Fwd: Re: [8u] RFR: 8129419: heapDumper.cpp: assert(length_in_bytes > 0) failed: nothing to copy

[Andreas Eriksson]

[Adding jdk8u-dev list]

Hi,

I believe I need an 8u Reviewer as well, since Dmitry is only committer 
for 8u?
Could someone please take a look?

This is type cleanup for dumping a heap with long arrays.

Hotspot webrev: 
http://cr.openjdk.java.net/~aeriksso/8129419/webrev.jdk8.00/hotspot/
JDK webrev: http://cr.openjdk.java.net/~aeriksso/8129419/webrev.jdk8.00/jdk/

See below for changes compared to the JDK 9 version.

Regards,
Andreas


-------- Forwarded Message --------
Subject: 	Re: [8u] RFR: 8129419: heapDumper.cpp: assert(length_in_bytes 
 > 0) failed: nothing to copy
Date: 	Wed, 10 Feb 2016 20:38:21 +0300
From: 	Dmitry Samersoff <dmitry.samersoff at oracle.com>
To: 	Andreas Eriksson <andreas.eriksson at oracle.com>, 
serviceability-dev at openjdk.java.net



Andreas,

Looks good for me!

-Dmitry

On 2016-01-28 20:36, Andreas Eriksson wrote:
> Hi,
>
> Please review this backport of JDK-8129419
> <https://bugs.openjdk.java.net/browse/JDK-8129419>.
>
> Hotspot webrev:
> http://cr.openjdk.java.net/~aeriksso/8129419/webrev.jdk8.00/hotspot/
> JDK webrev: http://cr.openjdk.java.net/~aeriksso/8129419/webrev.jdk8.00/jdk/
>
> There are three changes compared to JDK 9:
>
> 1) The hprof parser is in the JDK repo instead of top repo.
> Otherwise these changes are the same.
>
> 2) On Solaris os::write has some UseVMInterruptibleIO logic that failed
> because it assumed it was called by a JavaThread.
> I changed it to skip UseVMInterruptibleIO logic if called by a
> non-JavaThread.
>
> 3) The "heap dump file created" message were using sprintf and pragmas
> because of legacy reasons.
> Code after the change is the same as JDK 9 code.
>
> Regards,
> Andreas


-- 
Dmitry Samersoff
Oracle Java development team, Saint Petersburg, Russia
* I would love to change the world, but they won't give me the sources.