>>>>> "Casey" == Casey Marshall <[EMAIL PROTECTED]> writes:
Casey> Yeah, it won't work to have both implementations in the provider, Casey> anyway (the second one added would just override the first). I didn't Casey> remove the old MD5 or SHA-1 classes because I know some code uses Casey> those classes directly (I think in GCJ, there may be others). The Casey> versions from GNU Crypto should be faster, or at least just as fast; Casey> we did put effort into optimizing these hashes. I think for Classpath it would be fine to pick the best implementation, and delete the others. Gcj can be fixed up easily enough at the next import, or can maintain a small divergence here if need be. (FWIW gcj uses gnu.java.security.provider.MD5 directly... but we don't depend crucially on a particular implementation.) Tom _______________________________________________ Classpath mailing list [email protected] http://lists.gnu.org/mailman/listinfo/classpath
