Raif S. Naffah wrote: > the attached patch adds support for GNU MP in BigInteger > if/when configured.
How/why is the native version better? Is it really worthwhile to complicate the code this way? Where are the benchmarks that prove the native code is faster? I assume it is already widely know that I dislike native code (because of how IKVM works), but let me try to explain why native code sucks for all VMs: - Some VMs are hardened against stack overflow, this means that stack overflow in Java or VM code will not result in terminating the process, but simply a StackOverflowException. Is the native code you are introducing similarly harneded? - Some VMs are hardened against OOM, this means that memory allocation failures (both in Java and in the VM) will not result in terminating the process, but simply a OutOfMemoryError. Is the native code you are introducing similarly harneded? - Native memory allocation potentially fragments the heap and this can have significant performance implications. - Can you prove there aren't any security holes in your native code? Most people do not understand the (security) implications of using native code in combination with handles and most of the code out there is simply wrong (as is the current patch). See [1]. Regards, Jeroen [1] http://www.hpl.hp.com/personal/Hans_Boehm/misc_slides/java_finalizers.pd f
