable test code to maintain separation between server and client code.
<img width="1241" alt="image" src="https://github.com/enso-org/enso/assets/919491/5377f57f-18f0-4a50-a8ab-9331862ca547">
Random.Seed doesn't work in the GUI and the TEXT_ONLY tag doesn't do anything so it was incorrectly showing up in the component browser.
This MR makes Random.Seed private to hide it from the GUI and completely removes the TEXT_ONLY tag which is unused and unimplemented.
- Correctly applying code edits when connecting to argument placeholders or disconnecting existing arguments.
- Prevented the placeholders from temporarily flickering right after the edit operation.
- Rewritten the placeholder argument placing logic to match the old GUI exactly - the argument should never move right after being connected.
- Fixes#8494
- Fixed some cases of ports either not existing, or existing where they shouldn't.
Implements first two points of #8745
1. The fix for drop-down was simple, just stop click propagation
2. The fix for connections was much more complicated. It turned out, that it's about keeping track of hovered ports; when picking an option in WidgetSelection makes the drop-down disappear - but that won't emit `pointerleave` event, so the port was still deemed hovered. Changed the mechanism for tracking hovered port to more "centralized" one.
- Significantly flattens directory structure of `lib/dashboard/`
# Important Notes
- Basic testing done on:
- dashboard's `npm run dev` which (since quite recently) uses Vite.
- specifically: `npm run dev` in `app/ide-desktop/lib/dashboard`, OR `npm run dashboard:dev` in `app/ide-desktop`
- dashboard's bundle script (`npm run build`) which uses ESBuild.
- GUI2's own entry point (GUI2's `npm run dev`).
- `./run ide build`
- `./run ide watch`
- `./run ide2 build`
- `./run gui watch`
[Screencast from 2024-01-05 12-18-55.webm](https://github.com/enso-org/enso/assets/3919101/f083512f-f698-42d1-b43a-4e50546b958a)
* Before most widgets didn't show in placeholders, because the argument name widget did not create further widgets. Now it try in more cases, and ignores the problem if no widget is found.
* WidgetInput is no longer a set of classes - instead it contains a set of common fields, and every widget can just extend it with new fields. Adjusted priorities accordingly.
* Updated widget definitions: they better try to match types, and take care about default values.
# Important Notes
Setting value on placeholder breaks widgets. The issue is also present in develop (when connecting to placeholder). To restore node, you have to reopen the project.
Improve `commitEdit` API.
# Important Notes
- `commitEdit` now warns if it is called before a previous edit has been observed. The preferred approach is to ensure all related changes reside in the same `MutableModule`.
- Remove `root` parameter. I included it to avoid a null check/fallible operation, but it's better to repeat the check and simplify the API.
* Merged ArgumentAst and ArgumentPlaceholder into single class.
* Created `AnyWidget` input being a "general use" widget input. Most wigets try to match with it; the `Argument` input is now solely for WidgetArgument(Name) or those handling arguments in a specific way (like selector which want to show on arg name click).
* dynamic config is now part of widget input, and is properly propagated through vector editor/function widgets.
# Important Notes
The widgets still does not work perfectly:
* The chosen options often don't have argument placeholders - that's because we don't display them for constructors. Needs to be added on our side, or engine should send us methodCall info for constructors.
* There are issues with engine's messages sent to us. This makes widgets does not set up (so there is no drop-down, or vector adds `_` instead of default). I'm investigating them and going to fill issues.
More cases in E2E tests checking all ways of using component browser.
The tests found one actual bug: sometimes we displayed only a few entries after opening CB because the scroller size was not refreshed (we assumed it will be available in the next tick). Refactored Component Browser so it does not use nextTick anymore.
Further AST integration and API refinement before introducing the new synchronization mechanism.
closes#8609
# Important Notes
API changes:
- Renamed `Ast.parseExpression` / `Ast.parse` to better reflect their usage:
- `Ast.parse` returns a block or a single expression, depending on its input; this is convenient for expressions expected to be single-line.
- `Ast.parseBlock` always treats its input as a block; this is suitable for parsing a file.
- `astExtended` is no longer needed to access span information. `Ast.span` provides access to the value when appropriate (the project's committed modules have spans; uncommitted changes or other parsed expressions don't).
- `SourceRange` is now used everywhere in place of `ContentRange`--the two types had the same definition.
Features:
- Fix CodeEditor viewing.
Implementation improvements:
- Updated widget update handlers (#8545) implementation to Ast API.
- Integrated `imports` more thoroughly with new AST APIs.
- More tests.
Implements #8525
[Screencast from 18-12-23 14:40:29.webm](https://github.com/enso-org/enso/assets/1428930/c7db2a9d-0020-4bce-b44c-d6712df58098)
# Important Notes
Previously, we re-used the API from the dashboard. But this mechanism is brittle and is currently even broken with the new GUI. So instead, I just used the Vue equivalent toast API, which seems easier to use.
- Closes https://github.com/enso-org/cloud-v2/issues/804
- The GUI should now properly configure deep links and auth settings, among other things, because the logic has been ported over from the GUI1 runner.
- The porting was very basic; it may be a good idea to refactor it later, but that is out of scope of this PR.
- This also means that `runner/` should be safe to remove along with GUI1 in the future, as all code relevant to GUI2 now resides in `app/gui2/runner`.
# Important Notes
- The built `ide2` has been tested on Windows, but should also be tested on macOS.
- Should *also* be fully tested in `npm run dev`.
E2E were failing on develop, now they're fixed.
1. There was a problem in CB which I don't know by what was triggered (But this test is fixed/extended anyway in #8559
2. Otherwise, the test scenes didn't have any nodes due to some races in initialization.
- The module is edited by changing the AST, not the text representation.
- `IdMap`s no longer need to be maintained in parallel with the module content; they are snapshots produced as needed from the ASTs.
- Simplistic synchronization is in place until #8237: Edits are never merged; if two edits are started from the same state, one will be overwritten.
- Closes#8398
# Important Notes
- The original error caused by a failing `text/openFile` (`openTextFile`) is still present, but (seemingly?) harder to repro now
Fixes parts of #8135
- Fixed rendering of the documentation panel when no entry is selected.
- Multiple improvements of tags in the documentation panel:
- Group color for `GROUP` tag
- Custom styling for `DEPRECATED`, `UNSTABLE`, and `ADVANCED` tags.
- Sorting
- Collapsing tags that exceed one line with a button
- Filtering out some tags (`ICON` and `TEXT_ONLY` for now)
- Filtering out private items from the documentation
- Fixing group colors (was incorrectly assuming unique group names) and fallback color (was incorrectly replacing every character in the name with `-` before hashing, which was hilarious).
- Fixing ComponentBrowser demo scene.
- Minor fixes for mock database in demo scenes (removed duplicated entry, added an additional one for testing tags)
https://github.com/enso-org/enso/assets/6566674/10f4f5d0-9332-40c6-93c6-8263eb5fd85a
- Closes#8072
- Implement handlers for the corresponding buttons on the circular menu
- Add missing icons and styles
- Add functionality to match and extract ASTs
# Important Notes
None
- Partially addresses #8228
- Moves several files into `@/util/vue` and `@/util/data`
- Makes imports more consistent:
- Remove redundant trailing `.ts`
- Convert *some* imports to use `@/` rather than `./`
- Ideally it should be all imports other than the ones outside of `src/`, but I can't be sure I've found all of them
- Merge *some* duplicated imports (caused by one being an `import type`)
# Important Notes
None
Partially addresses #8495
* Delete key not working in CB and in Code Editor
* Cursor not placed at the end when adding new node with source node
* Garbage added when edited node while having another one selected.
* Closing CB when navigating.
* Too eager selecting component after filtering update.
* Premature node creation when dropping edges
* Discarding changes when clicking-off CB
Implements #6166.
# Important Notes
- More consistent handling of `default` arguments. `default` is a valid identifier, and only has special meaning when it isn't bound in scope. Since distinguishing the builtin `default` from an identifier called `default` cannot be done until alias analysis has been performed, `default` is now represented in the AST as a regular identifier.
- `TreeToIr`: Remove `insideTypeAscription`. It was only used for bug-for-bug compatibility with the old parser during the transition.
Fixes#8433
* Adds E2E test to `test` script in gui2 (without server)
* Add options to `run` script to specify what test should be run: `unit` (default), `e2e`, or `ci` (which runs both unit and e2e without watching/spawning report server).
* The CI test step now checks e2e tests.
### Important Notes
~~One of e2e tests was disabled because it caught the regression on develop: #8476 ~~
- Closes#8460
- Add new magnet axes 32px below the bottom of the previous node, and 32px above the top of the next node
- Fix a bug where magnet alignment and "zoom to all" take into account nodes that no longer exist
- Fix a bug where "zoom to all" breaks completely right after deleting a node
# Important Notes
- Snapping with bounded cross axes were discussed during refinement, but are out of scope of this PR.
- Closes#6263
# Important Notes
- Still needs to be tested on Windows and macOS
- Currently does not check whether <kbd>Ctrl</kbd> is pressed to forcibly upload the file
Add CB preview panel, using engine's `executeExpression` method to get input's evaluation result without changing actual code. Also, visualization evaluation failures are reported.
# Important Notes
* **The preview does not work yet due to engine issues**.
* The full screen preview does not work properly (it closes Component Browser). Will make a separate issue for that.
* The evaluated expression calls the visualization preprocessor in such a way it handles built-in visualizations properly, but not in general case. Will start discussion with engine how to fix that.
Introduce new AST type; use it to replace current uses of AstExtended; for now edits and synchronization are implemented on the old mechanisms (text edits / RelativeRange id map).
# Important Notes
- Edit-related code is commented out until the next PR because it is incompatible with the transitional IdMap-based synchronization.
- Closes#6206
- Zoom to all/zoom to selected in visualizations, are now both `Mod+A`
- Implement Zoom to all/zoom to selected for the graph - shortcut is `Mod+Shift+A`, as `Mod+A` is "select all nodes"
- Animate zoom+pan
# Important Notes
None
- Closes#8386
- Attempts to execute `<expr>.default_visualization` to query the engine for the *correct* fallback type
- If that is not possible, falls back to checking `inputType` - first for an exact match, then falling back to `Any` (i.e. the text/JSON visualization)
- Does not decide fallback based on the shape of the returned JSON
# Important Notes
Contains pretty significant refactors of `VisualizationMetadata` to allow it to be unset.
- Closes#8261
- Add vector editor widget
- Adding element to end
- Dragging elements
- Removing elements by dragging away
# Important Notes
- Both <kbd>Left Click</kbd> and <kbd>Ctrl</kbd> + <kbd>Left Click</kbd> are supported to begin a drag.
- Just <kbd>Left Click</kbd> alone is not sufficient as some widgets interact via left click
- Minor fixes for CodeMirror highlighting (the `!` to stop recursion was making the entire import line purple, and preventing strings inside operators from being highlighted)
- Change some icons that contain dark blue, to use `currentColor` instead
- Add execution time to hover tooltip (the LS is sending it, so might as well display it)
# Important Notes
None
- Closes#8288
- Show both `Panic`s and `DataflowError`s on nodes
- Update Code Editor diagnostics to use the same functionality to properly display `DataflowError`s
# Important Notes
- Does not include warnings
- ✔️ Done basic testing to ensure that errors *should* disappear and reappear as appropriate
- Closes#8234
- Make node expand to full width of visualization
- Have a fixed minimum width for visualizations, to be able to show at least the toolbar
- Clip viz-defined toolbars to the width visualization (don't let them overflow)
- We can consider wrapping viz-defined toolbars onto a separate line, later down the line
# Important Notes
None
- Closes#8404
# Important Notes
- All relevant documentation has already been updated in #8358.
- I've tested that the keyboard shortcuts send appropriate messages over the LS connection, and have confirmed that a non-error response is received, but I'm not sure that the profiling actually works.
- There seem to be no entries in `~/.enso/`, nor in `~/enso/`
* The "main view" (when no self type/module nor pattern is specified) contains all methods defined in top modules of every library
* Entries defined inside any "Internal" module from stdlib will be hidden by default; they will be visible only when self type is specified or the module path leading to such module internals.
* Fixed an issue with missing groups sometimes - we must wait with asking for groups for first executionComplete message.
* Modules and local variables have different default icon.
Fixed broken edge dragging and creating nodes from ports. Added basic support for multiple output ports, driven by already existing analysis of the port binding structure. Those constructs are not yet supported by the engine (hence the error in code), but the IDE has easier time already dealing with ports as individual binding expressions, not whole nodes.
<img width="865" alt="image" src="https://github.com/enso-org/enso/assets/919491/73126593-05c0-4553-ba6d-dad97d083c48">
Also improved the ability to interpret applied method arguments. Different cases of dynamic, static calls or partially applied functions are now properly supported.
<img width="961" alt="image" src="https://github.com/enso-org/enso/assets/919491/ffe02d79-841c-411d-a218-de89c2522f7b">
- Closes#8179
# Important Notes
- ⚠️ These tests are currently *not run* on any CI workflow.
- There is some unused code for mocking the PM. This has been intentionally kept, as this may be useful in the future.
Note that this may be useful for testing the dashboard, however the dashboard is currently only tested in cloud mode
- that is, without the backend switcher, and with only the remote backend available. As such, currently it uses HTTP API mocks, and no PM mock.
Fixes an exception when opening the code editor. `indexedDB.cmp` doesn't accept `undefined` values. Moved its usage to a separate function to guard against that.
Part of #8372
* Sorting was broken (sort's argument was removed)
* searching by initials now does not skip words
* icons: now we use same logic as existing nodes.
* local entities has worse score if "showLocals" option is toggled off.
# Important Notes
Please be aware that there is still an important issue with groups; sometimes they're not loaded. Reported [here](8376)
As suggested by @JaroslavTulach.
- Shows diagnostics sent by `executionContext/executionStatus`
- Shows `Panic`s and `DataflowError`s sent by `executionContext/expressionUpdates`
# Important Notes
None
Fixes#8334
There were two issues: one is bad code generated by vite. I needed to rephrase Rect.Zero definition. The second were missing widgets - that was caused by the fact that Vue components have a bit different properties than on test builds, apparently.
Fixes#8252
As we use proper alias analysis, we no longer display connections to overshadowed identifiers.
# Important Notes
I've changed the meaning of edge's "source" - now it's expression ID of the constituted identifier, not the node ID. This will allow proper handling nodes like `[x, y] = Main.get_point` once the engine will support them.
- Renames all `ReactiveIndex` and `ReactiveDb`s into the format `<key>To<value>`
- `ReactiveMapping`s (in `graphDatabase.ts`) currently have not been renamed - I'm not sure they need to be
- Removes various methods that have been made unnecessary by this change
- Simplifies some other methods
- Changes `SuggestionDb` to extend `ReactiveDb`
# Important Notes
None
- Closes#8071
# Important Notes
There is currently no way to predict the width a node, taking into account the width of widgets.
This should probably be done in another task.
Addresses several issues found during book club.
- Add placeholder icons for visualizations (see screenshot)
- Also add the corresponding export to visualizations, and relevant support in the visualization store and visualization metadata DB.
- Show icon both on the visualization selector button, and in each visualization selector entry
- Adjust gaps between visualization selector entries
- Port table viz fixes (#8102) to Vue
- Fix height of table rows to avoid cutting off descenders on letters like `y` and `g`
- Make the space and enter keys work on visualization toolbar buttons, if they are selected
- `.blur()` code editor when clicking outside of it
- And similarly `.blur()` visualization selector
- Move selection brush on scroll by scaling the delta of `scrollTop` and `scrollLeft`
- Note that mouse position is technically still incorrect when scrolling past the end. I think this is fine - the new behavior is less broken, plus I am not aware of any way to fix this.
# Important Notes
None
- Closes#8304
# Important Notes
- A better long-term solution may be to remove COOP/COEP/CORP headers instead, but that is out of scope of this PR.
- Feel free to create a new issue, or discuss elsewhere, if this is something we want to do eventually
