- Added the
enso install dependencies
command to the launcher which installs any project dependencies, ensuring thatenso run
will not need to download any libraries (#1981). Additionally, made thelibrary/preinstall
endpoint able to install any transitive dependencies of the library. - Updated the Scala compiler and dependencies (#3214).
- Updated to GraalVM 21.3.0 (#3258).
- Extended language server API to allow accessing the package definition, and to get the available component groups #3286.
- Added support for overloaded conversions. This allows the
from
method to be implemented with several overloads. (#3227)
- Added experimental support for automatic parallelization of computations
(#2000). The functionality is
hidden behind the
--with-auto-parellelism
flag to the interpreter. - Added support for compiling Enso packages without execution (#1998). This allows the distribution of precompiled libraries which greatly improves language start-up time.
- Added support for computing least-squares linear regression on tables (#2003).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added support for reading the compiler's intermediate representation from disk, as well as round-tripping that IR (#1996). The compiler will now take advantage of cached IR where available to drastically reduce start-up time, and will write these caches if they do not exist or are invalid.
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added support for serializing the compiler's intermediate representation to disk (#1991). This is currently disabled by default, but in the future will allow much faster startup of the runtime.
- Added support for fetching tables from Google Spreadsheets. (#1976).
- Added support for certain statistical functions of table columns (#1990).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added
Column.set_index
to allow you to explicitly change the index for a column (#1982). Also expandedTable.set_index
to accept columns, rather than just names of columns in the table. - Added support for Amazon Redshift in the database library (#1985).
- Implement
library/preinstall
endpoint, allowing the IDE to request a library to be installed asynchronously before importing it, so that adding the import does not seem to freeze the compiler (#1972).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added support for regular expressions
(#1968). They are available as
part of the
Text.match
,Text.matches
,Text.find
,Text.split
andText.replace
methods, and are provided by a lower-level API with more control that can be found in theStandard.Base.Data.Text.Regex
module.
- Implement Language Server endpoints for getting and setting library metadata (#1967).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Implement
File.list
which allows to list files inside of a directory and filter them using glob patterns (#1961). - Implement a host of missing helpers in the Table library (#1963).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Fixed a bug where reading binary and text files would be 100 times slower than expected (#1949).
- Added the ability to specify cell ranges for reading XLS and XSLX spreadsheets (#1954).
- Updated the Simple Library Server to make it more robust; updated the edition configuration with a proper URL to the Enso Library Repository, making it possible for new libraries to be downloaded from it (#1952).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Fixed a bug where visualizations would sometimes randomly fail to compute, due to thread interrupts (#1939).
- Fixed an issue where both host and polyglot interop would not properly propagate error information (#1941).
- Fixed inaproppriate parsing of code blocks in documentation.
- Documentation in IDE now shows names of suggestions (#1904).
- Implemented a basic library uploader
(#1933). It implements the
library/publish
endpoint of the Language Server and adds apublish-library
subcommand to the Launcher. - Implemented an HTTP endpoint that resets the time that the language server has spent idle (#1938).
- Implemented the mechanism for updating the editions cache (#1944), allowing the engine to download and use editions other than the one bundled with a given release.
- Added a notification about the successful program execution to the language server API (#1945).
- Added support for writing tables to file as JSON (#1937).
- Added support for parsing CSV files with too many headers declared (#1942).
- Added a single entry point method for all table serialization modes (#1946).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Update the included project templates for better display and usability in the IDE.
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Made the project manager more resilient to edition resolution failures (#1929).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Updated the included project templates to better work with the IDE.
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Fixed an issue where the websocket buffer in the language server could overflow (#1923).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Fixed a bug with module documentation where it would associate the wrong doc-block with the module (#1919).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added support for writing tables to XLSX spreadsheets (#1906).
- Added documentation for the new searcher categories (#1910).
- Fixed a bug where CSV files with very long lines could not be parsed (#1914).
- Added ordering comparisons for
Time
,Time_Of_Day
,Date
andDuration
(#1916).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added support for documenting modules directly (#1900).
- Added support for creating projects from a template (#1902).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added support for the
ALIAS
tag in documentation blocks for use by the searcher (#1896).
- Implemented a basic library downloader (#1885), allowing the downloading of missing libraries.
- Added support for reading XLS and XLSX spreadsheets (#1879).
- Added support for serializing tables into CSV files. (#1894).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Implement parts of the new Language Server API related to library support (#1875). Parts of the API are still mocked internally, but they are supported externally for testing purposes.
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Ensure that the module used by a visualization is preloaded when the visualization is being attached (#1857).
- Fix an issue with the
HostClassLoader
getting into a broken state (#1867).
- Implemented an HTTP endpoint returning the time that the language server has spent idle (#1847).
- Fix a bug where the
project/list
endpoint would fail if any of the projects referenced an edition that does not exist anymore (#1858).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Implemented changes to the import and export syntax, requiring to provide the
project namespace, or use the new
project
keyword to import from the current project (#1806). - Fixed a bug where unresolved imports would crash the compiler (#1822).
- Implemented the ability to dynamically load local libraries (#1826). Currently, it only supports the loading of local libraries, but will be integrated with the editions system soon.
- Integrated the library loading mechanism with the editions system (#1832).
- Added namespace information to project manager messages (#1820).
- Fixed a bug where the Project Manager would not preinstall the Graal runtime if the engine was already installed and only its runtime was missing (#1824).
- Extended content root mechanism to provide the home directory and filesystem roots on startup (#1821). It now also supports dynamically adding content roots and notifies the IDE when a new content root is added.
- Connected the documentation generator with Enso compiler and suggestion database, making the documentation generated before being sent to the IDE, using a faster Scala-based generator instead of a ScalaJS-based one on IDE's side, also enabling us to connect many AST elements with docs. See (#1744.
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Upgraded the underlying runtime to GraalVM 21.1.0 (#1738). This brings a raft of bug-fixes and improvements to how quickly Enso can reach its peak performance.
- Added support for bidirectional dataflow tracking to the
DataflowAnalysis
pass (#1748). This will allow the interpreter to perform more detailed analyses in the future to enable optimisations and new features.
- Added support for higher-kinded types in suggestions in the language server (#1712). This allows the searcher to make more accurate suggestions when working with collection types.
- Fixed an issue where symlinks were not extracted properly when installing a runtime for Enso (#1718).
- Implemented log masking (#1732). This feature masks personally identifiable information in the logs, such as code literals, computed values, and user environment variables.
- Added support for evaluating one-shot expressions on the result values of arbitrary expressions (#1749). This is very useful for enabling more advanced introspection in the IDE.
- Added the
workspace/projectInfo
endpoint to the language server (#1759). This allows the IDE to get information about the running project in contexts where the project manager isn't available or works differently. - Added the
file/checksum
endpoint to the language server (#1787). This allows the IDE to verify the integrity of files that it has transferred. The checksum is calculated in a streaming fashion so the checksummed file need not be resident in memory all at once. - Added support for reading and writing byte ranges in files remotely (#1795). This allows the IDE to transfer files to a remote back-end in a streaming fashion.
- Added support for multiple content roots in the language server (#1800). It is not yet exposed to the IDE, as this will be done as part of future work.
- Modified the
package.yaml
format in preparation for the library ecosystem (#1797). Theengine-version
field has been deprecated in favour of anedition
field that allows to set up the engine version and dependency resolution using the upcoming Edition system. New tools will still be able to read the old format, but upon modification, they will save changes in the new format. As theedition
file did not exist in the older version, old tools will actually correctly load the migrated package file (as we allow for unknown fields), but they will not know how to interpret the newedition
field and so will fall back to using thedefault
engine version, which may be unexpected. Ideally, after migration, the project should be used only with the new tools. The affected tools are the Launcher and the Project Manager. - Added documentation and a minimal tool for hosting custom library repositories (#1804).
- Added
documentationHtml
field to Suggestions database entry (#1791)
- Overhauled the examples throughout the standard library (#1707, #1725, and #1731). These examples all now conform to a standard format and have been tested to work.
- Made some miscellaneous fixes to the
HTTP
portion of theBase
library that fix a few bugs (#1722). - Removed reflective access when loading the OpenCV library (#1727). Illegal reflective access operations were deprecated and will be denied in future JVM releases.
- Overhauled the types we use for errors throughout the standard library (#1734). They are now much more informative, and should provide more clarity when things go wrong.
- Re-wrote the documentation generator for the Enso website from Python into Scala (#1729). This has greatly improved the performance, enabling us to generate the documentation structure for the entire standard library 8-10 times faster than before.
- Implemented Standard Library methods for controlling default visualizations in the graphical interface (#1786).
- Adding a pipeline for automatic nightly builds
(#1689). During the night after
each workday any new changes to the
main
branch are built and released as a nightly build. The nightly builds can be useful to preview in-development features, but they should not be relied on as they are not considered stable. Only the 3 latest nightly builds are kept, so the nightly versions become obsolete very quickly.
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added a feature that allows the tooling to install GraalVM language implementations to existing runtimes (#1660).
- Fixed an issue that would prevent the language server from starting if projects had clashing identifiers (#1665).
- Added support to the language server for suggesting the module types themselves.
- Added support for reporting errors in visualisation code, making it much simpler to write new visualisation preprocessors (#1671). Previously the preprocessor would fail without any information as to what went wrong.
- Fixed an issue where the language server's update state could become desynchronised with the IDE's one (#1691). This meant that the IDE and language server didn't agree on what had been sent, and hence the IDE would miss out on certain updates.
- Added a schema version to the suggestions database, allowing the tooling to detect out-of-date versions and upgrade them (#1703).
- Added detailed logging to the tooling boot sequence to help us debug issues that users are seeing (#1704).
- Fixed some inconsistent naming around the
Maybe
type (#1666). - Added the
.sum
method for vectors of numeric types (#1702).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added support for the Python and R runtimes to the bundled runtime (#1644).
- Added a feature to ensure that suggestions are ranked by type specificity, with the more specific suggestions being ranked first (#1629).
- Fixed a raft of small issues in the runtime server that caused bugs in the engine's interaction with the IDE (#1633).
- Fixed an issue where the suggestions database would get out of sync when a project was renamed (#1647).
- Fixed some bugs in the vector constructors that prevented them from working correctly on certain inputs (#1650).
- Added support to the launcher and project manager for installing companion runtimes alongside Enso (#1651).
- Added some additional useful methods to the
Standard.Table
library (#1628). - Added a method to perform basic type inference on JSON, allowing converting
Geo-JSON to a
Table
(#1632). - Performed a comprehensive overhaul of the standard library documentation (#1641). It now has a standard format.
- Fixed an issue where we were accidentally archiving two copies of some runtime components (#1631). Downloads should now be smaller.
- Fixed an issue where a panic would be improperly cached, resulting in no updates being sent to the IDE (#1611).
- Added a feature to provide searcher suggestions for types compatible with the
type of
this
(#1613).
- Added a prototype of a library for working with images (#1450).
- Added histogram and scatter-plot visualisation support for the
Table
library (#1608). - Fixed a bug in the implementation of
join
in the database library where it would join on the wrong table when doing a multiple-join (#1614). - Fixed an outdated example for the
File.read
function.
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Fixed miscellaneous crashes in the interpreter (#1588).
- Fixed an issue where the documentation for builtins wasn't getting indexed (#1575). The docs should now show up in the searcher!
- Added support for visualising database tables to the
Database
library (#1582). - Reworked the
Process
library to work better in the IDE (#1591). - Added a proper visualisation for
Array
and improved the one forVector
(#1588).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Added rudimentary support for interoperability with Python (#1541). Due to limitations of the underlying implementation (GraalPython), this does not currently work on windows. We are working to have some means of supporting Python interop on Windows.
- Added rudimentary support for interoperability with R (#1559). Due to limitations of the underlying implementation (FastR), this does not currently work on windows. We are working to have some means of supporting R interop on Windows.
- Fixed a performance issue that occurred due to the interpreter observing deeper scopes than necessary during server-controlled execution (#1564). Execution of lambdas in the IDE is no longer far slower than it should be.
- Fixed an issue where interrupts during the execution of polyglot Java code would cause the host classloader to break, preventing further execution (#1574). Please note that the fix that has been put in place is suboptimal, and means that we are currently unable to interrupt host code during its execution. We intend to fix this as soon as a fix for the host classloader has been merged upstream. You can track the associated issue in GraalVM here.
- Fixed an issue where the interpreter would crash due to project name shadowing (#1571).
- Added support for lazy initialization of the language server (#1535). This ensures that it behaves properly on systems where the working directories are on lazily-mounted NFS volumes.
- Fixed an issue where the unified logging infrastructure would disconnect, preventing it from gathering diagnostic logs (#1563). It now sends periodic keepalive messages to ensure that the connection has not timed out.
- Fixed project name validation in the project manager when renaming projects (#1570).
- Added support for materializing data from databases in the database library (#1546). You can now use this library to connect to your data sources (currently only SQLite and Postgres, but support for further backends is planned).
- Reorganized the standard library in order to support plans for its future evolution (#1571).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Fixed another issue where the parser would crash on partial issues, causing issues for both the Engine and IDE (#1523).
- Made panic messages short, fixing an issue where retention would cause ballooned memory usage while the full message contents were waiting to be logged (#1528).
- Fixed an issue where dynamic dependencies were analysed incorrectly, leading to missing updates for the IDE (#1532).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Fixed an issue where the parser would crash on partial inputs, causing issues for both the engine and IDE (#1509).
- Fixed a problem where
Type_Error
s would not be displayed properly when pretty printed (#1504). - Fixed an issue with
_
desugaring where it would not desugar correctly when used in function position (#1512).
- Fixed an issue where suggestions were sometimes not being provided for modules
other than
Base
(#1507). - Fixed a few issues where expression and value updates were not sent when they should be (#1516, #1522, and #1508).
- Fixed a bug where sorting boolean columns in a
Table
would produce incorrect output (#1505).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Fixed another issue where dependency analysis was operating at too fine a granularity (#1495).
- Moved all user-facing errors to in-Enso errors, allowing them to be presented properly in the IDE, and interacted with by users (#1487).
- Fixed an issue where the runtime server would not send correct expression payloads for dataflow errors (#1484).
- Added "pretty" representations to all Error types, allowing for better display in the IDE (#1498).
- Updated the Table library with a raft of additional features (#1489). This includes table concatenation, direct indexing and column aggregation, as well as a general clean-up of the API pre-stabilisation.
- Added a flexible sorting mechanism to the Table library (#1471).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Improved the robustness of the project manager and language server in the presence of rapid restarts (#1463).
- Significantly improved the efficiency of visualising large tables through zero-cost translation to Enso's vectors (#1476).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Significantly improved the way that dataflow errors and panics are handled in the interpreter. They now flow through the program in a far more seamless fashion, and panics are supported properly in the IDE without the whole program dying (#1433).
- Significantly improved the syntax and semantics for FFI with JVM languages, making it far more of a first-class citizen in Enso (#1443).
- Added support for polyglot JavaScript definitions to Enso (#1451). These allow users to write JavaScript inside Enso, and seamlessly call between Enso and JS code.
- Fixed an issue where executing a host value could result in a
NullPointerException
due to a missing null check (#1413). - Fixed an issue where dataflow analysis was incorrectly tracking usages of undefined variables. This resulted in problems for the runtime server (#1421).
- Added support for collection of profiling information about the running program to the language server protocol (#1407). Initial support is for collection of execution-time information.
- Updated the default
main
in a new Enso project to be more IDE friendly (#1419). - Added support for panic sentinels in the runtime instrument, allowing the language server to trace the expressions affected by a panic while still executing others (#1436).
- Added support for checking the Enso version for a particular project in the project manager, allowing the IDE to improve compatibility with multiple versions (#1454).
- Updated the way that we use dataflow errors in the standard libraries, making
our
Base
functionality much more amenable to working in the IDE (#1446).
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Removed Uniform Function Call Syntax, making the language much more predictable and allowing the compiler to provide improved diagnostics for common mistakes.
- Return correct qualified names for modules at runtime, ensuring that the compiler and interpreter agree.
- Improve the handling of bundled components with the Project Manager, ensuring rapid startup and easy integration with the IDE.
- Fixed the reflection configuration for the Project Manager, fixing a bug where it was unable to extract archives on Windows.
- The Language Server now uses qualified names in its messages, fixing a class of bugs where the IDE and Tooling did not agree on what a given expression was.
- Fixed mis-handling of tags in the documentation parsing infrastructure.
- Implemented a stub file for the functionality built into the interpreter. This allows us to provide comprehensive documentation about this functionality for display in the IDE and for reading by library users.
- Added aggregation functionality to the Table library, allowing users to group their data.
- This is a beta release, so please see the issue tracker for issues opened before the release date.
The initial version of the Enso language, with most language features functioning.
- Flexible and concise syntax for the construction of functional programs, including pattern matching and lambdas.
- Support for currying, named and defaulted arguments, and operator sections.
- Support for user-defined data-types with fields and dynamically-dispatched methods on them.
- Importing and working with Java code in a seamless fashion using polyglot imports.
- Functional monadic state and error handling, through the use of data errors and panics.
- Opt-in tail-call optimisation.
- The initial version of the interpreter and runtime.
- Nothing.
- The initial version of the Enso Launcher and Project Manager, supporting:
- Installation and management of Enso releases, and the GraalVM runtimes on which they depend.
- Aggregation of logs from the various Enso service components.
- Basic project management functionality.
- Initialisation and set-up of a language server for a specific project.
- The initial version of the Enso Language Server, supporting:
- Dynamic introspection and modification of the running Enso program.
- Caching of intermediate values in computations, ensuring that only necessary parts of the program are recomputed on a change.
- Intelligent suggestions based on semantic analysis of the code.
- Attaching visualisation code to values in the running Enso program.
- The initial version of
Base
, the core library, supporting:- Functionality for working with core types like
Integer
,Decimal
, andText
. - Common data structures such as
List
,Vector
, andMap
. - Support for working with
JSON
data. - Support for working with
HTTP
endpoints. - Support for interacting with files and processes on the local machine.
- Support for working with polyglot entities.
- Support for metaprogramming the Enso language.
- Functionality for working with core types like
- The initial version of the Enso
Table
library for working with tabular data. - The initial version of the Enso
Test
library, containing testing and benchmarking utilities.
- A list of stabilised APIs and/or features.
- Nothing.
- This is a beta release, so please see the issue tracker for issues opened before the release date.
- Nothing