Closed techsy730 closed 3 years ago
Actually, I need to revert that second commit. In both the in the JDK and FastUtils, no spliterator implementation backed by an iterator tries to use the iterator in multiple threads. So it is actually safe to use to use them as stream sources.
Ooooook... looks like that part was only partially written. I know, the conditional code is a bit crazy in the hash-based classes. Let me know when this is read to merge as I want to release 8.5.2.
Ok, I think we are good now.
But do we have any tests for AbstractMap
? As in a map type (possible test only) that doesn't implement more then the bare minimum of methods require to extend AbstractMap
?
EDIT: Actually, similar question for the other Abstract helpers
Have a look at AbstractInt2IntMapTest.
In general, look in the IDE for AbstractTest.
keySet
orvalues
in a LinkedOpenHashMap would follow the proper order (instead falling back to regular OpenHashMap order).forEachRemaining
methods. This saves a lot of implicit casts by the generics supporting bridge methods the compiler inserts.spliterator()
methods