RFR: 4511638: Double.toString(double) sometimes produces incorrect results
bpb at openjdk.java.net
Fri Apr 9 16:42:16 UTC 2021
On Fri, 9 Apr 2021 16:33:43 GMT, Raffaello Giulietti <github.com+70726043+rgiulietti at openjdk.org> wrote:
>> Forgot to add that other changes in the code are the use of switch expressions and the use of instanceof patterns.
> here's some background information for those that didn't follow the mailing list for the last couple of years.
> Some enjoyable properties of the novel algorithm:
> * No intermediate objects are instantiated.
> * Loop-free core algorithm.
> * Only int and long arithmetic.
> * No divisions at all.
> * 17.7x speedup (jmh) (https://mail.openjdk.java.net/pipermail/core-libs-dev/2020-April/065921.html).
> * Randomized, yet reproducible deep diving tests (jtreg).
> * Clear, unambiguous spec.
> * All floats have been tested to fully meet the spec.
> * Fully documented in (https://drive.google.com/file/d/1luHhyQF9zKlM8yJ1nebU0OgVYhfC6CBN/view) and/or in comments.
> See (https://mail.openjdk.java.net/pipermail/core-libs-dev/2019-September/062580.html) for some (invented) Q&A. The last Q&A deals with your investment in time for an informed review.
@rgiulietti Please issue the `/csr/` command here . Speaking of which, does the CSR  need to be updated?
More information about the core-libs-dev