RFR (XS) 8221307 : String.substring() OOB exception on start index reports improper information

Ivan Gerasimov ivan.gerasimov at oracle.com
Tue Aug 13 19:55:07 UTC 2019


Hello!

A simple benchmark did not detect any negative impact of delegating 
substring(int) to substring(int, int):

Before fix:
Benchmark                     Mode  Cnt   Score   Error Units
StringSubstring.from26toEnd0  avgt   12  20.279 ± 0.306 ns/op
StringSubstring.from26toEnd1  avgt   12  20.677 ± 1.924 ns/op

After fix:
Benchmark                     Mode  Cnt   Score   Error Units
StringSubstring.from26toEnd0  avgt   12  19.835 ± 0.135 ns/op
StringSubstring.from26toEnd1  avgt   12  20.784 ± 1.791 ns/op


So here's the updated webrev: 
http://cr.openjdk.java.net/~igerasim/8221307/01/webrev/

Will you approve it to go?

With kind regards,

Ivan


On 8/8/19 6:56 AM, Claes Redestad wrote:
> Hi,
>
> perhaps a stupid question, but why isn't String::substring(int) 
> calling substring(beginIndex, length())? That'd ensure consistent error
> messages, but otherwise preserve semantics.
>
> /Claes
>
> On 2019-08-08 02:17, Ivan Gerasimov wrote:
>> Hello!
>>
>> The exception thrown by substring(int) may look confusing.
>> For example it produces "String index out of range: -1" when the 
>> index is length+1.
>>
>> It is proposed to make the error message more clear, similar to what 
>> we have for substring(int, int).
>>
>> BUGURL: https://bugs.openjdk.java.net/browse/JDK-8221307
>> WEBREV: http://cr.openjdk.java.net/~igerasim/8221307/00/webrev/
>>
>> Would you please help review?
>>
>
-- 
With kind regards,
Ivan Gerasimov



More information about the core-libs-dev mailing list