On Tue, 4 Mar 2025 19:37:32 GMT, Vladimir Ivanov <vaiva...@openjdk.org> wrote:
>> test setup was updated to generate data of requested size. > > Vladimir Ivanov has updated the pull request incrementally with one > additional commit since the last revision: > > JDK-8350811 [JMH] test foreign.StrLenTest failed with > StringIndexOutOfBoundsException for size=451 test/micro/org/openjdk/bench/java/lang/foreign/StrLenTest.java line 149: > 147: while (lorem.length() < size) { > 148: lorem += lorem; > 149: } This is matter of taste, but I would prefer StringBuilder instead: StringBuilder builder = new StringBuilder(lorem.length()+size); for (int l = 0; l<size; l+=lorem.length()) { builder.append(lorem); } return builder.substring(0, size); test/micro/org/openjdk/bench/java/lang/foreign/StrLenTest.java line 181: > 179: > 180: static class SlicingPool { > 181: final MemorySegment pool; This seems ok; though I was suspicious but I can't measure a difference: with the default parameters for size of 5, 20, 100, arena was 10x-200x+ times bigger. Now it fits exactly. I measured 1x, 10x, 100x (for size 1500 and 15000) and results appear consistent. So, this is just a note; looks good ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/23873#discussion_r1982330072 PR Review Comment: https://git.openjdk.org/jdk/pull/23873#discussion_r1982357391