[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: A query on the classloader architecture
From: |
Stuart Ballard |
Subject: |
Re: A query on the classloader architecture |
Date: |
Fri, 09 Jul 2004 11:52:17 -0400 |
User-agent: |
Mozilla Thunderbird 0.6 (X11/20040605) |
Jeroen Frijters wrote:
I'm sorry, but you lost me. I clearly don't understand what you mean.
Can you please clarify your original statement (the one I quoted above)?
I think the point is that Sun's VM returns null for stuff that (in IKVM)
comes from IKVM.GNU.Classpath.dll, but not for anything else. In
particular, Sun's VM will never return null for the classloader of any
application code.
Perhaps a simple rule that might have the effect you want would be to
use the null classloader for the following assemblies:
1) IKVM.Runtime.dll
2) IKVM.GNU.Classpath.dll
3) IKVM.JNI.*.dll
4) Anything in the GAC
and a non-null classloader for everything else.
Am I right that assemblies in the GAC must never have references to
other assemblies that aren't? And that IKVM-compiled Java assemblies are
never suitable for GACing? If both these things are true, then no class
in the set listed above can ever refer to any class outside that set
(assuming the references of the IKVM assemblies are limited to each
other and stuff in the GAC).
Stuart.
--
Stuart Ballard, Senior Web Developer
NetReach, Inc.
(215) 283-2300, ext. 126
http://www.netreach.com/
- Re: A query on the classloader architecture, (continued)
RE: A query on the classloader architecture, Jeroen Frijters, 2004/07/09
RE: A query on the classloader architecture, Jeroen Frijters, 2004/07/09
RE: A query on the classloader architecture, Jeroen Frijters, 2004/07/09
RE: A query on the classloader architecture, Jeroen Frijters, 2004/07/09
RE: A query on the classloader architecture, Jeroen Frijters, 2004/07/11