You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@harmony.apache.org by Pavel Pervov <pm...@gmail.com> on 2007/02/16 14:17:46 UTC

Re: [classlib] compatibility nuances

<Copying from JIRA>

I've experimented with RI VMs.

Jrockit uses 'stack list' to store methods it reads from class file (last
declared comes first).
HotSpot seems to use some sort of hash map to store method. Statics are
always reported first, moving 'static public void main' from the beginning
of the class to the end changes the order of reporting non-static methods
drastically.

So, "nice" is nice, but I would suggest to treat this as non-bug difference
and close HARMONY-953 as such.

WBR,
-- 
Pavel Pervov,
Intel Enterprise Solutions Software Division

Re: [classlib] compatibility nuances

Posted by Gregory Shimansky <gs...@gmail.com>.
On Friday 16 February 2007 16:59 Zakharov, Vasily M wrote:
> I've also checked the issue, using my experience of the same kind
> with RMI Compiler, and I also came to conclusion that figuring RI
> algorithm out in this case is hopeless.
>
> So +1 for closing H-953 as NBD.

+1 here.
Closed it right now.

> -----Original Message-----
> From: Pavel Pervov [mailto:pmcfirst@gmail.com]
> Sent: Friday, February 16, 2007 4:18 PM
> To: harmony-dev@incubator.apache.org
> Subject: Re: [classlib] compatibility nuances
>
> <Copying from JIRA>
>
> I've experimented with RI VMs.
>
> Jrockit uses 'stack list' to store methods it reads from class file
> (last
> declared comes first).
> HotSpot seems to use some sort of hash map to store method. Statics are
> always reported first, moving 'static public void main' from the
> beginning
> of the class to the end changes the order of reporting non-static
> methods
> drastically.
>
> So, "nice" is nice, but I would suggest to treat this as non-bug
> difference
> and close HARMONY-953 as such.
>
> WBR,

-- 
Gregory

RE: [classlib] compatibility nuances

Posted by "Zakharov, Vasily M" <va...@intel.com>.
I've also checked the issue, using my experience of the same kind
with RMI Compiler, and I also came to conclusion that figuring RI
algorithm out in this case is hopeless.

So +1 for closing H-953 as NBD.

 Vasily


-----Original Message-----
From: Pavel Pervov [mailto:pmcfirst@gmail.com] 
Sent: Friday, February 16, 2007 4:18 PM
To: harmony-dev@incubator.apache.org
Subject: Re: [classlib] compatibility nuances

<Copying from JIRA>

I've experimented with RI VMs.

Jrockit uses 'stack list' to store methods it reads from class file
(last
declared comes first).
HotSpot seems to use some sort of hash map to store method. Statics are
always reported first, moving 'static public void main' from the
beginning
of the class to the end changes the order of reporting non-static
methods
drastically.

So, "nice" is nice, but I would suggest to treat this as non-bug
difference
and close HARMONY-953 as such.

WBR,
-- 
Pavel Pervov,
Intel Enterprise Solutions Software Division