Stefan Bodewig wrote:

now that we've switched to JDK 1.4, we don't need JCE (and probably
some other things that have been swallowed by 1.4) as a standalone
project any longer.  If we really wanted to use bouncycastle's
version, we'd have to put it into the bootclasspath.  Is this worth
it?
I don't think so.

Using bountycastle's implementation has helped to identify a hardcoded
Sun provider reference in commons-httpclient a while ago, but with JDK
1.4 we wouldn't catch that as the provider is part of the core class
library anyway.
Given infinite resources (mostly people resources, less so of a machine resource consideration), there are all sorts of combinations that would be useful to explore. But my experience is that I have a hard enough time getting enough people to focus on one path, so my preference is not to divide up the effort.

Add to this the fact that we don't build the latest bouncycastle, so we can't use this to get early warning about future changes, and I suggest that it simply be dropped.

- Sam Ruby


--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to