Re: Classpath JNI changes - take 2

[Brian Jones]

Stephen Crawley <address@hidden> writes:

> I tried Stuart Ballard's suggestion of putting all JNI calls EXCEPT
> for the Kissme implemented ones int GC regions.  It works.

Good.

> With that in mind, the attached classpath.diffs.gz files give my 
> revised proposal for changes to jni.h in Classpath.  Basicly, the
> changes allow the following:
> 
>   1  They define placeholders for the JDK 1.4 extensions to JNI

Why placeholders?  Why not the actual methods themselves?

>   2  They allow JVM implementations to #include jni.h to avoid 
>      mismatched function tables.

This has to do with the methodID/fieldID changes?

>   3  They allow JVM implementations to define custom extensions  
>      to the function table.

Why would a JVM choose to expose functionality in this way?  For
example could there be a kissme.h instead to accomplish the same
thing?

Brian
-- 
Brian Jones <address@hidden>
http://www.haphazard.org/~cbj/