Looks good. ForkJoinPool.java —
356 * readers must tolerate null slots. Worker queues are at odd 357 * indices. Shared (submission) queues are at even indices, up to 358 * a maximum of 64 slots, to limit growth even if array needs to 359 * expand to add more workers. Grouping them together in this way 360 * simplifies and speeds up task scanning. Minor typo: “even if array” -> “even if the array” I don’t claim to understand most of the updates to ForkJoinPool, there is more going on here from looking at the CVS logs (helpful to tease apart various changes), which is ok. Paul. > On Apr 5, 2018, at 9:43 AM, Martin Buchholz <marti...@google.com> wrote: > > http://cr.openjdk.java.net/~martin/webrevs/jdk/jsr166-integration/overview.html > > 8200728: Docs (Comparison of Stack and Deque methods) for Deque is not > correct' src/main/java/util/Deque.java > http://cr.openjdk.java.net/~martin/webrevs/jdk/jsr166-integration/Deque-Stack/index.html > https://bugs.openjdk.java.net/browse/JDK-8200728 > > 8200520: forkjoin tasks interrupted after shutdown > http://cr.openjdk.java.net/~martin/webrevs/jdk/jsr166-integration/ForkJoinPool-shutdown/index.html > https://bugs.openjdk.java.net/browse/JDK-8200520 > > 8200258: Improve CopyOnWriteArrayList subList code > http://cr.openjdk.java.net/~martin/webrevs/jdk/jsr166-integration/CopyOnWriteArrayList-subList/index.html > https://bugs.openjdk.java.net/browse/JDK-8200258 > > 8197531: Miscellaneous changes imported from jsr166 CVS 2018-04 > http://cr.openjdk.java.net/~martin/webrevs/jdk/jsr166-integration/miscellaneous/index.html > https://bugs.openjdk.java.net/browse/JDK-8197531 >