On Fri, 5 Mar 2021 14:24:34 GMT, Claes Redestad <redes...@openjdk.org> wrote:
> This patch removes the CharacterData.isOtherUppercase and isOtherLowercase > methods. It also exploits the fact that isOtherUppercase is always false for > all codepoints in the CharacterDataLatin1 range for a small speed-up. > > I have no means to test if this is correct on PPC, which has intrinsics for > isLowerCase/isUpperCase, but unless I'm reading the code wrong the intrinsic > for isLowerCase on PPC already appears to effectively do the fused logic of > isLowerCase(ch) || isOtherLowerCase(ch) since it handles the two values where > isLowerCase and isOtherLowercase disagrees (0xaa, 0xba), which means this > change should make the intrinsic and the java code be in better agreement. Marked as reviewed by naoto (Reviewer). ------------- PR: https://git.openjdk.java.net/jdk/pull/2846
