[
https://issues.apache.org/jira/browse/GROOVY-12685?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18137593#comment-18137593
]
ASF GitHub Bot commented on GROOVY-12685:
-----------------------------------------
codeconsole opened a new pull request, #3203:
URL: https://github.com/apache/groovy/pull/3203
The forked compiler for the core test sources runs with `groovycTest_mx=1g`.
On master, `:compileTestGroovy` fails with "Java heap space" at 900m and passes
at 1g, so there is little headroom. Master plus the closure fixes in #3171 to
#3202 fails at 1g (three runs) and passes at 1100m, and CI uses the value from
`gradle.properties`, so landing that batch would fail CI's test compile.
This raises `groovycTest_mx` to 1536m. `groovyJUnit_mx` is already 2g, and
the Gradle daemon is unaffected because the compiler runs in its own fork.
> Raise the heap of the forked compiler for the core test sources
> ---------------------------------------------------------------
>
> Key: GROOVY-12685
> URL: https://issues.apache.org/jira/browse/GROOVY-12685
> Project: Groovy
> Issue Type: Task
> Components: build
> Reporter: Scott Murphy Heiberg
> Priority: Major
>
> The forked compiler for the core test sources runs with {{groovycTest_mx=1g}}
> (gradle.properties). On master, {{:compileTestGroovy}} fails with "Java heap
> space" at 900m and passes at 1g, so there is little headroom: the
> closure-packing and closure-runtime fixes open as apache/groovy#3171 to #3202
> add enough test code that master plus all of them fails at 1g (three runs)
> and passes at 1100m. CI uses the default from gradle.properties, so landing
> that batch, or any other sizeable set of tests, will fail CI's test compile.
> Proposed: raise {{groovycTest_mx}} to 1536m. {{groovyJUnit_mx}} is already
> 2g, and the Gradle daemon's 2g is unaffected because the compiler runs in its
> own fork.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
