[Jim Laskey]

Wonder if you should not move the test into StringLatin1.newString and StringUTF16.newString to return "" if subLen [len] == 0. Then there would be an across the board benefit.


> On Feb 25, 2020, at 11:43 AM, Сергей Цыпанов <sergei.tsypanov at yandex.ru> wrote:
> 
> Hello,
> 
> current implementation of String.substring(int,int) has a fast-path for the case when beginIndex = 0 and endIndex = length.
> 
> I think there should be similar fast-path for the case beginIndex = endIndex (asuming both are valid):
> 
> diff --git a/src/java.base/share/classes/java/lang/String.java b/src/java.base/share/classes/java/lang/String.java
> --- a/src/java.base/share/classes/java/lang/String.java
> +++ b/src/java.base/share/classes/java/lang/String.java
> @@ -1900,10 +1900,13 @@
>     public String substring(int beginIndex, int endIndex) {
>         int length = length();
>         checkBoundsBeginEnd(beginIndex, endIndex, length);
> -        int subLen = endIndex - beginIndex;
>         if (beginIndex == 0 && endIndex == length) {
>             return this;
>         }
> +        int subLen = endIndex - beginIndex;
> +        if (subLen == 0) {
> +            return "";
> +        }
>         return isLatin1() ? StringLatin1.newString(value, beginIndex, subLen)
>                           : StringUTF16.newString(value, beginIndex, subLen);
>     }
> 
> With this call to str.substring(n,n) will be allocation-free.
> 
> Regards,
> Sergey Tsypanov