The addresses the following case:
local l1: Listing<String> = new { 1 }
l2: Listing<Int> = (l1) { 2 }
The member in `l1` should be checked against its owner; to check against `Listing<String>`.
This adds a new feature to build a dependency graph of Pkl programs, following the SPICE outlined in https://github.com/apple/pkl-evolution/pull/2.
It adds:
* CLI command `pkl analyze imports`
* Java API `org.pkl.core.Analyzer`
* Pkl stdlib module `pkl:analyze`
* pkl-gradle extension `analyze`
In addition, it also changes the Gradle plugin such that `transitiveModules` is by default computed from the import graph.
This adds support for adding the `const` modifier to object members.
Such object members are also required to have the `local` modifier applied.
This follows SPICE-0011.
- Update dependencies by deleting lock files and regenerating them with `gw updateDependencyLocks`.
Deleting lock files avoids strange `some.library:some.older.version=default` entries.
Most updated dependencies are test dependencies.
- Handle breaking changes in library commonmark.
- Fix test to close PackageServer exactly once.
This problem surfaced because JUnit 5.11 changed override rules for lifecycle methods,
resulting in too many instead of too few close() calls.
- Bump msgpack version
- Bump clikt version
- Bump Gradle plugin versions
This changes how the language performs typechecks for mappings and
listings.
Currently, Pkl will shallow-force any Mapping and Listing to check it
the type parameter (e.g. Listing<Person> means each element is checked
to be an instance of Person).
This changes the language to check each member's type when the member
is accessed.
This also adjust test runner to handle thrown errors from within tests.
With the change to make mapping/listing typechecks lazy, we can now
correctly handle thrown errors from within a single test case.
This adjusts the test runner to consider any thrown errors as a failure
for that specific test case.
To make error messages from Pkl eval easier to read, this change uses
the Jansi library to colour the output, making it quicker and easier to
scan error messages and understand what's happened.
The Jansi library also detects if the CLI output is a terminal capable
of handling colours, and will automatically strip out escape codes if
the output won't support them (e.g. piping the output somewhere else).
Instead of bundling Pkl's built-in CA certificates as a class path resource and loading them at runtime,
pass them to the native image compiler as the default SSL context's trust store.
This results in faster SSL initialization and is more consistent with how default certificates
are handled when running on the JVM.
Further related improvements:
- Remove HttpClientBuilder methods `addDefaultCliCertificates` and `addBuiltInCertificates`.
- Remove pkl-certs subproject and the optional dependencies on it.
- Move `PklCARoots.pem` to `pkl-cli/src/certs`.
- Fix certificate related error messages that were missing an argument.
- Prevent PklBugException if initialization of `CliBaseOptions.httpClient` fails.
- Add ability to set CA certificates as a byte array
- Add CA certificates option to message passing API
* Add `--proxy` and `--no-proxy` CLI flags
* Add property `http` to `pkl:settings`
* Move `EvaluatorSettings` from `pkl:Project` to its own module and add property `http`
* Add support for proxying in server mode, and through Gradle
* Add `setProxy()` to `HttpClient`
* Add documentation
This adds a language change that requires references from typealiases
to the enclosing module to be `const`.
This is required because typealiases are not late-bound.
Rationale is laid out in SPICE-0007.
Also:
* Update documentation to reflect new rules.
* Fix `Project.pkl`; mark method `const` to not break said rule.
This adds changes to support loading project dependencies in non-file based projects.
The design for this feature can be found in SPICE-0005: https://github.com/apple/pkl-evolution/pull/6
Changes:
* Consider all imports prefixed with `@` as dependency notation.
* Bugfix: fix resolution of glob expressions in a local dependency.
* Adjust pkl.Project:
- Allow local dependencies from a scheme-local paths.
- Disallow certain evaluator settings if not loaded as a file-based module.
* Breaking API change: `ProjectDependenciesManager` constructor now requires `ModuleResolver` and `SecurityManager`.
* Split MemberNode into (Regular/Shared)MemberNode
SharedMemberNode enables generating non-constant object members
at run time without generating an unbounded number of Truffle root nodes.
* Invert shouldRunTypeCheck to match its name
* Introduce VmObjectBuilder
Introduce VmObjectBuilder, a uniform way to build `VmObject`s
whose `ObjectMember`s are determined at run time.
Replace some manual object building code with VmObjectBuilder.
Add some assertions to fix IntelliJ warnings.
* Improve implementation of globbed read/import nodes
- Leverage SharedMemberNode to have a single Truffle root node per globbed
read/import expression instead of one root node per resolved glob element.
- Remove caching in ReadGlobNode because it only works correctly if glob pattern is a string *constant*.
- Remove caching in StaticReadNode (now ReadGlobElementNode/ImportGlobElementNode)
because it seems unnecessary and the implementation doesn't look quite right.
* Simplify code
* Fix ClassCastException when reflecting on globbed import
* Fix caching of globbed reads
Problem:
The result of a globbed read depends not only on the glob pattern but also on the current module URI.
However, ResourceManager does not take this into account when caching globbed reads, causing wrong results.
Changes:
- Cache globbed reads per read expression.
This is also how globbed imports are cached, except that import URIs are constant.
- Make ReadGlobNode and ImportGlobNode code as similar as possible.
- Reduce code duplication by inheriting from AbstractReadNode.
- Add some tests.
* Don't expose JDK internal classes; instead solve msgpack issue with `--initialize-at-run-time`.
* Use quick build mode for non-release builds: 40% faster compilation, 20% smaller executable.
* Remove options that were commented out.
* Also run ServerTest against native executable
This adds support for Windows.
The in-language path separator is still `/`, to ensure Pkl programs are cross-platform.
Log lines are written using CRLF endings on Windows.
Modules that are combined with `--module-output-separator` uses LF endings to ensure
consistent rendering across platforms.
`jpkl` does not work on Windows as a direct executable.
However, it can work with `java -jar jpkl`.
Additional details:
* Adjust git settings for Windows
* Add native executable for pkl cli
* Add jdk17 windows Gradle check in CI
* Adjust CI test reports to be staged within Gradle rather than by shell script.
* Fix: encode more characters that are not safe Windows paths
* Skip running tests involving symbolic links on Windows (these require administrator privileges to run).
* Introduce custom implementation of `IoUtils.relativize`
* Allow Gradle to initialize ExecutableJar `Property` values
* Add Gradle flag to enable remote JVM debugging
Co-authored-by: Philip K.F. Hölzenspies <holzensp@gmail.com>
This changes the file paths to use characters that are safe for Windows.
Channges the output of the following:
* Package cache directory
* Generated pkl-doc files
* Kotlin generated code
Unsafe characters are encoded as (<hex>).
For example, the colon character `:` is encoded as `(3a)`.
Additionally, this changes the cache directory prefix (package-1 to
package-2).
Follows the design of https://github.com/apple/pkl-evolution/pull/3
This is technically a breaking change, but follows the intended use-case
and documentation of settings.
* Require that the global settings file at `~/.pkl/settings.pkl`
amends stdlib module `pkl:settings`, or otherwise sets `output.value`
to an instance of `pkl.Settings`.
* Simplify the object mapping of `PklSettings.java`
When using spread syntax with typed object, properties that are fixed
or const cannot be assigned or amended.
This should be checked in the checkTypedProperty() function.
- Update google-java-format to a version compatible with JDK 21 and run "gw spotlessApply".
- Fix wrong test assumption
JavaCodeGenerator writes a properties file using java.util.Properties,
which doesn't guarantee order of entries.
- Fix most deprecation warnings
- Add CI job for JDK 21
- Change HttpClient to follow all redirects except HTTPS to HTTP.
- Run language snippet tests with --no-cache and real PackageServer
instead of pre-seeded cache.
This increases HTTP test coverage and enables testing of package redirects.
- Change PackageServer to return 301 for request paths starting with /HTTP301/
and 307 for request paths starting with /HTTP307/.
- Update some outdated test package checksums that apparently weren't verified.
- Fix and clean up the pkl-commons-test build script.
- Change tests to read test packages/certs directly from
the file system instead of packaging and reading them
from the class path.
- Update expected checksums of some test packages.
- Fix a conflict between Pkl's and Gradle's
Kotlin libraries in the pkl-gradle project.
- Fix build deprecation warnings.
- Ensure Gradle distribution integrity with `distributionSha256Sum`.
- Manually verify integrity of Gradle wrapper added by this commit.
This fixes an assertion error that gets thrown if:
1. A module declares a class as abstract
2. An amending module does not use that abstract class as a type
Underneath the hood, the modifiers of the class/typelias object member
is considered different from the modifiers on the VmClass/VmTypeAlias
values.
The object model skips forcing any explicitly members that are explicitly
declared hidden, abstract, or local.
However, it _should_ evaluate any abstract classes found in a module.
So, it's incorrect to apply the same modifiers on the class to the object member.
This is a comprehensive solution to the "flaky PackageServer tests"
problem. It rules out port conflicts and imposes no limits on test
parallelism. The same solution can be used for other test servers
in the future.
Major changes:
- Turn `PackageServer` from a singleton into a class that is
instantiated per test class or test method.
- Start the server the first time its `port` property is read.
Bind the server to an ephemeral port instead of port 12110.
- For every test that uses `PackageServer`, pass the server port to
`--test-port`, `HttpClient.Builder.setTestPort`, the `CliBaseOptions`
or `ExecutorOptions` constructor, or the Gradle plugin's `testPort` property.
Wire all of these to `RequestRewritingClient`'s `testPort` constructor parameter.
- Enhance `RequestRewritingClient` to replace port 12110 with `testPort`
in request URIs unless `testPort` is -1 (its default).
- Introduce `ExecutorOptions.Builder`.
This makes executor options more comfortable to create
and allows to hide options such as `testPort`.
- Deprecate the `ExecutorOptions` constructor to steer users towards the builder.
- Get rid of `ExecutorOptions2`, which is no longer needed.
- Clean up `EmbeddedExecutorTest` with the help of the builder.
* pkl-excutor tests: symlink 0.25.0 distribution into pkl-executable/build
* Use `IoUtils.getPklHomeDir` in HttpClientBuilder
* Simplify Exceptions.java
* Enable testing for older pkl-executor distribution
Moving to java.net.http.HttpClient brings many benefits, including
HTTP/2 support and the ability to make asynchronous requests.
Major additions and changes:
- Introduce a lightweight org.pkl.core.http.HttpClient API.
This keeps some flexibility and allows to enforce behavior
such as setting the User-Agent header.
- Provide an implementation that delegates to java.net.http.HttpClient.
- Use HttpClient for all HTTP(s) requests across the codebase.
This required adding an HttpClient parameter to constructors and
factory methods of multiple classes, some of which are public APIs.
- Manage CA certificates per HTTP client instead of per JVM.
This makes it unnecessary to set JVM-wide system/security properties
and default SSLSocketFactory's.
- Add executor v2 options to the executor SPI
- Add pkl-certs as a new artifact, and remove certs from pkl-commons-cli artifact
Each HTTP client maintains its own connection pool and SSLContext.
For efficiency reasons, It's best to reuse clients whenever feasible.
To avoid memory leaks, clients are not stored in static fields.
HTTP clients are expensive to create. For this reason,
EvaluatorBuilder defaults to a "lazy" client that creates the underlying
java.net.http.HttpClient on the first send (which may never happen).
