JMH 1.5.1

Aleksey Shipilev aleksey.shipilev at oracle.com
Tue Feb 3 07:24:09 UTC 2015


Hi Nils,

That usually happens when jmh dependencies versions are out of sync.
E.g. jmh-core is at 1.5, but jmh-generator-annprocess is 1.4.1. Please
check that.

I had just bootstrapped a Scala benchmark from
jmh-scala-benchmark-archetype, and it worked flawlessly on 1.5.1.

Thanks,
-Aleksey.

On 03.02.2015 06:23, Nils Kilden-Pedersen wrote:
> Since 1.5 I’ve been getting this error:
> 
> |java.lang.AbstractMethodError:
> org.openjdk.jmh.generators.annotations.APGeneratorDestinaton.getResource(Ljava/lang/String;)Ljava/io/Reader;|
> 
> Granted, I’m using Gradle and Scala, so it may be related to that, since
> no one else has reported this.
> 
>> 
> On Fri, Jan 30, 2015 at 12:36 PM, Aleksey Shipilev
> <aleksey.shipilev at oracle.com <mailto:aleksey.shipilev at oracle.com>> wrote:
> 
>     Hi,
> 
>     JMH 1.5.1 is released and available at Maven Central (thanks to Evgeny,
>     as usual). This is a routine maintenance release, and it hopefully does
>     not concern most of the audience.
> 
>     Summary of changes:
> 
>      * Sometimes, when the referenced (private) classes are not available
>     during annotation processing, javac throws a cryptic error and
>     terminates. Normally it means unsatisfied dependencies, but this also
>     affects some trivial cases like referencing a standard Swing's JFrame in
>     a JMH benchmark. We can work this around, fixed as:
>        https://bugs.openjdk.java.net/browse/CODETOOLS-7901278
> 
>      * Up to this point, SecurityManager performance tests appeared to work
>     by coincidence. We did a few tight-ups in the JMH code, done more tests,
>     and provided samples how to use this properly:
>        https://bugs.openjdk.java.net/browse/CODETOOLS-7901280
> 
>     http://hg.openjdk.java.net/code-tools/jmh/file/tip/jmh-samples/src/main/java/org/openjdk/jmh/samples/JMHSample_33_SecurityManager.java
> 
>      * Incremental compilation support, that was added recently, prints the
>     warning messages when it overwrites the existing benchmark entries in
>     the metadata file. This trips projects that require -Xlint:all -Werror,
>     and we should really make these messages diagnostic. Fixed:
>        https://bugs.openjdk.java.net/browse/CODETOOLS-7901284.
> 
>      * More tests, as prompted by CODETOOLS-7901278:
>        https://bugs.openjdk.java.net/browse/CODETOOLS-7901279
> 
>     Enjoy!
> 
>     Thanks,
>     -Aleksey.
> 
> 




More information about the jmh-dev mailing list