avoid full (re)compilation after cold scalafix invocations #411
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
bjaglin
force-pushed
the
store-analysis
branch
2 times, most recently
from
65df750
to
d63114b
Compare
bjaglin
changed the title
bjaglin
commented
Apr 22, 2024
Comment on lines
-751
to
-754
| // prevent storage of the analysis with relaxed scalacOptions - despite not depending explicitly on compile, | ||
| // it is being triggered for parent configs/projects through evaluation of dependencyClasspath (TrackAlways) | ||
| // in the scope where scalafix is invoked | ||
| analysis.withHasModified(false) |
There was a problem hiding this comment.
this turned out to be unnecessary, since there is no scenario where a relaxed invocation would override a prior, regular one and invalidate the cache
bjaglin
changed the title
bjaglin
changed the title
bjaglin
commented
Apr 22, 2024
| @@ -583,7 +585,7 @@ object ScalafixPlugin extends AutoPlugin { | |||
| case "--stdout" => | |||
| // --stdout cannot be cached as we don't capture the output to replay it | |||
| throw StampingImpossible | |||
| case "--tool-classpath" => | |||
| case tcp if tcp.startsWith("--tool-classpath") => | |||
There was a problem hiding this comment.
this PR exposed that a scripted was passing for the wrong reason: a scalafix invocation was failing because of a cache miss recompilation while it should have failed in scalafix itself
sbt-scalafix/src/sbt-test/skip-windows/caching/test
Lines 186 to 193 in 9ceb2c5
|
transient failure in inconfig scripted |
bjaglin
commented
Apr 22, 2024
| @@ -56,7 +56,7 @@ scriptedSbt := { | |||
| if (jdk >= 21) | |||
| "1.9.0" // first release that supports JDK21 | |||
| else | |||
| (pluginCrossBuild / sbtVersion).value | |||
| "1.3.3" // get https://github.com/sbt/sbt/issues/1673 to avoid race conditions | |||
There was a problem hiding this comment.
See #411 (comment). I can't tell for sure, but I think that storing analysis files slightly changed the timing of concurrent executions of compile, causing that new transient failure.
bjaglin
force-pushed
the
store-analysis
branch
2 times, most recently
from
ccdab2b
to
67ce03d
Compare
bjaglin
changed the title
bjaglin
changed the title
bjaglin
changed the title
bjaglin
commented
Apr 23, 2024
Comment on lines
+5
to
+6
| > compile | ||
| > checkLastCompilationCached |
There was a problem hiding this comment.
users with no fatal warnings calling scalafix before compile no longer have to compile everything twice
bjaglin
commented
Apr 23, 2024
Comment on lines
+23
to
+26
| > scalafix --check RemoveUnused | ||
| -> checkLastCompilationCached | ||
| > scalafix --check RemoveUnused | ||
| > checkLastCompilationCached |
There was a problem hiding this comment.
successive scalafix invocations on a cold environment (where no compile was run successfully) now benefit from incremental compilation, no matter which flags are used in the build
bjaglin
commented
Apr 23, 2024
Comment on lines
-41
to
-44
| TaskKey[Unit]("checkZincAnalysisPresent") := { | ||
| val isPresent = (Compile / previousCompile).value.analysis.isPresent() | ||
| assert(isPresent, "zinc analysis not found") | ||
| } |
There was a problem hiding this comment.
I removed this helper for whitebox testing since the blackbox one is enough
The Zinc analysis is now saved on compilations triggered by scalafix invocations. This allows subsequent scalafix invocations not to re-trigger compilations, and subsequent regular compilations to be either incremental (in case the build contains `-Xfatal-warnings` or alike) or a full cache hit otherwise.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Follow-up of https://twitter.com/channingwalton/status/1778119590486118908 which puts the finger on a behavior that can be very irritating in large builds
This revisits #152, leveraging
scalafixInvoked(which has the same semantics as thescalafixRunExplicitlymentioned in the old PR, and didn't cause any issue for 3 years) in order to trigger a regularcompilefromscalafixand thus store zinc analysis to benefit later compilations.