In message <000d01c2611e$922bc390$0d00a8c0@luckplane>, "Dean Arnold" writes: >So the $64K question: is the latest version J2ME compatible ?
No. >If not, how difficult would it be to make a J2ME >compatible version ? It's very easy to make the source code modifications if you limit yourself to the .regex package (and maybe the .perl as well). The hard part is to set up an acceptable build system and conditional compilation mechanism to simultaneously support multiple versions for the various incarnations of the Java platform. Which reminds me, I need to get back to Didge about his proposed Ant task that uses Velocity macros for conditional compilation. At any rate, the changes for J2ME may be a little more than they used to be. It used to be that all you needed to do was remove 'implements java.io.Serializable' and replace Character.isWhitespace(char) calls with a custom implementation (see http://www.savarese.org/talks/hjug-042000/img24.htm). Now you would also have to change the Hashmaps to Hashtables, the ArrayLists to Vectors, and there may be some issue with references to Collections Framework interfaces. At any rate, it's not infeasible to support J2ME and JDK 1.1.x with #ifdefs, but there has not been sufficient demand to motivate the process of evaluating and deciding upon an appropriate way to handle conditional compilation. We've talked about it and I've ranted about Java's lack of real conditional compilation since 1996. But you may be the first person with enough of an itch to start scratching. In any case, J2ME support is another way to distinguish ORO from J2SE 1.4's java.util.regex and I think it's a good thing. So, as Jon is fond of saying, you just volunteered for the job. daniel -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
