If we do down this path, we should look at replacing (or enhancing) our UUID implementation with the UUID class that is now part of JDK 5 ( http://java.sun.com/javase/6/docs/api/java/util/UUID.html). I just heard from a user that the random number version of the UUID (version 4) provides for more efficient index processing. This is part of the Java implementation. We should take a look at that as an alternative.
Kevin On Sun, Mar 16, 2008 at 8:26 PM, Kevin Sutter <[EMAIL PROTECTED]> wrote: > Patrick, > I would be +1 for getting rid of the JDK 1.4 requirement. > > Thanks, > Kevin > > > On Sun, Mar 16, 2008 at 2:52 PM, Evan Ireland <[EMAIL PROTECTED]> wrote: > > > Sounds like a good plan, based on the JDK 1.4 EOL schedule. > > > > http://java.sun.com/j2se/1.4.2/ > > > > > -----Original Message----- > > > From: Patrick Linskey [mailto:[EMAIL PROTECTED] > > > Sent: Sunday, 16 March 2008 12:06 p.m. > > > To: [email protected] > > > Subject: JDK 1.4 support? > > > > > > Hi, > > > > > > What do you guys think about discontinuing our JDK 1.4 > > > support? I've made a few attempts to make our concurrent map > > > utilization pluggable to allow us to take advantage of the > > > JDK improvements in concurrency, but have had a hard time > > > doing so in the past. It occurs to me that one easy way to do > > > this would be to just get rid of JDK 1.4 support in OpenJPA > > > 1.1, which would let us switch directly to the Java 5 > > > java.util.concurrent package, without requiring cumbersome > > > wrappers etc. > > > > > > Thoughts? > > > > > > -Patrick > > > > > > -- > > > Patrick Linskey > > > 202 669 5907 > > > > > > > >
