Wayne Meissner wrote:
Of course, a gym bag full of cash is always useful in persuading me to
do the extra work and change the license to whatever is most
convenient.

There ya go, Fletch, just cough up some dough.

JFFI in jruby isn't a short term solution anyway - its not going to go
in until after the alleged branching off of jruby-2.0 (aka
JRuby-Forever :)

Actually, as long as we don't break compatibility, there's no reason JFFI couldn't replace JNA sooner. It's just work.

We really need to make JRuby _not_ depend on native code, since there
are platforms like OpenBSD, NetBSD, FreeBSD, etc that would require
jna-posix and JNA/JFFI to be ported to them, and that will become a
huge PiTA overhead if we had to support each one with native code.

So it's probably good for "the long now" of JRuby if we do the extra lifting to isolate native bits and failover. jna-posix already does this, so we should probably look at doing similar for other APIs. In many cases, it may make sense to simply move those libraries into jna-posix, like unix sockets. Or maybe spawn a few new libraries and conditionally load them when on platforms that support them.

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

   http://xircles.codehaus.org/manage_email


Reply via email to