When a file referenced by a task's inputs/outputs (e.g. Compile / resources +=
file("nope.txt")) does not exist, hashing the task's cache key threw a
NoSuchFileException deep inside sjsonnew serialization. It surfaced as an opaque
sjsonnew.SerializationException that dumped the entire input list, with the real
cause buried several `Caused by:` levels down, so users routinely mistook it for
a corrupt cache and reached for `clean`.
ActionCache.mkInput now catches the hashing failure, detects a NoSuchFileException
anywhere in the cause chain (ActionCache.findMissingFile), and throws a
MessageOnlyException naming the file:
[error] file referenced by the build does not exist: nope.txt
util-cache gains a dependency on util-control (a leaf module, no cycle) for
MessageOnlyException.
Fixes#9217.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
update held a process-global lock around coursier resolution and artifact
fetching whenever a logger was set OR coursier was not in fallback mode. That
lock exists only to serialize coursier's interactive progress bar, which is
rendered solely when no custom logger is supplied and coursier is not in
fallback mode. The loggerOpt.nonEmpty clause therefore over-serialized the
common non-interactive case (IntelliJ re-imports, CI, any non-TTY run, where
sbt supplies a quiet debug-only logger), making update scale with the number
of modules rather than the number of distinct artifacts.
Fixes#5508.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
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.
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: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fixes#9009.
dependencyMode := DependencyMode.PlusOne strips every internal
project dependency from the compile classpath, even direct ones --
LibA/compile fails with Not found: CoreClass despite
LibA.dependsOn(Core). Direct mode happens to work today, but only by
accident (filterByDirectDeps matches against allDependencies, which
includes projectDependencies -- the bug surface is one step removed).
Root cause: ClasspathImpl.filterByPlusOne resolves the
direct-dep key set from UpdateReport, which only contains externally
resolved (LM) modules. Internal project refs never appear there, so
resolvedDirectKeys is empty for internal entries; plusOneKeys is
likewise empty; the final filter strips every internal classpath
entry whose moduleIDStr identifies an internal project. Diagnosis on
the issue by eureka0928.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fixes#6886.
dependencyTree, dependencyBrowseTree, and inspect tree re-explore
the same node once per incoming edge. In a DAG with N levels and M
children per node the rendered output is O(M^N) -- the OP needed
>16 GB heap, #7360 has a 6 GB heap dump, and Friendseeker's analysis
on the issue showed the exponential re-traversal directly.
Fix: track a visited set across the renderer's recursion. The first
time a node is encountered it is rendered in full; on subsequent
visits the entry collapses to a one-line +- <id> (*) (ASCII) or a
<id> (*) leaf (JSON), matching Maven's dependency:tree (*)
convention. Cycle detection (separate parents set, (cycle) marker)
is unchanged.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fixes#2998
When test is run at the prompt in a multi-project build,
sbt aggregates <proj>/test across subprojects. Each subproject's
TestResultLogger emits its own pass/fail summary inline, but the
aggregation layer never restated which projects failed. The final
"sbt.TestsFailedException" line surfaced no project names or failing
test classes, so on CI users had to scrollback-grep through thousands
of log lines to find what broke.
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
**Problem**
We need some tooling to debug caching issues.
**Solution**
This adds an exeprimental execution log support,
which shows input and output of cached tasks.
**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
Bumps io to released 1.12.0 which contains the atomic-write changes
to IO.copyFile, IO.transfer(InputStream, File), and IO.jar/zip.
ActionCacheStore.putBlob now uses IO.transfer instead of duplicating
staging logic.
Eugene Yokota
kenji yoshida
BrianHotopp
Anatolii Kmetiuk
Anatolii Kmetiuk