RFR: 8325730: StringBuilder.toString allocation for the empty String

[Claes Redestad]

On Tue, 20 Feb 2024 17:00:14 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

>> JDK-8282429 accidentally removed an optimization (JDK-8240094) that ensured StringBuilder/StringBuffer::toString returns `""` when the builders are empty.
>> 
>> 
>> Name                         Cnt     Base      Error   Test   Error   Unit  Change
>> StringBuffers.emptyToString    5   12,289 ±    0,384  9,883 ± 0,721  ns/op   1,24x (p = 0,000*)
>>   :gc.alloc.rate                 1862,398 ±   57,647  0,007 ± 0,000 MB/sec   0,00x (p = 0,000*)
>>   :gc.alloc.rate.norm              24,000 ±    0,000  0,000 ± 0,000   B/op   0,00x (p = 0,000*)
>>   :gc.count                        31,000             0,000         counts
>>   :gc.time                         21,000                               ms
>> StringBuilders.emptyToString   5    4,146 ±    0,567  0,646 ± 0,003  ns/op   6,42x (p = 0,000*)
>>   :gc.alloc.rate                 9208,656 ± 1234,399  0,007 ± 0,000 MB/sec   0,00x (p = 0,000*)
>>   :gc.alloc.rate.norm              40,000 ±    0,000  0,000 ± 0,000   B/op   0,00x (p = 0,000*)
>>   :gc.count                        96,000             0,000         counts
>>   :gc.time                         64,000                               ms
>>   * = significant
>
> test/micro/org/openjdk/bench/java/lang/StringBuffers.java line 49:
> 
>> 47: 
>> 48:     @Benchmark
>> 49:     public String emptyToString() {
> 
> Do we actually need to remove the `substring` test here?

I couldn't figure out why we'd want to have `String::substring` micros in a test `StringBuffers` (there's also a `StringSubstring` micro), though I could deal with that as an explicit cleanup RFE.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17931#discussion_r1496266477