If I understood you idea correctly, that would be thread-unsafe
In general before we get rid of that class or decide to keep it, I think we should work out an approach to performance optimizations. It's a rare case when perofmance might be significantly improved for free, a regular price is: additional complexity, more code, lack of readability, additional memory footprint etc. Performnace optimization is not only looking for hotspots and tuning them but also a hard everyday work on tuning just a regular stuff. Though at the same time we can't afford megabytes of code that is a perofomace optimization for a single never-used scenario We develop a system, rather than a regular application and different users will run different apps on it. That is why users may have hotspots differ from those that we can find with benchmarks we can have. So let's discuss how we approach performance in general and that return to this specific class. Thanks, Mikhail 2006/7/24, Anton Luht <[EMAIL PROTECTED]>:
Hello, > * As Alex and other have mentioned, just use a simple wrapper object with > correct equals/hashCode. There can be a single instance of such object - not creation of a new key before lookup but assigning two keys' values :) -- Regards, Anton Luht, Intel Middleware Products Division --------------------------------------------------------------------- Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
--------------------------------------------------------------------- Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
