Review Request 121831: libksysguard: process.h: encapsulate private fields

[Gregor Mi]

> On Jan. 21, 2015, 10:10 a.m., Dominik Haumann wrote:
> > processcore/process.h, line 40
> > <https://git.reviewboard.kde.org/r/121831/diff/3/?file=343193#file343193line40>
> >
> >     Is virtual needed here?

Does it hurt here to have it virtual? I thought, if in doubt 'virtual' should be used.


> On Jan. 21, 2015, 10:10 a.m., Dominik Haumann wrote:
> > processcore/processes_linux_p.cpp, line 171
> > <https://git.reviewboard.kde.org/r/121831/diff/3/?file=343197#file343197line171>
> >
> >     Don't you change behavior here?
> >     
> >     Before, we just wrote into the varialbes.
> >     
> >     Now, we use the setters, which also sets 'changes |= Process::Gids;'
> >     
> >     Is that maybe an issue? I myself don't know the code well enough to see this here.

True. Thanks for noticing. The changes will be read in ProcessModelPrivate::processChanged when a change is signaled by KSysGuard::Process::processChanged. Then GUI model updates are triggered. So probably the worst that can happen is one additional GUI update when processChanged is emitted.


- Gregor


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/121831/#review74463
-----------------------------------------------------------


On Jan. 25, 2015, 12:01 p.m., Gregor Mi wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://git.reviewboard.kde.org/r/121831/
> -----------------------------------------------------------
> 
> (Updated Jan. 25, 2015, 12:01 p.m.)
> 
> 
> Review request for KDE Base Apps, Dominik Haumann, Eike Hein, and John Tapsell.
> 
> 
> Repository: libksysguard
> 
> 
> Description
> -------
> 
> In process.h there are several public fields. This RR introduces a d-ptr.
> 
> 
> (In a separate commit I would add the .reviewboardrc file)
> 
> What is the current policy on using small C++ macros as done in this RR? Use it (code is more compact and readable) or don't use it (better for debugging)?
> 
> 
> Diffs
> -----
> 
>   .reviewboardrc PRE-CREATION 
>   processcore/processes_base_p.h 71b8a9cc6ee14bf7934a0a9d3199b257b5ce1be7 
>   processcore/processes_linux_p.cpp 898d4fa491873fe95a8b32a5c1b85642b2e46ad5 
>   processcore/processes_solaris_p.cpp f054df4b1e762e9cbec1ff8dea78f467b878bee0 
>   processui/ProcessFilter.cpp ec520593fb67c777d56817f2493d40dc5ade0347 
>   processui/ProcessModel.cpp 53bc041110c9cdb686fef783895104969b661889 
>   processui/ksysguardprocesslist.cpp 4dc142e864d8353ceafc3a6735ffa81e48291420 
>   processui/scripting.h 2445c0ab0d81af3283c0f6e9c5f349a3d70b0de9 
>   processui/scripting.cpp 76291b0ae0a26e486aa81a4ca3976ff4a47cb3c0 
>   tests/processtest.cpp f9b36e9a3a3c2048b51f1d935f8c40de2ad8a9b8 
>   CMakeLists.txt cefc86f12be684e195bd148641483e9e1734e636 
>   processcore/process.h b6695c0ed301dc5f0fad8ba847da811f19ebfd9a 
>   processcore/process.cpp a38b8be71da1a51cb87f636664ebac817b1d20ab 
>   processcore/processes.h d09c3265333fe7e2702deaa910c5fbe4bc3ac9e6 
>   processcore/processes.cpp 580df8db152040f1ad075430fdce08fe7ad4ae2d 
>   processcore/processes_atop_p.cpp 24c76e3e35f62bd8e9e705ad32cc11cbd3662601 
> 
> Diff: https://git.reviewboard.kde.org/r/121831/diff/
> 
> 
> Testing
> -------
> 
> Compiles and runs. Data is still shown; no visible error. Unit tests succeed.
> 
> 
> Thanks,
> 
> Gregor Mi
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20150125/c6cdc902/attachment.htm>