Forum issue: Could not create SecurityManager

[Luke Daley-2]

http://forums.gradle.org/gradle/topics/java_lang_internalerror_could_not_create_securitymanager_jarjar_org_gradle_process_internal_child_bootstrapsecuritymanager

Anyone have any ideas?

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[Radim Kubacki]

Seems weird to me. We're using JarJar to create self-contained Jar for toolingApi but these internal classes are not repackaged and even then the repackaging does not use 'jarjar.' prefix (it has 'org.gradle.jarjar').

My ideas: 1) there is another JarJar involved (part of Jenkins plugin or their own code) and it screws thing up 2) as a part of our JarJar-ing something is not processed correctly (something like META-INF/services). Is it the only report of this kind?

-Radim

On Mon, Feb 10, 2014 at 11:56 PM, Luke Daley <[hidden email]> wrote:

http://forums.gradle.org/gradle/topics/java_lang_internalerror_could_not_create_securitymanager_jarjar_org_gradle_process_internal_child_bootstrapsecuritymanager

Anyone have any ideas?

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

   http://xircles.codehaus.org/manage_email

[Adam Murdoch]

On 10 Feb 2014, at 2:56 pm, Luke Daley <[hidden email]> wrote:

http://forums.gradle.org/gradle/topics/java_lang_internalerror_could_not_create_securitymanager_jarjar_org_gradle_process_internal_child_bootstrapsecuritymanager

Anyone have any ideas?

There are possibly several issues here, all with the same symptom. The jvm does a poor job of writing any diagnostics when something goes wrong creating the system SecurityManager, which we use to bootstrap the test process.

You might start with getting someone to run with -i, so we can see exactly what jvm args ended up being used to launch the process.

It’s not necessarily a problem with the jvm args, though. There might be a problem in our caching so that two processes are stomping over each other when they create the small jar we create to bootstrap, so that the problem happens in one build but then goes away.

--

Adam Murdoch
Gradle Co-founder
http://www.gradle.org
VP of Engineering, Gradleware Inc. - Gradle Training, Support, Consulting
http://www.gradleware.com

[Luke Daley-2]

Adam Murdoch wrote:

> On 10 Feb 2014, at 2:56 pm, Luke Daley <[hidden email]
> <mailto:[hidden email]>> wrote:
>
> - hide quoted text -- show quoted text -
>> http://forums.gradle.org/gradle/topics/java_lang_internalerror_could_not_create_securitymanager_jarjar_org_gradle_process_internal_child_bootstrapsecuritymanager
>>
>> Anyone have any ideas?
>
> There are possibly several issues here, all with the same symptom. The
> jvm does a poor job of writing any diagnostics when something goes
> wrong creating the system SecurityManager, which we use to bootstrap
> the test process.
>
> You might start with getting someone to run with -i, so we can see
> exactly what jvm args ended up being used to launch the process.

Have asked for this.
> It’s not necessarily a problem with the jvm args, though. There might
> be a problem in our caching so that two processes are stomping over
> each other when they create the small jar we create to bootstrap, so
> that the problem happens in one build but then goes away.
Would a logical step be adding some test coverage to verify this isn't
happening?

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email