RFR: 8291641: Optimize StackTraceElement.toString() [v2]

[David Schlosnagle]

On Tue, 2 Aug 2022 05:29:38 GMT, Quan Anh Mai <duke at openjdk.org> wrote:

>> David Schlosnagle has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Address comments
>
> Since constructing a `String` from a `StringBuilder` requires an array copy anyway, you don't need to be exact in the capacity of the builder and overshooting is fine. This can simplify the length calculation a lot. Thanks.

Good call @merykitty , I updated with my take on simplifying the length estimate, let me know if that reads better.

JMH as of https://github.com/openjdk/jdk/pull/9665/commits/d3774a69b1e6f316ff8b4af257fb2e95e5c4c3dd for comparison to previous, so still a ~50% improvement over existing state on my x64 test environment (`Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz`).

Benchmark                               Mode  Cnt       Score     Error  Units
StackTraceElementBench.printStackTrace  avgt   15  144011.319 ± 959.263  ns/op
StackTraceElementBench.toString         avgt   15      67.874 ±   0.485  ns/op

-------------

PR: https://git.openjdk.org/jdk/pull/9665