JDK-8273642: The UUID class does not implement RFC4122 fully (UUID type 5 is missing).

Roger Riggs roger.riggs at oracle.com
Thu Nov 11 22:45:27 UTC 2021

Hi Mario,

I linked and closed https://bugs.openjdk.java.net/browse/JDK-8273642 as 
a duplicate.

The hex string->binary conversions in UUID are already highly optimized.

I don't think the any of the recently added support for Hexadecimal in 
java.util.HexFormat would/could make it any faster.

There might be a trivial space gain by refactoring to have a single 
table used for the digit conversions in both.
(UUID.nibbles vs HexFormat.DIGITS) but it would be a very low value change.

Thanks, Roger

On 11/11/21 1:34 PM, Mario Pavlov wrote:
> Thank you, Alan. I'll try to see if there is opportunity for further
> optimization, although I'm currently looking at the implementation in JDK
> 16 and I don't see obvious things that can be optimized :)
> Also is https://bugs.openjdk.java.net/browse/JDK-8246516 duplicate of
> https://bugs.openjdk.java.net/browse/JDK-8273642 ?
> On Thu, Nov 11, 2021 at 8:03 PM Alan Bateman <Alan.Bateman at oracle.com>
> wrote:
>> On 11/11/2021 17:17, Mario Pavlov wrote:
>>> Hi Aleksey,
>>> Thanks for your answer.
>>> In that case maybe I should start with something even smaller, e.g.
>>> https://bugs.openjdk.java.net/browse/JDK-8192784
>>> What do you think?
>> UUID.fromString(String) was re-implemented in JDK 15 with a more optimal
>> implementation so I suspect JDK-8192784 can be closed. One thing you
>> could do is do some benchmarks to see if there is any further
>> opportunity and bring it to core-libs-dev for discussion. Note that
>> identifying enhancements in JBS that can be closed is a contribution in
>> itself.
>> -Alan

More information about the jdk-dev mailing list