RFR (M) : JDK-8016029 : test runtime/6878713/Test6878713.sh failed
Mikhailo Seledtsov
mikhailo.seledtsov at oracle.com
Fri Jul 19 08:40:36 PDT 2013
Hi Remi,
Thank you for advice.
Misha
On 7/18/2013 4:35 PM, Remi Forax wrote:
> One ASM developer here,
> Mikhailo, you can use the ASMifier that creates, from a class file, a
> Java file that will generate the class file
> (you will just have to change the imports to use the ASM version
> embedded in the jdk).
> see http://asm.ow2.org/doc/faq.html#Q10
>
> cheers,
> Rémi
>
> On 07/18/2013 09:59 PM, John Rose wrote:
>> Good; thanks, Misha. — John
>>
>> On Jul 18, 2013, at 12:55 PM, Mikhailo Seledtsov
>> <mikhailo.seledtsov at oracle.com> wrote:
>>
>>> Hi John, Coleen,
>>>
>>> John, thank you for the advice on using
>>> jdk.internal.org.objectweb.asm.*. I did not know about this capability.
>>> Hopefully we can use this approach in this particular test, as well
>>> as other similar tests. This could be a great help for rewriting the
>>> bytecode based tests in a source form. In case if for some reason it
>>> does not work out, I will include the jasm as a separate file, with
>>> my comments.
>>> Unfortunately the source jasm file was not available for this
>>> test, only a jar file. This is the case for many other similar
>>> byte-code based tests. I did use tools to disassemble the jar into a
>>> jasm, in order to look at it and to understand what the code does.
>>> Hence, the only jasm I could include is the reverse-engineered one;
>>> however I agree that it is still better than having binary in the
>>> source, since I can include some comments with jasm code explaining
>>> what it does.
>
More information about the hotspot-runtime-dev
mailing list