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.
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.
# 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.
Rewrites node input component. Now the input is composed of multiple widget components arranged in a tree of views with automatic layout. That allows creating complex UI elements on top of the node itself, and further widget positions will be automatically adapted to that. The tree roughly follow the span tree, as it is built by consuming its nodes and eagerly creating widgets from them. The tree is rebuilt every time the expression changes, but that rebuild process reuses as much previously created widgets as possible, and only updates their configuration as needed. Each widget type can have its own configuration options that can be passed to it from the parent, or assigned based on configuration received from the language server.
<img width="773" alt="image" src="https://user-images.githubusercontent.com/919491/233439310-9c39ea88-19bc-43da-9baf-1bb176e2724e.png">
# Important Notes
For now, all span-tree updates are sent over to the shared Frp endpoint of the whole tree, so there is no mechanism for intercepting them by the parent widgets. One idea would be to use existing bubbling/capturing events on widget display objects for that purpose, but I think existing implementation is simpler and more convenient, and we can always easily change that if we have a use for it.
There are some issues with performance due to much more display objects being created on the graph. Expect it to be a little worse, especially at initialization time.
Fixes#6317
The `drop` method is available in the WASM object. This can be tested by typing `ensoglApp.wasm.drop()` in the dev console - all objects should be removed and all connections closed.
# Important Notes
* This PR fixed serveral leaks by this occasion
* A new tool for tracking leaks was added to prelude's `debug` module.
Implementation of https://www.pivotaltracker.com/story/show/184012743https://user-images.githubusercontent.com/919491/214082311-cf49e43c-1d1f-4654-903c-a4224cd954d8.mp4
This is also a step towards more general widget support. The widget metadata is queried using `Meta.get_annotation` method through a dedicated visualization. For now only `Single_Choice` case is handled, and always all suggestions are is returned.
# Important Notes
There are limitations as to which node segments receive a widget. Only chain method calls are supported now (`thing.method` syntax), and only outside of lambda scope. Widgets in lambdas will require support for visualisations of lambda subexpressions, which is currently missing in the engine. The IDE technically tries to place the widgets there, but the data never arrives. It should work once the engine support is added.
This PR includes a mock for `Meta.get_annotation` call that only supports `Table.at` method. Real implementation is a separate task that is already in progress.
This PR implements HTML generation from documentation IR for all suggestion database entries and replaces the old documentation panel with a newer one.
Additional adjustments to the looks of the documentation would be applied separately in a future PR. This PR focuses on the fastest possible delivery of a usable documentation panel. We want to test it in real-world use cases and gather feedback for future improvements.
Documentation demo scene with mocked data:
https://user-images.githubusercontent.com/6566674/213436313-88753ed8-346f-423e-956e-7db39f5dc266.mp4
Component browser with actual engine-provided data:
https://user-images.githubusercontent.com/6566674/213436375-d0ec074b-f7a6-4deb-a7de-3adee999cc86.mp4
# Important Notes
- Fixed language protocol data structures.
- Scrolling to the selected method is also implemented here.
- Also, the selected item is highlighted with yellow.
- Only some pieces of information we have are displayed. For example, we don't display return types for methods or types of arguments.
- A bunch of code related to previous implementation is removed, but probably not all of it.
The fix consists of two parts:
1. All the "review-apply" and "store temporary md" actions in the searcher controller are now guarded by an ignored transaction.
2. Because some of the temporary state may reach the UR frames assigned to other actions, added a bunch of code for removing all temporary expressions from the code and use it after restoring a frame. We may consider using it after project load as well.
### Important Notes
Added a useful method "log_err" to ResultOps (so every Result will have those).
Logging: Replace tracing with an efficient logging implementation, with 0-runtime cost for disabled log levels. (https://www.pivotaltracker.com/story/show/183755412)
Profiling: Support submitting `profiler` events to the User Timing Web API, so that measurements can be viewed directly in the browser. (https://www.pivotaltracker.com/story/show/184003550)
# Important Notes
Logging interface:
- The macros (`warn!`, etc.) now take standard `format_args!` arguments (the tracing implementations accepted a broader syntax).
- Compile-time log levels can now be set through the CLI, like so:
`./run ide start --log-level=trace --uncollapsed-log-level=info`
Profiling:
- The hotkey Ctrl+Alt+Shift+P submits all `profiler` events logged since the application was loaded to the Web API, so that they can then be viewed with the browser's developer tools. Note that standard tools are not able to represent async task lifetimes or metadata; this is a convenient interface to a subset of `profiler` data.
- As an alternative interface, a runtime flag enables continuous measurement submission. In the browser it can be set through a URL parameter, like http://localhost:8080/?emit_user_timing_measurements=true. Note that this mode significantly impacts performance.
Fixes https://www.pivotaltracker.com/story/show/184216698
Reduced impact of node dropdown widgets on load times by deferring creation of grid views until each widget is opened. This also improves node editing time, as the dropdowns are not recreated immediately.
This approach of lazy initialization now caused a significant lag when opening the dropdown. Two major causes of the lag spike is glyph generation (msdfgen, `new_glyph`) and shader compilation (happened every time, because each dropdown has unique layer stack). To reduce the impact of that, the shader compiler now caches the shaders based on generated shader source. Glyph creation hasn't been changed and is still slow. The startup performance is now roughly where it was before introducing widgets.
This PR is a draft PR while I learn EnsoGL. The eventual goal is to implement the projects list portion of the cloud dashboard in this PR. This PR will implement part of https://www.pivotaltracker.com/n/projects/2539513/stories/183557950
### Important Notes
This PR is still really rough and contains a lot of hacks & hard-coded values. The FRP usage is also likely to be suboptimal and need fixing.
Fix issues noted here: https://github.com/enso-org/enso/pull/3678#issuecomment-1273623924
- Time complexity of an operation during line-redrawing scaled quadratically with number of lines in a change; now linear.
- Time complexity of adding `n` selections to a group was `O(n^2)`. Now it is `O(n log n)`, even if the selections are added one by one.
Also fix a subtle bug I found in `Group::newest_mut`: It returned a mutable reference that allowed breaking the *sorted* invariant of the selection group. The new implementation moves the element to invalidated space before returning a reference (internally to `LazyInvariantVec`), so that if it is mutated it will be moved to its correct location.
### Important Notes
New APIs:
- `NonEmptyVec::extend_at` supports inserting a sequence of elements at a location, with asymptotically-better performance than a series of `insert`s. (This is a subset of the functionality of `Vec::splice`, a function which we can't safely offer for `NonEmptyVec`).
- `LazyInvariantVec` supports lazily-restoring an invariant on a vector. For an invariant such as *sorted* (or in this case, *sorted and merged*), this allows asymptotically-better performance than maintaining the invariant with each mutation.
This PR fixes the `code_to_insert` method of entry to insert valid code according to the newest language version. Also created a separate method for getting imports required by given entry.
Now, method entries do not add imports (except when they are extensions), and are insterted with place for this type `_.method`. Static methods and constructors are inserted with the type name, and proper import for type is added.
There are some additional work done:
* The ReferentName and NormalizedName were removed, as we are now case-sensitive.
* All QualifiedName structures were replaced with new one in `name` module, as there is no longer functional difference between type qualified name and module qualified name.
* The QualifiedName structure removes "Main" module segment where it is not necessary, thus simplifying our code base and avoiding potential issues.
* Added macro `mock_suggestion_database` which should make creating consistent mocks of SuggestionDatabase much simpler.
* Fixed bug where the visualization preview show no value for some time.
https://user-images.githubusercontent.com/3919101/202750275-0d378d5f-1482-4637-bdcd-c428a9eac0d4.mp4
# Important Notes
The tests in controller/searcher.rs file are not of the best quality, but those will be overhauled anyway when implementing my next task.
[ci no changelog needed]
[Task link](https://www.pivotaltracker.com/story/show/182675703)
This PR implements the actual integration of the breadcrumbs panel with the component list panel. A special breadcrumbs controller (`controller::searcher::breadcrumbs`) is tracking the currently opened module with a list of its parents. The searcher presenter uses the API of the controller to sync the displayed list of breadcrumbs with the controller state.
https://user-images.githubusercontent.com/6566674/193064122-7d3fc4d6-9148-4ded-a73e-767ac9ac83f8.mp4
# Important Notes
- There is an `All` breadcrumb displayed at all times at the beginning of the list. It will be replaced with a section name as part of [Section Title on Component Browser's Breadcrumbs Panel](https://www.pivotaltracker.com/story/show/182610561) task.
- I changed the implementation of `project::main_module_id`, `project::QualifiedName::main_module`, and `API::main_module` so that they are logically connected which each other.
- I adjusted the Breadcrumbs View to avoid "appearance" animation glitches when opening new modules. `set_entries` was replaced with the `set_entries_from` endpoint.
Show custom icons in Component Browser for entries that have a non-empty `Icon` section in their docs with the section's body containing a name of a predefined icon.
https://www.pivotaltracker.com/story/show/182584336
#### Visuals
A screenshot of a couple custom icons in the Component Browser:
<img width="346" alt="Screenshot 2022-07-27 at 15 55 33" src="https://user-images.githubusercontent.com/273837/181265249-d57f861f-8095-4933-9ef6-e62644e11da3.png">
# Important Notes
- The PR assigns icon names to four items in the standard library, but only three of them are shown in the Component Browser because of [a parsing bug in the Engine](https://www.pivotaltracker.com/story/show/182781673).
- Icon names are assigned only to four items in the standard library because only two currently predefined icons match entries in the currently defined Virtual Component Groups. Adjusting the definitions of icons and Virtual Component Groups is covered by [a different task](https://www.pivotaltracker.com/story/show/182584311).
- A bug in the documentation of the Enso protocol message `DocSection` is fixed. A `text` field in the `Tag` interface is renamed to `body` (this is the field name used in Engine).
Implement generation of Java AST types from the Rust AST type definitions, with support for deserializing in Java syntax trees created in Rust.
### New Libraries
#### `enso-reflect`
Implements a `#[derive(Reflect)]` macro to enable runtime analysis of datatypes. Macro interface includes helper attributes; **the Rust types and the `reflect` attributes applied to them fully determine the Java types** ultimately produced (by `enso-metamodel`). This is the most important API, as it is used in the subject crates (`enso-parser`, and dependencies with types used in the AST). [Module docs](https://github.com/enso-org/enso/blob/wip/kw/parser/ast-transpiler/lib/rust/reflect/macros/src/lib.rs).
#### `enso-metamodel`
Provides data models for data models in Rust/Java/Meta (a highly-abstracted language-independent model--I have referred to it before as the "generic representation", but that was an overloaded term).
The high-level interface consists of operations on data models, and between them. For example, the only operations needed by [the binary that drives datatype transpilation](https://github.com/enso-org/enso/blob/wip/kw/parser/ast-transpiler/lib/rust/parser/generate-java/src/main.rs) are: `rust::to_meta`, `java::from_meta`, `java::transform::optional_to_null`, `java::to_syntax`.
The low-level interface consists of direct usage of the datatypes; this is used by [the module that implements some serialization overrides](https://github.com/enso-org/enso/blob/wip/kw/parser/ast-transpiler/lib/rust/parser/generate-java/src/serialization.rs) (so that the Java interface to `Code` references can produce `String`s on demand based on serialized offset/length pairs). The serialization override mechanism is based on customizing, not replacing, the generated deserialization methods, so as to be as robust as possible to changes in the Rust source or in the transpilation process.
### Important Notes
- Rust/Java serialization is exhaustively tested for structural compatibility. A function [`metamodel::meta::serialization::testcases`](https://github.com/enso-org/enso/blob/wip/kw/parser/ast-transpiler/lib/rust/metamodel/src/meta/serialization.rs) uses `reflect`-derived data to generate serialized representations of ASTs to use as test cases. Its should-accept cases cover every type a tree can contain; it also produces a representative set of should-reject cases. A Rust `#[test]` confirms that these cases are accepted/rejected as expected, and generated Java tests (see Binaries below) check the generated Java deserialization code against the same test cases.
- Deserializing `Code` is untested. The mechanism is in place (in Rust, we serialize only the offset/length of the `Cow`; in Java, during deserialization we obtain a context object holding a buffer for all string data; the accessor generated in Java uses the buffer and the offset/length to return `String`s), but it will be easier to test once we have implemented actually parsing something and instantiating the `Cow`s with source code.
- `#[tagged_enum]` [now supports](https://github.com/enso-org/enso/blob/wip/kw/parser/ast-transpiler/lib/rust/shapely/macros/src/tagged_enum.rs#L36-L51) control over what is done with container-level attributes; they can be applied to the container and variants (default), only to the container, or only to variants.
- Generation of `sealed` classes is supported, but currently disabled by `TARGET_VERSION` in `metamodel::java::syntax` so that tests don't require Java 15 to run. (The same logic is run either way; there is a shallow difference in output.)
### Binaries
The `enso-parser-generate-java` crate defines several binaries:
- `enso-parser-generate-java`: Performs the transpilation; after integration, this will be invoked by the build script.
- `java-tests`: Generates the Java code that tests format deserialization; after integration this command will be invoked by the build script, and its Java output compiled and run during testing.
- `graph-rust`/`graph-meta`/`graph-java`: Produce GraphViz representations of data models in different typesystems; these are for developing and understanding model transformations.
Until integration, a **script regenerates the Java and runs the format tests: `./tools/parser_generate_java.sh`**. The generated code can be browsed in `target/generated_java`.
This PR introduces a new structure delivered by Searcher Controller: The Component List.
The Component list is hierarchical, and its structure reflects how the components are displayed in Component Browser - only "submodule" section at this point, other sections will be covered in next tasks.
This does not introduce anything visual; the structures are tested in unit tests.
Add a method in `SuggestionDatabase` allowing to find a suggestion entry by a fully qualified path, working faster than a linear search through all the entries in the `SuggestionDatabase`.
https://www.pivotaltracker.com/story/show/181852566
# Important Notes
- **Testing:** when testing the PR with the newest currently available nightly version of the Engine, you may observe the following warnings in the Chrome JS Developer Console (the numerical values may differ):
index.ts?ab16:289 WARN app/gui/src/model/suggestion_database.rs:61 An existing suggestion entry id at Standard.Base.Nothing.Nothing.is_nothing was overwritten with 768.
index.ts?ab16:289 WARN app/gui/src/model/suggestion_database.rs:61 An existing suggestion entry id at Standard.Base.Data.Numbers.Integer.up_to was overwritten with 936.
index.ts?ab16:289 WARN app/gui/src/model/suggestion_database.rs:61 An existing suggestion entry id at Standard.Base.Data.Numbers.Integer.down_to was overwritten with 937.
index.ts?ab16:289 WARN app/gui/src/model/suggestion_database.rs:61 An existing suggestion entry id at Standard.Base.Data.Text.Text.== was overwritten with 971.
This is a result of bugs in the Standard Library. Those are planned to be addressed by:
- #3480 short-term (to fix the specific bugs currently present in the Standard Library),
- https://www.pivotaltracker.com/story/show/182283983 long-term (to improve the Engine such that it disallows introducing this category of bugs in the Standard Library in the future).
(For more details, see also: https://discord.com/channels/401396655599124480/978929754138877962.)
As a result of the bugs mentioned above, the Engine is responding with some invalid replies. In case of such invalid replies, warnings are emitted in the JS Dev Console. Other than the warnings, the code is expected to work correctly with the Engine for the cases when the Engine returns correct data.
- A `HashMapTree` was used for storing the map. A quick back-of-the-envelope estimation of memory usage of a simpler alternative (a one-level `HashMap`) would put the alternative at ~1MB (~10k entities × averaged ~100 bytes per entity path), which was considered too much in a discussion with @farmaazon.
- The `HashMapTree::remove` method deletes a whole subtree of a `HashMapTree`. An alternative removal method was implemented for use in `suggestion_database::QualifiedNameToIdMap` to better match `HashMap` entry removal semantics.
- In case of path collisions, a warning is emitted.
- Paths are treated case-sensitively.
- The new method of the `SuggestionDatabase` type is currently only used in unit tests. This matches the explicit requirement in the Task's description.
- JS Developer Console logging was enabled for the `tracing` package. The `WARN` level was picked to match the default level enabled in the "old" logging infrastructure.
[ci no changelog needed]
Implement a command that launches the application, runs a series of steps (a "workflow"), writes a profile to a file, and exits.
See: [#181775808](https://www.pivotaltracker.com/story/show/181775808)
# Important Notes
- The command to capture run and profile is used like: `./run profile --workflow=new_project --save-profile=out.json`. Defining some more workflows (collapse nodes, create node and edit value) comes next; they are implemented with the same infrastructure as the integration-tests.
- The `--save-profile` option can also be used when profiling interactively; when the option is provided, capturing a profile with the hotkey will write a file instead of dumping the data to the devtools console.
- If the IDE panics, the error message is now printed to the console that invoked the process, as well as the devtools console. (If a batch workflow fails, this allows us to see why.)
- New functionality (writing profile files, quitting on command, logging to console) relies on Electron APIs. These APIs are implemented in `index.js`, bridged to the render process in `preload.js`, and wrapped for use in Rust in a `debug_api` crate.
