At 6:56 PM -0300 6/9/11, Jim Laskey wrote:
>Thank you. Do you know which OpenJDK build this corresponds to - b146?
>
>Cheers,
>
>-- Jim
>
mlvm is based off of bsd-port and looking here the last tag bsd-port merged was
jdk7-b145:
http://hg.openjdk.java.net/bsd-port/bsd-port/jdk
___
I've been testing the new version with Dynalink, which "traditionally" had test
failures due to bugs in MLVM.
The good news is that of 6 formerly failing tests concerned with vararg
invocation, 4 now pass.
The less good news is that two still fail, although with different messages.
The bad
On Jun 9, 2011, at 2:46 PM, Stephen Bannasch wrote:
>> On Jun 9, 2011, at 12:55 PM, Stephen Bannasch wrote:
>>
>>> -XX:+EnableInvokeDynamic -XX:-UseRicochetFrames
>>
>> These flags are now obsolete. I suggest testing without them. -- John
>
> Thanks, that fixed the tests -- all jdk/test/java
On 06/10/2011 12:15 AM, Szymon Jachim wrote:
Hello,
I'm trying to run one of the examples using JDK 1.7 preview (build
145) and I'm getting:
"Exception in thread "main" java.lang.ClassFormatError: Class file
version does not support constant tag 15 in class file[...]"
Which build of JDK7 an
Hello,
I'm trying to run one of the examples using JDK 1.7 preview (build 145)
and I'm getting:
"Exception in thread "main" java.lang.ClassFormatError: Class file version
does not support constant tag 15 in class file[...]"
Which build of JDK7 and ASM framework should I use to experiment with t
>On Jun 9, 2011, at 12:55 PM, Stephen Bannasch wrote:
>
>> -XX:+EnableInvokeDynamic -XX:-UseRicochetFrames
>
>These flags are now obsolete. I suggest testing without them. -- John
Thanks, that fixed the tests -- all jdk/test/java/lang/invoke tests now pass
with this invocation:
$ jtreg -XX:+Un
I uploaded the latest mlvm built on Mac OS X 10.6.7.
http://www.concord.org/~sbannasch/mlvm/java-1.7.0-internal-mlvm-2011_06_09.tar.gz
I excluded the coro patch because of compilation failures:
export davinci=$(pwd) guards="buildable testable /coro"
All jdk/test/java/lang/invoke test pas
On Jun 9, 2011, at 12:55 PM, Stephen Bannasch wrote:
> -XX:+EnableInvokeDynamic -XX:-UseRicochetFrames
These flags are now obsolete. I suggest testing without them. -- John
___
mlvm-dev mailing list
mlvm-dev@openjdk.java.net
http://mail.openjdk.java
Tests need to be updated to be in sync with the state of the implementation.
Rémi
On 06/09/2011 09:55 PM, Stephen Bannasch wrote:
> I'm getting 5 failed java/lang/invoke tests with the latest mlvm code.
>
> I'm setting up my guards this way to exclude coro which now causes a
> compilation failur
I´m a lurker here... so sorry for making this my first post.
In case you guys haven´t heard, an old but "undeservedly unknown
programming language" (sic), IBM NetRexx
(Rexx running on the JVM) is now open source. Since yesterday, at least.
http://www.brightsideofnews.com/news/2011/6/8/ibm-hand
I'm getting 5 failed java/lang/invoke tests with the latest mlvm code.
I'm setting up my guards this way to exclude coro which now causes a
compilation failure on my Mac -- see
https://gist.github.com/1017562 for coro failure.
export davinci=$(pwd) guards="buildable testable /coro"
I'm runn
Not to temper your enthusiasm, but the obvious answer here is that the
perf gains are largely from avoiding a reflective invocation. To me
that doesn't make the results any less spectacular, since they were
far easier to wire in than hand-generated stubs (which Groovy does and
which JRuby can do if
On Thu, Jun 9, 2011 at 12:50 PM, Christian Thalinger
wrote:
> Sorry I'm a bit behind but is this still a problem or have you fixed it
> already?
>
> It looks like it works:
>
> @ 24
> java.lang.invoke.MethodHandle::invokeExact (32 bytes) inline (hot)
>
On Jun 7, 2011, at 10:44 PM, Charles Oliver Nutter wrote:
> That's pretty much it. The test I ran looks like this:
>
> def foo
> 1
> end
>
> 100_000.times {
> foo
> }
>
>
> The body of the block passed to "times" jits (in Hotspot) and appears
> to inline the "foo" method just fine, but the bod
On Jun 8, 2011, at 11:10 PM, Attila Szegedi wrote:
> Woo-hoo!
I agree :-)
>
> On Jun 8, 2011, at 9:48 PM, Charles Oliver Nutter wrote:
>
>> Hello friends! I have another update!
>>
>> I've just landed preliminary work to make JRuby directly bind Ruby to
>> Java calls that were normally done vi
15 matches
Mail list logo
