On Mon, 8 Apr 2024 16:47:24 GMT, Paul Sandoz <psan...@openjdk.org> wrote:
>> @PaulSandoz @AlanBateman I've added a commit to this PR which removes the >> use of Gatherer for Stream::flatMap, but instead implements flatMap for all >> of the pipelines using the same encoding which Gatherer would use. It seems >> very competitive performance-wise, and resolves at least one open JBS-issue >> with flatMap (will look to see if it resolves more than that) > >> @PaulSandoz @AlanBateman I've added a commit to this PR which removes the >> use of Gatherer for Stream::flatMap, but instead implements flatMap for all >> of the pipelines using the same encoding which Gatherer would use. It seems >> very competitive performance-wise, and resolves at least one open JBS-issue >> with flatMap (will look to see if it resolves more than that) > > That's a rather clever use of `allMatch`! > > Did you observe performance improvements using `@Stable` on the `cancel` > field? It's really hard to predict in the abstract (since the default value > of the field will be read in proportion to the number of elements until the > stream is cancelled). @PaulSandoz @AlanBateman This should be ready for review now, if you find some time :) ------------- PR Comment: https://git.openjdk.org/jdk/pull/18625#issuecomment-2051317333
