> > I disagree with the "Much simpler". Everytime a java programm just > execs an external programm "some how" handles its output and tries to > do something intelligent with it, somewhere a small cute kitten dies. >
As they do when you blindly assume some native library exists on a random user's machine on an unknown version of Linux, is a version you know how to talk to, and has kept binary compatibility with the way the IDE wants to talk to it, and won't, say, segfault the JVM instead. A script failing is much less dramatic. As someone who runs Gentoo Linux and configures my USE flags liberally to de-bloat my OS by compiling common libraries with many features I have no use for absent, I can tell you for sure that just because a library has the same name as the one you're looking for does *not* mean it contains all the calls a developer might think it does based on whatever happened to be on their machine when they wrote a feature. A script failing can't take the JVM with it. -Tim
