8229022: BufferedReader performance can be improved by using StringBuilder

[Daniel Fuchs]

Hi Brian,

Looks good to me.

On 02/10/2019 01:13, Brian Burkhalter wrote:
> While the performance improvement that I measured for this proposed change [1] using JMH was only in the 2% to 8% range as opposed to the 13% claimed in the issue description, given the simplicity of the change [2] it is probably worth doing. All use of the StringBuilder is within a synchronized block within a single package-scope method, so there should be no problem with access by multiple threads.

The StringBuffer is a local in the readLine method and
doesn't escape - so whether or not it's accessed within a synchronized
block should not matter. It's safe to replace it with StringBuilder
indeed!

best regards,

-- daniel

> 
> Thanks,
> 
> Brian
> 
> [1] https://bugs.openjdk.java.net/browse/JDK-8229022
> [2] diff
> 
> --- a/src/java.base/share/classes/java/io/BufferedReader.java
> +++ b/src/java.base/share/classes/java/io/BufferedReader.java
> @@ -314,7 +314,7 @@
>        * @throws     IOException  If an I/O error occurs
>        */
>       String readLine(boolean ignoreLF, boolean[] term) throws IOException {
> -        StringBuffer s = null;
> +        StringBuilder s = null;
>           int startChar;
>   
>           synchronized (lock) {
> @@ -372,7 +372,7 @@
>                   }
>   
>                   if (s == null)
> -                    s = new StringBuffer(defaultExpectedLineLength);
> +                    s = new StringBuilder(defaultExpectedLineLength);
>                   s.append(cb, startChar, i - startChar);
>               }
>           }
>