diff options
author | Stefan Zeiger <szeiger@novocode.com> | 2016-08-12 18:48:04 +0200 |
---|---|---|
committer | Stefan Zeiger <szeiger@novocode.com> | 2016-08-12 18:48:04 +0200 |
commit | 00617591ab87d98514c9e60a267ae69bb45d1d11 (patch) | |
tree | 7cc8a27cae3f50e601b72f15008557093ce34af1 | |
parent | 366f45b54938895be74b248af4ba69f5976d639a (diff) | |
download | scala-00617591ab87d98514c9e60a267ae69bb45d1d11.tar.gz scala-00617591ab87d98514c9e60a267ae69bb45d1d11.tar.bz2 scala-00617591ab87d98514c9e60a267ae69bb45d1d11.zip |
Split “partest run” off from “partest pos neg jvm”
We’ve seen several OOM failures in “run” tests lately. Maybe going back
to more separate partest calls will help. Now that everything is
launched from the same sbt instance and test results are always
accumulated, this should not have any negative impact on build
performance or usability.
-rw-r--r-- | build.sbt | 6 |
1 files changed, 4 insertions, 2 deletions
@@ -771,7 +771,8 @@ lazy val root: Project = (project in file(".")) testAll := { val results = ScriptCommands.sequence[Result[Unit]](List( (Keys.test in Test in junit).result, - (testOnly in IntegrationTest in testP).toTask(" -- run pos neg jvm").result, + (testOnly in IntegrationTest in testP).toTask(" -- run").result, + (testOnly in IntegrationTest in testP).toTask(" -- pos neg jvm").result, (testOnly in IntegrationTest in testP).toTask(" -- res scalap specialized scalacheck").result, (testOnly in IntegrationTest in testP).toTask(" -- instrumented presentation").result, (testOnly in IntegrationTest in testP).toTask(" -- --srcpath scaladoc").result, @@ -789,7 +790,8 @@ lazy val root: Project = (project in file(".")) // All attempts to define these together with the actual tasks due to the applicative rewriting of `.value` val descriptions = Vector( "junit/test", - "partest run pos neg jvm", + "partest run", + "partest pos neg jvm", "partest res scalap specialized scalacheck", "partest instrumented presentation", "partest --srcpath scaladoc", |