RFR(M): 8039407 : Regression test needed for JDK-8015396

[David Chase]

Thanks for comments, I'll clean it up.

I don't fully understand the hex values, and I was also not able to reproduce the bug, though the guy who submitted it claimed that this was his test case that worked for him.  It was not even 100% clear to me whether he was seeing it in compiler or interpreter, but I'll stick it in a loop, just in case, plus a short sleep (1 second) to give the compiler time to run before one last try.

On 2014-04-17, at 3:32 PM, Vladimir Kozlov <vladimir.kozlov at oracle.com> wrote:

> Hi David,
> 
> Can you use next style for the comment?
> 
> /*
> *
> */
> 
> And it would be good to see corresponding hex values in the comment which you use later in the code.
> 
> The comment needs more explanation what it is all about and why you are listing particular Windows machines configurations.
> 
> This test will test only Intepreter and not compiled jited code. You need at lest 3 iterations for compiled code to be used even with -Xcomp. I would suggest to create separate methods and call them from main() in a loop.
> 
> About bug itself. Was you able to reproduce the failure? It is not clear from report.
> 
> Thanks,
> Vladimir
> 
> On 4/17/14 11:30 AM, David Chase wrote:
>> Any chance of getting some reviews on this?
>> 
>> On 2014-04-11, at 7:56 AM, David Chase <david.r.chase at oracle.com> wrote:
>>> Changed to use Double.isNaN, retested with and without injected failure.
>>> 
>>> webrev: http://cr.openjdk.java.net/~drchase/8039407/webrev.01/
>> 
>> 
>> Original RFR:
>>> Bug 8015396 needs a regression test; I cannot reproduce it so there is no "fix" to check in,
>>> hence the separate bug to cover adding the test.  The test was included with the bug report,
>>> and has been adapted here to jtreg.
>>> 
>>> webrev: http://cr.openjdk.java.net/~drchase/8039407/webrev.00/
>>> bug: https://bugs.openjdk.java.net/browse/JDK-8039407
>>> 
>>> testing: I artificially stuffed a "NaN" into the result value and observed that the test would then
>>> report a failure.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20140417/2f88523f/signature.asc>