Alex Bowden wrote: > So perhaps XJ3D may make VRML97.jar redundant after all! (because it was the > file loading that was so slow as to be the basic reason to keep it in use). > Well done.
Ta. I can't really take credit for it because it was an accident really...... There's a couple of small issues that we've found that have slowed the code up a fraction from the original email. We found a couple of corner cases where we did run into re-entrancy problems with the field parser. Our simple solution is just a global mutex, but we are going to work with parser pooling for the final M5 release to give us that speed back and also work nicely with multi-threading. In the way we have it now, we've effectively disabled the nice ability to parallelize the event model evaluation over multiple CPUs if the machine has them. > What would be the simplest way to give this a try. Is the early M5 expected > soon? Take a look at the other message just posted by Alan. There's a dev release we shoved out to the Xj3D website a couple of hours ago. Basically a zip file with the JARs in it. Grab, unzip, place in the dir where you already have Xj3D and you should be fine. (I hope.....). As usual, this is dev, not formal release, so YMMV as to whether it works as well as M4 or not. We actually have been through some very major rearrangements of code so that our codebase now roughly lines up with the X3D draft spec organisation of components and profiles. We've also re-organised the way capability bits are set and removed a bunch of bits from the setting that we don't think we need. Our tests seem to run the code cleanly, but there's is probably a bunch of corner cases that we haven't tested. From off the top of my head, the place where we are probably likely to encounter problems is in the grouping nodes. I rewrote the code so that if we detect a static load of the nodes (ie in a Loader where the LOAD_BEHAVIORS flag is not set) then we remove all the BranchGroups. A bunch of this functionality is not yet exposed to the end user, but we may run into problems somewhere. We plan to expose all of that to the end user by the real M5 release. -- Justin Couch http://www.vlc.com.au/~justin/ Java Architect & Bit Twiddler http://www.yumetech.com/ Author, Java 3D FAQ Maintainer http://www.j3d.org/ ------------------------------------------------------------------- "Humanism is dead. Animals think, feel; so do machines now. Neither man nor woman is the measure of all things. Every organism processes data according to its domain, its environment; you, with all your brains, would be useless in a mouse's universe..." - Greg Bear, Slant ------------------------------------------------------------------- =========================================================================== To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff JAVA3D-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
