> On July 2, 2015, 12:28 a.m., Ian Downes wrote:
> > src/linux/perf.cpp, line 473
> > <https://reviews.apache.org/r/36114/diff/1/?file=997704#file997704line473>
> >
> >     Is there some way to have different extract functions without coding a 
> > single, specific kernel version in the name, e.g., even extractv1, 
> > extractv2, ... seems preferable

used extractv1.

I think I can take advantage of the virtual functions in OO to do that, but not 
sure it's worth the complication.


> On July 2, 2015, 12:28 a.m., Ian Downes wrote:
> > src/linux/perf.cpp, lines 531-532
> > <https://reviews.apache.org/r/36114/diff/1/?file=997704#file997704line531>
> >
> >     I don't understand these conditions, e.g., surely 
> >     ```
> >     (version > Version(3,12,0)) == (version >= Version(3,0,0) && version > 
> > Version(3, 12, 0)
> >     ```
> >     for all possible values of version, unless I'm not understanding how 
> > versions are compared...?
> >     
> >     For this type of selection I think it's clearest/simplest to start from 
> > the lowest in the ordering:
> >     ```cpp
> >     if (a < x) {
> >       //
> >     } else if (a < y) {
> >       //
> >     } else if (a < z) {
> >       //
> >     } else {
> >       //
> >     }
> >     ```

Don't quite follow your comment.

normally, kernel should go from 2.6.39 to 3.0 directly, but some kernel does 
2.6.40 for 3.0, i.e., 'uname -r' can't be trusted.

We care about two bands:
1. 3.0 - 3.12 (vanilla), or its equivalant 2.6.40 to 2.6.52 (customized) -> v1; 
2. >3.12 or its equivalent > 2.5.52 -> v2.


- Chi


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/36114/#review90170
-----------------------------------------------------------


On July 1, 2015, 10:44 p.m., Chi Zhang wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/36114/
> -----------------------------------------------------------
> 
> (Updated July 1, 2015, 10:44 p.m.)
> 
> 
> Review request for mesos, Ian Downes, Paul Brett, and Cong Wang.
> 
> 
> Bugs: mesos-2834
>     https://issues.apache.org/jira/browse/mesos-2834
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> perf: added another extract function to support the new perf format after 
> v3.12.
> 
> 
> Diffs
> -----
> 
>   src/linux/perf.cpp 697b75e846a43d4f106ad8f39a18882836d7dc02 
> 
> Diff: https://reviews.apache.org/r/36114/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Chi Zhang
> 
>

Reply via email to