Deduplicate executable jar package entries. In the issue, the duplicate was something like:
(resource_managed/.../sbt/sbt.autoplugins, "sbt/sbt.autoplugins")
(classes/.../sbt/sbt.autoplugins, "sbt/sbt.autoplugins")
This is because copyResources copied autoplugin metadata to the classes directory.
Solution: when composing the packageBinMappings, differentiate between artifacts that come from class directory and from outside of it. If there is a duplicate, prefer resource that comes from outside of it.
Co-authored-by: Anatolii Kmetiuk <anatoliykmetyuk@gmail.com>
The original fixture pinned
scalaOrganization := "io.github.scala-wasm"
scalaVersion := "3.8.3-RC1-wasm-bin-SNAPSHOT"
and ran compile. The wasm fork never published a release of
scala3-library_3 -- only a snapshot to Sonatype's central-snapshots
repo, which has ~90-day retention. #8732 merged 2026-02-21; the
snapshot was GC'd ~2026-05-22 and develop CI began failing
deterministically.
Fixes#9259.
Co-authored-by: BrianHotopp <brihoto@gmail.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* [2.x] feat: Add cacheVersion setting for global cache invalidation
**Problem**
There was no escape hatch to invalidate all task caches when needed.
**Solution**
Add `Global / cacheVersion` setting that incorporates into the cache key
hash. Changing it invalidates all caches. Defaults to reading system
property `sbt.cacheversion`, or else 0L. When 0L, the hash is identical
to the previous behavior (backward compatible).
Fixes#8992
* [2.x] refactor: Simplify BuildWideCacheConfiguration and add cacheVersion test
- Replace auxiliary constructors with default parameter values
- Add unit test verifying cacheVersion invalidates the cache
* [2.x] fix: Restore auxiliary constructors for binary compatibility
* [2.x] test: Improve cacheVersion scripted test and add release note
- Scripted test now verifies cache invalidation via a counter
that increments only when the task body actually executes
- Add release note documenting the cacheVersion setting
**Problem**
Current MixedAnalyzingCompiler analysis cache caches using the last write,
assuming all writing happens via it. However, this doesn't work with
sbt 2.x caching system where the gz file under it can switch.
**Solution**
This switches to using caffeine caching, which includes file size
and file timestamp for local analysis caching.
**Problem**
Some tests are failing.
**Solution**
* Port EvictionErrorSpec to lm-coursier
* Port EvictionWarningSpec to lm-coursier
* Port InclExclSpec to lm-coursier
* Port ZincComponentCompilerSpec to lm-coursier
* Apparently we still need lmIvy
**Problem**
When usePipelining := true in a multi-project build, sbt appends
-Ypickle-java and -Ypickle-write <path>/early.jar to scalacOptions
for fast parallel compilation. These flags were being leaked into the Scala
REPL, causing failure on console
**Solution**
Strip pipelining flags before forwarding to the REPL in Deafults.scala.
**Problem**
In sbt 2.x, forking still creates layered classloader in the worker process,
which doesn't work for some tests.
**Solution**
This provides an escape hatch to emulate the sbt 1.x semantics of
using the system classpath for testing.
**Problem**
Projects defined in subdirectory build.sbt files are
not correctly resolved to the base directory relative to the file.
**Solution**
Call resolveBase.
Bug 1: bgRun forks with inheritIO() instead of LoggedOutput
bgRunTask / bgRunMainTask resolve their fork options via (run / forkOptions), which inherits run / connectInput := true. That has two downstream consequences:
1. ForkRun.configLogged (in run/src/main/scala/sbt/Run.scala) skips installing OutputStrategy.LoggedOutput(log) whenever config.connectInput == true.
2. Fork.apply (in run/src/main/scala/sbt/Fork.scala) sees connectInput && outputStrategy == StdoutOutput and takes the interactiveFork path, which calls jpb.inheritIO().
Bug 2: Background-job log relay drops messages after the spawning task ends
Even after Bug 1 is fixed, LoggedOutput routes the forked process's stdout into the background ManagedLogger, whose appenders include a shared RelayAppender. The relay calls CommandExchange.logMessage(event), which uses isChannelOwner(c) to pick the target channel.
Closes#4399
- Add subscribeToAll to InitializeOption (protocol); default true for backward compatibility.
- CommandExchange: send broadcast notifyEvent/logMessage only to channels with subscribeToAll.
- TestServer: support subscribeToAll parameter for tests; AbstractServerTest: subscribeToAllForTest.
- ClientSubscriptionTest: assert default client receives build/logMessage when command runs.
- Scripted test server/client-subscription: run show name to exercise server client path.
Summary
- Adds support for passing JVM arguments inline to `run`, `runMain`, `bgRun`, `bgRunMain`, and `fgRun`/`fgRunMain` using `--` as a delimiter
- Syntax: `run <jvmArgs> -- <appArgs>` (e.g., `run -Xmx2G -Dapp.mode=debug -- arg1 arg2`)
- Fully backward compatible — without `--`, all arguments are treated as app args as before
- When `fork` is `false`, a warning is logged that JVM arguments will be ignored
**Problem**
When using scala-native, we can't enable semanticdb etc
because of the platform-level platform.
**Solution**
Since compilerPlugin(...) is always going to be JVM,
just add an explicit JVM platform to the given ModuleID.
* Add Test configuration to evictionWarningOptions
* Add Test configuration to default evictionWarningOptions
* Deduplicate eviction errors based on callers.
* Update eviction error test for semantic versioning.
* Group evictions by configuration and update test.
Co-authored-by: Zainab Ali <zainab.ali.london@gmail.com>
**Problem**
When managedScalaInstance is set to false, we stopped creating
ScalaInstance, but the documentation says the user can
pull in their own dependencies.
**Solution**
Attempt to construct a ScalaInstance from update report
even when managedScalaInstance is set to false.
In sbt 2.x, running batch commands through the thin client (sbtn) suppresses stack traces for all tasks because the server's shell command unconditionally sets state.interactive = true. This causes LogManager.defaultTraceLevel to return -1 (suppressed) even when the client explicitly signals non-interactive (batch) mode via Attach(interactive=false).
This fixes the shell command to check the originating NetworkChannel's interactive flag before setting state.interactive, so thin client batch commands correctly get Int.MaxValue trace level and display full stack traces.
**Problem**
When dependencyMode := Direct is set, the filtering was applied at
the managedClasspath level, which removed transitive dependencies
from all downstream classpaths including Test / dependencyClasspath.
This caused runtime test failures because transitive deps like
hamcrest-core (pulled in by junit) were missing.
**Solution**
Move the dependencyMode filtering from managedClasspath to a new
filteredDependencyClasspath task, and wire dependencyPicklePath
(the classpath used by the compiler) to use it. Runtime classpaths
like dependencyClasspath and fullClasspath remain unfiltered,
preserving all transitive dependencies for test execution.
Fixes#8989
Co-authored-by: Dream <42954461+eureka928@users.noreply.github.com>
**Problem**
Like File, Path normally captures the absolute path,
so likely not a good candidate for caching.
**Solution**
Ban it.
---------
Co-authored-by: Anatolii Kmetiuk <anatoliikmt@proton.me>
**Problem**
There seems to be multiple problems around metabuild reloading (reload plugins).
1. Originally reported issue 9006 states that the build can't come back
from reload plugins.
2. During the course of fixing, I discovered that when reload plugins is used
the project name is set to "project" instead of foo-build etc.
3. Also there was a bug in the rootPaths when reload plugins is used,
which results in class not defined error.
**Solution**
1. Fix the plugin context so reload plugin still behaves like a metabuild.
2. Fix the rootPaths.
**Problem**
sbt always includes all transitive dependencies on the classpath.
This makes it easy to accidentally depend on transitive dependencies
without declaring them, leading to fragile builds that break when
a library changes its own dependencies.
**Solution**
Add a `dependencyMode` setting with three modes:
- DependencyMode.Transitive (default) — current behavior, all
transitive dependencies on the classpath
- DependencyMode.Direct — only declared dependencies plus
scala-library on the classpath
- DependencyMode.PlusOne — declared dependencies plus their
immediate transitive dependencies plus scala-library
Fixessbt/sbt#8942
Co-authored-by: Dream <42954461+eureka928@users.noreply.github.com>
Fixes unresolved dependency path reporting for Coursier (`ResolveException.failedPaths`) and adds a stable scripted regression.
This PR addresses 5168 by reconstructing unresolved dependency caller chains from the Coursier resolution graph and attaching them to `ResolveException.failedPaths`. That allows unresolved warnings to show the full path from the missing module up to the root project.
It also adds and stabilizes `lm-coursier/unresolved-path` scripted coverage by:
- using a local test Maven repo fixture (no flaky remote test dependency)
- checking update stream output via an sbt task (`checkLog`) instead of shell `grep`
- asserting the unresolved path includes missing module, transitive caller, and root project
Co-authored-by: bitloi <89318445+bitloi@users.noreply.github.com>
Summary
- Fix `SbtParser.importsToLineRanges` to prepend `import` keyword when missing from Dotty's `Import` AST node source spans
- Add unit test for comma-separated import parsing
Co-authored-by: Renzo <170978465+RenzoMXD@users.noreply.github.com>
sbt 2.x allows `dependsOn(...)` between subprojects with mismatched
Scala versions without any warning or error. This can lead to confusing
classpath issues at compile or runtime, especially now that Scala 3.8+
has dropped backward TASTy compatibility with 2.13.
Per review feedback, move the Scala version mismatch check from
compileTask to projectDependenciesTask, where PR #8681 already handles
Scala version mixing logic. This provides earlier detection and keeps
the validation co-located with cross-version resolution.
Generated-by: Copilot
Co-authored-by: dev-miro26 <121471669+dev-miro26@users.noreply.github.com>
When using the arguments file (`@argsfile`) mechanism for forked runs,
double quotes inside arguments were not escaped, causing the JVM's
argument file parser to strip them. For example, passing `{"a":1}` as
an argument would result in `{a:1}`.
Escape `"` as `\"` in `createArgumentsFile`, matching the existing
backslash escaping, so the JVM correctly round-trips quoted arguments.
Fixessbt/sbt#7129
Co-authored-by: BrianHotopp <brihoto@gmail.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Treat explicit plugin toggles as last-call-wins for the same plugin.
This avoids contradictory include/exclude states when disablePlugins(X) is followed by
enablePlugins(X) (and vice versa), aligning behavior with normal override expectations.
Apply the same semantics to ProjectMatrix and add regression coverage:
- unit tests in main/src/test/scala/ProjectSpec.scala
- scripted test in sbt-app/src/sbt-test/project/i1926-disable-enable-plugin
When Test / autoScalaLibrary := false, build the non-fork test classloader
without the ScalaInstance layer: use Flat strategy and project classpath
only (rawCP), so tests do not see sbt's scala-library and avoid
NoSuchMethodError / version mismatch (e.g. scala/scala build).
- ClassLoaders.testTask: read autoScalaLibrary; when false use Flat and
rawCP-only fullCP, and do not add si.libraryJars to exclude.
- Add scripted test tests/autoScalaLibrary-test-loader that runs test
with Test / autoScalaLibrary := false.
**Problem**
`testFull` can select an aggregate-only key before a directly defined root key.
In issue #8772 this causes root project tests to be skipped, so a failing root test can pass unexpectedly.
**Solution**
Prefer directly defined keys in `Act.select`, and only fall back to aggregate-only candidates when no direct key exists.
Add scripted regression test `tests/i8772-root-project-testfull` and keep `project/extra-projects-key-aggregate` behavior intact.
Change the behavior when a cached task's output type contains
java.io.File: instead of silently skipping the cache, throw a
compile-time error with a message recommending xsbti.VirtualFileRef,
xsbti.HashedVirtualFileRef, or xsbti.VirtualFile, and linking to the
caching documentation.
Internal sbt tasks that return File types are wrapped with
Def.uncached to opt out of caching.
Fixes https://github.com/sbt/sbt/issues/8762
Generated-by: GitHub Copilot (Claude)
* [2.x] fix: Reject java.io.File as cached task output type
Use @transient on File-returning keys in Keys.scala instead of
wrapping tasks with Def.uncached in Defaults.scala. This ensures
build users who rewire these tasks also get caching skipped
automatically.
- Add @transient to 19 File-returning taskKey definitions in Keys.scala
- Revert Def.uncached wrappers from Defaults.scala
- Error at compile time when File is used as cached task output type
- Update error message to recommend @transient and link to docs
- Update scripted tests to use @transient approach
Fixes#8762
Generated-by: GitHub Copilot (Claude Opus 4.6)
**Problem**
`scalaOrganization` was ignored during compiler bridge resolution because `ZincLmUtil.getDefaultBridgeModule` hard-coded `ScalaArtifacts.Organization` (`org.scala-lang`).
**Solution**
Add a `scalaOrganization` parameter to `ZincLmUtil` methods, and passing `scalaOrganization` value to those methods.
Generated-by: Claude Opus 4.6
When `semanticdbEnabled := true` is set on Scala 2.x projects, the
`doc` task fails because `${CSR_CACHE}` placeholders in scalacOptions
(specifically the `-Xplugin:` path for the semanticdb compiler plugin)
are not resolved before being passed to Scaladoc.
This fix resolves virtual file references (containing $) in
scalacOptions before passing them to the Scaladoc bridge, matching
what zinc's MixedAnalyzingCompiler already does for compilation
(see sbt/zinc#1545).
Fixessbt/sbt#8740
Generated-by: GitHub Copilot (Claude Opus 4.6)