> On June 11, 2013, 9:59 p.m., Simeon Bird wrote:
> > Should something similar be done for Nepomuk2::ResourceData::removeProperty 
> > as well?
> 
> Simeon Bird wrote:
>     ping?
> 
> Vishesh Handa wrote:
>     Thanks for reminding me.
>     
>     I was just taking a look at this patch again, and are you sure it's 
> thread safe? I notice that the resource manager is being locked before 
> accessing any of its members, but in this case we're just accessing the model 
> which has a mutex inside it. So it should be safe, right?
>     
>     I don't want to commit something right before 4.12 release that causes a 
> lot of crashes.
>     
>     Also added the same code to removeProperty

I *think* it is thread-safe, because the only part of the rm we are accessing 
is the model, which is only set once on initialization, which is guaranteed to 
happen before the storage service is advertised.

But now I have another worry: what happens if d->overrideModel is set? Can 
m_model be null in that case?
(is it ever set? I forget what it is for right now...)

Also I totally agree about being safe just before the release - I think its 
thread safe, but I wouldn't bet my house on it. 
How about delaying the commit until just after the 4.11 release, so it has a 
month of testing before finding its way into 4.11.1?


- Simeon


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/110938/#review34193
-----------------------------------------------------------


On June 10, 2013, 3:44 p.m., Vishesh Handa wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/110938/
> -----------------------------------------------------------
> 
> (Updated June 10, 2013, 3:44 p.m.)
> 
> 
> Review request for Nepomuk and Simeon Bird.
> 
> 
> Description
> -------
> 
>    ResourceData: Do not send the dbus message unless the mainmodel is valid
>     
>     This way we avoid all those thousands of debug messages of
>     org.kde.nepomuk.DataManagement not existing because Nepomuk is not
>     running.
>     
>     BUG: 269442
> 
> 
> Diffs
> -----
> 
>   libnepomukcore/resource/resourcedata.cpp ba3f6f4 
> 
> Diff: http://git.reviewboard.kde.org/r/110938/diff/
> 
> 
> Testing
> -------
> 
> not really
> 
> 
> Thanks,
> 
> Vishesh Handa
> 
>

_______________________________________________
Nepomuk mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/nepomuk

Reply via email to