RFR(XS): 8014068: TEST_BUG: compiler/ciReplay/TestSA.sh fails on Windows: core wasn't generated

[Igor Ignatyev]

> Dump replay.diff on output instead of "SEE replay.diff". When you run
> several tests scratch directory will be overwritten and replay.diff gone.
>
> Did you verified that the test work on Windows with hs_err_pid<pid>.mdmp
> cores?
Yes, I did. I reran compiler/ciReplay on the hosts that are used during 
nightly.

http://cr.openjdk.java.net/~iignatyev/8014068/webrev.02/

Best regards,
Igor Ignatyev

On 05/13/2013 11:03 PM, Vladimir Kozlov wrote:
> Dump replay.diff on output instead of "SEE replay.diff". When you run
> several tests scratch directory will be overwritten and replay.diff gone.
>
> Did you verified that the test work on Windows with hs_err_pid<pid>.mdmp
> cores?
>
> Vladimir
>
> On 5/13/13 11:23 AM, Igor Ignatyev wrote:
>> corrected.
>>
>> http://cr.openjdk.java.net/~iignatyev/8014068/webrev.01
>>
>> Best regards,
>> Igor Ignatyev
>>
>> On 05/13/2013 09:07 PM, Vladimir Kozlov wrote:
>>> Next output redirection:
>>>
>>>   2>&1 > replay.diff
>>>
>>> should be:
>>>
>>>   > replay.diff 2>&1
>>>
>>> otherwise error will be not in file.
>>>
>>> Vladimir
>>>
>>> On 5/13/13 9:28 AM, Igor Ignatyev wrote:
>>>> Hi all,
>>>>
>>>> Please review patch.
>>>>
>>>> Problem:
>>>> incorrect processing of core location on windows hosts due to using
>>>> invalid variable.
>>>>
>>>> Fix:
>>>> using of $OS was replaced by using of $VM_OS variable
>>>> fixed processing core locations for *nix: added core.<pid>
>>>> also slight cosmetic changes:
>>>>    - added check of OS before 'ulimit' change
>>>>    - not supported by all hosts '--brief' flags was removed
>>>>
>>>> webrev: http://cr.openjdk.java.net/~iignatyev/8014068/webrev.00/
>>>> jbs: https://jbs.oracle.com/bugs/browse/JDK-8014068
>>>>