The change adds logic that will attempt a few retries when executing `gu` (GraalVM updater) commands. Previously, if it failed, it failed. Retries should help with the most common case - occassional network hiccups.
Closes#6880.
# Important Notes
Note that I don't use an external library for retries on purpose. Didn't want to introduce a yet another dependency for this tiny functionality.
Remove a leak detector added in #7038, that has a large (~10s) impact on startup performance. I missed the performance impact during QA, but noticed it after merge.
Fixes#6908
Fast typing leads to a race condition where the end of the node editing can happen where the state of the CB is not caught up yet. This leads to two issues:
1. The wrong selection is made, as the View of the Component Browser has not been updated yet, but the selection from there is used to determine how the node is created.
2. Processing of the keys is aborted early, while still waiting for additional key input.
This is now fixed.
https://github.com/enso-org/enso/assets/1428930/dce82bb3-1c25-4a67-988a-d74469dc8fcd
Fixes#6772
When detaching an existing edge by grabbing by a source port, the node's code is no longer immediately modified. It is only changed once the edge has been either connected or destroyed. When grabbing on the source side, the existing behavior is preserved. That way, we always have guaranteed place to keep the edge connected to.
https://github.com/enso-org/enso/assets/919491/49e560cb-0a29-4c6a-97ec-4370185b8c89
In general, the detached edges are now more stable, resilient to all kinds of expression modifications during the drag.
https://github.com/enso-org/enso/assets/919491/e62450ff-46b2-466f-ac33-f4f19e66ee1d
In case there is a situation where the currently dragged edge's port is destroyed (e.g. by Undo/Redo), instead of showing glitched port position it is simply dropped.
https://github.com/enso-org/enso/assets/919491/8fb089aa-a4a5-4a8c-92eb-23aeff9867b8
# Important Notes
The whole edge connection and view handling at the graph-editor view level has been completely rewritten. The edge endpoints are now identified using new `PortId` structure, that is not dependant on the span-tree. This prepares us for eventual removal of the span-tree in favour of manipulating AST directly. Right now those `PortId`s are still stored within the span-tree nodes, but it will be easy to eventually generate them on the fly from the AST itself. The widget tree has also already been switched to that representation where appropriate.
Additionally, I have started splitting the graph editor FRP network into smaller methods. Due to its absolutely enormous size and complexity of it, I haven't finished the split completely, and mostly edge-related part is refactored. I don't want to block this PR on this any longer though, as the merge conflicts are getting a bit unwieldy to deal with.
1. Fixed leaks of Scene. We were making several cycles of Rc references.
2. Fix removing event listeners: the remove listener option should take same options as those passed in addEventListener.
This should finally fix#6505
- Add type detection for `Mixed` columns when calling column functions.
- Excel uses column name for missing headers.
- Add aliases for parse functions on text.
- Adjust `Date`, `Time_Of_Day` and `Date_Time` parse functions to not take `Nothing` anymore and provide dropdowns.
- Removed built-in parses.
- All support Locale.
- Add support for missing day or year for parsing a Date.
- All will trim values automatically.
- Added ability to list AWS profiles.
- Added ability to list S3 buckets.
- Workaround for Table.aggregate so default item added works.
Fixes performance problems observed when creating/resolving errors (#6674):
|before|after|
|---|---|
|||
This also helps with #6637, although I haven't been able to reproduce the degree of slowness shown there so I can't confirm that this resolves that issue.
# Important Notes
- Disable visualizations until shown. [Faster startup, and all graph changes.]
- 6x faster message deserialization. [Saves 400ms when making a change with many visualizations open.]
- Fast edge recoloring. [Saves 100-150ms when disconnecting an edge in Orders.]
- Add a checked implementation of a `profiler` data structure, used instead of the fast `unsafe` version when `debug-assertions` are enabled.
Several small changes:
- Dropdowns: Populate `GridView` lazily (fixes#6865).
- Clear disconnected edges when editing node (fixes case 1 in #7018).
- Fix regression in node selection rendering (2nd bug in #6975).
- Update profiler docs. The hotkey to *prOfile without exiting* is now `Ctrl+Alt+O` (`Ctrl+Alt+P` has been requisitioned by the CB).
Node selection:
| Pre-`Rectangle` | This PR |
| --- | --- |
|  |  |
# Important Notes
- `Rectangle`: When `inset > border`, the extra space is now between the body and the border, not outside the border.
- More robust node layering logic. Now an inconsistent layer order cannot occur, even if something strange happens (like editing an expression and an edge at the same time).
- The dynamic drop down in the `drop_down` example scene doesn't show any entries before (or after) this, so I can't test the dynamic case.
close#6936
Changelog:
- add: new suggestion type Getter that is not exposed to the api
- update: do not return suggestion of type getter when doing a global search (without specifying self types)
Private suggestions and modules mentioned in the issue will be filtered out after we finish the work on the new (refined) exports algorithm.
# Important Notes
This PR adds facilities for controllers to be aware of what shortcut command is currently being processed. This allows grouping consequences of single user action into a single transaction without hard-coding it separately for all the separate paths case-by-case, which turned out to be challenging and error-prone.
Additionally, a number of minor fixes were carried over from #6877:
* workaround for #6718;
* avoiding creating spurious transactions when dealing with node positions;
* dropping any non-user user-triggered transactions that occur during the IDE project initialization.
Before, we prevented default on every keyboard event, making usage of HTML inputs impossible (both on the dashboard and inside visualizations).
# Important Notes
The list of shortcuts was based on comments in #6364 (where the `preventDefault` was introduced).
Implements #6544 (eliminates 10/42 of the constantly-displayed draw calls).
Fixes#6717. Improves startup CPU time by 5% (250ms, loading Orders on my dev box).
# Important Notes
- Edges: New implementation uses only Rectangle under most conditions.
- Node and action area: Replace some shapes with Rectangle.
- List view: Replace some shapes with Rectangle.
- Display object hierarchy: The lowest-level shape instance types no longer have their own display objects.
- Includes initial support for using `Rectangle` to display triangles.
Fixes#6763
The bug was caused by pushing stack frames in the engine first, then failing to get graph controller - in that case we didn't update graph, but kept the stack, so graph was not synchronized with stack.
This PR changes the approach: we try to open graph first, only then push frames to stack. If _any_ frame will fail, we try to pop those we pushed so far, to restore the previous state. The same fix was applied for leaving nodes.
Also, I realized, that running several "enter node" / "leave node" actions could mix push/pop operations, making a mess with our execution context state. I've added a mutex to ensure we won't do this: the contesting operation will simply fail.
# Important Notes
In case when _restoring_ state fails, I did not have any better idea than just trying until it succeed.
Now, quick typing in component browser and pressing "enter" should not cut off the last part typed. Fixes#6733https://github.com/enso-org/enso/assets/3919101/3979ed5a-ba4e-4e25-93e6-672e731b7bd8
On this occasion, also fixed "go-to-dashboard" button and "Unsupported engine version" being over the full-screen visualization. Fixes#6722
# Important Notes
I did a significant refactoring of Project View:
1. The huge `frp::extend` block was split into multiple `init` methods.
2. Remaining of the "Old searcher" were removed.
3. The "Edited" event from node's input is emitted only when in edit mode (it's consistent with other API terminology, and makes FRP for showing CB much simpler.
The code was _mostly_ moved around, but the check is advised anyway, as there were small changes here and there.
Implements #6792Fixes#6715Fixes#6052Fixes#5689
The dynamic dropdown widgets entries now can specify additional widget configuration as a list of `parameters` of the inner method call. That allows for creating smarter widgets within nested constructors, taking the outer widget's context into account.
<img width="772" alt="image" src="https://github.com/enso-org/enso/assets/919491/97c70654-9170-4cf0-ae4d-2c25c74caa96">
With the changes to the serialization logic, I have also adressed issues related to automatic label generation for both static and dynamic dropdown entries. For access chains (e.g. `Foo.Bar.Baz_Qux`), the label will now always contain only the last segment, and all underscores will be removed (e.g. `Baz Qux`). This also applies to dynamic entries where the label is not explicitly specified in method annotation.
<img width="265" alt="image" src="https://github.com/enso-org/enso/assets/919491/1abe6c77-010b-4622-b252-97cd1543cb48">
Additionaly, now the dynamic entries containing constructors will also be resolved within suggestion database, allowing us to automatically insert relevant import, shorten the actually used expression and wrap it with parentheses if required. That was required for nested widgets to show up, as we depend on properly resolved argument names to show them. The widget definitions in annotations no longer need to wrap the expressions manually. Instead, the constructors used in dropdown entries should be specified using fully qualified names, similarly to how we do it in tag values.
CC @jdunkerley - The dropdown entries containing just a constructor will no longer need added parentheses around them. Instead, the constructors should be specified using fully qualified names, similarly to how we do it in tag values.
<img width="389" alt="image" src="https://github.com/enso-org/enso/assets/919491/19944b5b-d0c7-43ac-bf17-ca1556e0b3f0">
Note that currently the import resolution is attempted even if the used constructor is is not specified using a fully qualified name. To accomplish that, the IDE is performing a more expensive search through whole suggestion database for matching type and module (e.g. in example above, we are searching for a match for `Aggregate_Column.First`). If there are multiple potential matches due to a name collision, it is undefined which one would be preferred. Effectively one will be picked at random. To avoid that, the libraries should over time transition to using fully qualified names wherever possible.
# Important Notes
I have removed the `payload` field from the span tree, and with it the generic argument on its nodes. This was already partially done on the branch with new design, on which I also had a few changes that turned out to be useful for this PR. So I pulled it in as well. It is a nice simplification that will ease our further work on removing the span-tree altogether. The biggest impact it had was on the node output port, where I had to store the port data outside of the span tree. This is the approach we would be taking when transitioning to AST anyway.
Empty edition (null value) was parsed as NaN, which was confusing. This change correctly detects the case before trying different fallback mechanisms.
Addresses invalid warning mentioned in #6806.
Only invalidate the graph editor view at most once per frame. On develop, this saves about 70ms (2%). Testing a recent backend without #6755 as a stress-test, this saves about 5s (45%). This reflects better scalability to large numbers of `SuggestionUpdate` messages.
Fixes#6630.
# Important Notes
- Also fix intermittent profiling failures occurring since the introduction of microtasks.
* Remove unused code: project management in component browser
* Encapsulate internal FRP logic of project list
* Collapse some code paths
* Open project passed on command line through presenter
A project name or ID that is passed on the command line was initialised
in the controller setup, before the presenters and views are set up.
Now, we fully initialise the IDE before opening a project so we have
control over the view while a project is being opened.
* Show a spinner in all cases of opening a project
* Let root presenter open/close projects when switching projects
* Change spinner to make progress over a fixed period
* Resolve issues when Project Manager API isn't available
* Bump wasm size limit
Fixes some of #6662
Issues addressed:
- `ide watch` and `gui watch` should now use the desktop platform
- error screen should now be shown when passing invalid options
- password (both creating password when registering, and resetting password) should now warn on invalid input
# Important Notes
Instead of checking whether `location.hostname === 'localhost'`, I've opted to use a constant defined by the build tool instead. This is to make it easier to merge the cloud IDE and desktop IDE entrypoints in the future, since it would be able to simply set `platform: Platform.cloud` in the build config.
Fixes#6787
# Important Notes
I can't get Project Manager compilation to work locally so I guess I'll be relying on CI to verify that it's working correctly?
Of course, QA should be able to catch any problems too - the websocket API hasn't been changed so it should work out of the box with the current dashboard.
Refactored the logic behind selecting appropriate widgets for span tree nodes. Now the bulk of it is moved into widget methods. When a given widget type is reporting to be not compatible with the expression, it will not be used even if the configuration was overriden using an method argument annotation. In that case, the usual logic for automatically selecting the appropriate widget will kick in.
* Run typecheck and eslint on Lint CI
* Address reviews; fix type errors in `.d.ts` files
* Remove unused parameter
* Run prettier
* Fix lint error
---------
Co-authored-by: Paweł Buchowski <pawel.buchowski@enso.org>
# Important Notes
The mouse handling changes involve an unfortunate huge hack, where we enable mouse events on the mouse shape during box selection. That way we know for sure that no other shape will be able to receive mouse enter event. Then the list editor widget is modified to only actually respond to events when its background is hovered. We will definitely want a more proper way to handle mouse event contention, but it's definitely out of scope for current bugfixing.
This change fixes the rather elusive bug where shutdown hooks could not be fired when shutdown was taking too long and termination was forced.
Under the circumstances described in detail in ticket #6515 there was a small chance that we could have a shutdown race condition. Essentially the messages received when client was disconnected and language server forced the termination could lead to language server not sending the public `ProjectClosed` message which triggers shutdown hook. Now we always do.
Also made sure that multiple `ProjectClosed` messages don't lead to firing multiple shutdown hooks, which was another possibility.
No tests as one would have to be able to introduce different delays in various message handlers to simulate the problem.
Having ability to do such chaos testing would be nice but it is beyond the scope of this ticket.
I was able to reproduce the problem 100% with my specially crafted setup so I'm fairly confident about the change.
Closes#6515.
close#6611
Changelog:
- update: run compiler passes on the `ascribedType` field of the constructor arguments
- update: suggestion builder uses the type information attached to `ascribedType`
- feat: resolve qualified names in type signatures
