JDK-8239787 : AArch64: String.indexOf may incorrectly handle empty strings

[Alexey Bakhtin]

Hello Leonid,

Thank you for review.
This issue was found in aarch64 version of string intrinsic. Intrinsics for other platforms are not affected. However, this test applicable for all platforms, so I've excluded platform restriction for this test.
The rest findings are also fixed.

Webrev: http://cr.openjdk.java.net/~bae/8239787/webrev.02/ <http://cr.openjdk.java.net/~bae/8239787/webrev.02/>

Thank you
Alexey

> On 29 Feb 2020, at 00:12, Leonid Mesnik <leonid.mesnik at oracle.com> wrote:
> 
> Hi
> 
> I have few comments about added test:
> 
> 1. Are there any reasons to limit to aarch64 only? If it is valid for other platforms then let run it on these platforms also.
> 
> 2. Please fix whit spaces in lines 33,34;
> 
> 3. The meaningful directory names are preferred rather than bugid. Not a strict requirement but usually it increase readability.
> 
> Leonid
> 
> On 2/28/20 3:47 AM, Alexey Bakhtin wrote:
>> Hello Yuri,
>> 
>> Here is webrev with fix and new jtreg test :
>> http://cr.openjdk.java.net/~bae/8239787/webrev.01/
>> 
>> Thank you
>> Alexey
>> 
>>> On 28 Feb 2020, at 11:45, Yuri Nesterenko <yan at azul.com> wrote:
>>> 
>>> /Hi Alexey, could you please provide a regression test with this fix?
>>> Thanks, --yan