RFR: 8168722: Unified Logging configuration output needs simplifying

Marcus Larsson marcus.larsson at oracle.com
Wed Feb 28 09:49:53 UTC 2018


Thanks for reviewing Robbin!

Marcus


On 2018-02-27 16:10, Robbin Ehn wrote:
> Looks good thanks for fixing this!
>
> It's awesome to be able to compress your log expression!
>
> /Robbin
>
> On 02/20/2018 09:51 AM, Marcus Larsson wrote:
>> Hi,
>>
>> Please review the following patch to make UL configuration 
>> descriptions more readable/usable. Instead of explicitly listing all 
>> tag sets with their corresponding levels (a=info,b=info, ..., 
>> x=debug, etc), a significantly shorter aggregate is now generated and 
>> printed instead (all=info,x=debug).
>>
>> Summary:
>> Initially, the configured level for the majority of the log tag sets 
>> is found and used for the "all=<level>" part of the config string. 
>> Then, using a greedy iterative algorithm, selections are generated 
>> and added to the configuration string to cover the descriptions for 
>> tag sets with levels deviating from the current description. In each 
>> step, the selection which covers the most deviating tag sets is 
>> chosen. This is repeated until there are no deviating tag sets 
>> remaining. For the curious, the problem can be thought of as a 
>> variation of the set cover problem [0].
>>
>> Issue:
>> https://bugs.openjdk.java.net/browse/JDK-8168722
>>
>> Webrev:
>> http://cr.openjdk.java.net/~mlarsson/8168722/webrev.00
>>
>> Tested with hs-tier 1-2, and locally with 
>> jtreg/{serviceability,runtime,gc}/logging.
>>
>> Thanks,
>> Marcus
>>
>> [0]: https://en.wikipedia.org/wiki/Set_cover_problem



More information about the hotspot-runtime-dev mailing list