RFR (XS): 8023134: Rename VM LogFile to hotspot_pid{pid}.log (was hotspot.log)

[Vladimir Ivanov]

Coleen, I haven't integrated it yet.
So, I'm open for discussion.

Best regards,
Vladimir Ivanov

On 9/10/13 9:27 PM, Coleen Phillmore wrote:
>
> I have to admit, I was not looking forward to having multiple of these
> hotspot.log turds all over my file systems.   I like Bengt's suggestion,
> below.  Can we decline to make this change?
>
> Coleen
>
> On 9/10/2013 8:31 AM, Bengt Rutisson wrote:
>> On 9/10/13 1:31 PM, Staffan Larsen wrote:
>>> I'm not sure this change is only to the better. Since every run of a
>>> debug-build of hotspot creates this file, there is now going to be an
>>> enormous amount of these files create. At least for me that is going
>>> to be a big nuisance since they will clutter up every directory where
>>> I run hotspot. Would it be possible to keep hotspot.log as the
>>> default name, and have a flag which multiprocess tests can use to
>>> include the pid in the name?
>>
>> I agree with Staffan. This seems to be a quite invasive change in the
>> sense that it will create lots of files in our test and development
>> environments.
>>
>> In fact, isn't it possible for the multiprocess tests that have issues
>> to just specify the LogFile as:
>>
>> -XX:LogFile=hotspot_pid%p.log
>>
>> That wouldn't require any changes to the VM at all.
>>
>> Thanks,
>> Bengt
>>
>>>
>>> /Staffan
>>>
>>> On 9 sep 2013, at 14:24, Vladimir Ivanov
>>> <vladimir.x.ivanov at oracle.com> wrote:
>>>
>>>> http://cr.openjdk.java.net/~vlivanov/8023134/webrev.00/
>>>> 7 lines changed: 4 ins; 0 del; 3 mod
>>>>
>>>> Added pid of the Java process to default name of VM log file:
>>>>     - old: hotspot.log
>>>>     - new: hotspot_pid{pid}.log
>>>>
>>>> It allows to avoid hotspot.log overwrite/corruption in multi-VM test
>>>> scenarios and simplifies hs_err/ciReplay/hotspot.log files matching.
>>>>
>>>> Thanks!
>>>>
>>>> Best regards,
>>>> Vladimir Ivanov
>>
>