'Re: Review Request: Plasma + Nepomuk - libplasma patch - attempt 1'

[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-panel-devel
Subject:    Re: Review Request: Plasma + Nepomuk - libplasma patch - attempt 1
From:       "Aaron Seigo" <aseigo () kde ! org>
Date:       2009-09-24 0:18:36
Message-ID: 20090924001836.18187.59109 () localhost
[Download RAW message or body]

> On 2009-09-23 21:28:50, Aaron Seigo wrote:
> > /trunk/KDE/kdelibs/plasma/context.h, line 78
> > <http://reviewboard.kde.org/r/1699/diff/1/?file=11891#file11891line78>
> > 
> > or perhaps:
> > 
> > forActivity(id)->setActiveContext(true);?
> > 
> 
> wrote:
> This looks appealing. The problem is that the setActiveContext would have to \
> trigger the context/activity changed signal in the manager.

that can be done behind the scenes quite easily. i'm thinking more about the public \
API here rather than the implementation; we often make some parts of the internal \
implementation a little "uglier" so that the public API is more consistent and \
pretty.

> On 2009-09-23 21:28:50, Aaron Seigo wrote:
> > /trunk/KDE/kdelibs/plasma/context.h, line 67
> > <http://reviewboard.kde.org/r/1699/diff/1/?file=11891#file11891line67>
> > 
> > ensureActivityExists?
> > 
> > or maybe ensureActivity and createActivity could be merged? maybe something like:
> > 
> > activityId(const QString &name, const QString &expectedId = QString());?
> > 
> > or perhaps these can remain as they are but be private API? it should only really \
> > be needed by Context when setting the current activity in that Context?
> 
> wrote:
> It would be ok to rename to ensureActivityExists
> 
> activityId is a bit to vague - it doesn't say what it does at all (it looks like a \
> getter for a field while it can create a new activity) 
> The idea was for an application/applet/whatever to be able to create a new activity \
> - then, plasma (and other apps) would react to that, by creating a new \
> plasma-activity.

well, the method fetches the id for a given activity name. in that sense, it seems \
like a getter. that it may need to create the activity in the nepomuk store is almost \
a detail?

as for creating a new activity, it really seems more natural, at least to me, to \
create Context objects and manage those. the only complication is that activities \
have IDs and they have names. *thinks*

you know, it seems to me that applications should ALWAYS be refering to activities by \
ID. i can't think of a single use case where application code should rely on the \
activity's name. that's something for the user, isn't it?

in plasma-desktop, we need to keep a mapping of Containment ID to Activity ID (or, \
Context::id()), but that's all.

> On 2009-09-23 21:28:50, Aaron Seigo wrote:
> > /trunk/KDE/kdelibs/plasma/context.h, line 83
> > <http://reviewboard.kde.org/r/1699/diff/1/?file=11891#file11891line83>
> > 
> > is this equivalent to currentContext()->name()?
> 
> wrote:
> hm... Context doesn't have a name. Containment has, and it can be different - if \
> the activity name is empty, the Containment name is some generic string \
> (Desktop/Panel...)

the name of a Context should be the name associated with the Activity, if any, \
correct? i don't think we should think too much bout Containment is doing (other than \
we want to map Containment <-> Context)

> On 2009-09-23 21:28:50, Aaron Seigo wrote:
> > /trunk/KDE/kdelibs/plasma/context.h, line 105
> > <http://reviewboard.kde.org/r/1699/diff/1/?file=11891#file11891line105>
> > 
> > is this equivalent to:
> > 
> > forActivity(id)->name(); ?
> 
> wrote:
> Again, the context doesn't have a name
> 
> wrote:
> what exactly is the difference between "the context" and "the activity"?

context is a combination of the activity and the physical location of the \
machine/user. it may eventually contain other items (presence?)

> On 2009-09-23 21:28:50, Aaron Seigo wrote:
> > /trunk/KDE/kdelibs/plasma/context.h, line 151
> > <http://reviewboard.kde.org/r/1699/diff/1/?file=11891#file11891line151>
> > 
> > setCurrentActivityId()? and then there could be a corresponding activityId().
> 
> wrote:
> the setCurrentActivity is an old function - so we can't really change the name. We \
> could deprecate it as well, and introduce the one with the ID but (IMO) that would \
> be too much... 
> I've mostly made decisions with a thought that I should change the api as little as \
> possible.

fair enough. :)

- Aaron

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviewboard.kde.org/r/1699/#review2433
-----------------------------------------------------------

On 2009-09-23 20:19:36, Ivan Cukic wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviewboard.kde.org/r/1699/
> -----------------------------------------------------------
> 
> (Updated 2009-09-23 20:19:36)
> 
> 
> Review request for Plasma.
> 
> 
> Summary
> -------
> 
> Nepomuk-based storage of plasma activities
> 
> 
> Diffs
> -----
> 
> /trunk/KDE/kdelibs/plasma/CMakeLists.txt 1027137 
> /trunk/KDE/kdelibs/plasma/applet.cpp 1027137 
> /trunk/KDE/kdelibs/plasma/containment.h 1027137 
> /trunk/KDE/kdelibs/plasma/containment.cpp 1027137 
> /trunk/KDE/kdelibs/plasma/context.h 1027137 
> /trunk/KDE/kdelibs/plasma/context.cpp 1027137 
> /trunk/KDE/kdelibs/plasma/private/activitieshandler.h PRE-CREATION 
> /trunk/KDE/kdelibs/plasma/private/activitieshandler.cpp PRE-CREATION 
> /trunk/KDE/kdelibs/plasma/private/context_p.h PRE-CREATION 
> /trunk/KDE/kdelibs/plasma/private/desktoptoolbox.cpp 1027137 
> /trunk/KDE/kdelibs/plasma/private/fallbackactivitieshandler.h PRE-CREATION 
> /trunk/KDE/kdelibs/plasma/private/fallbackactivitieshandler.cpp PRE-CREATION 
> /trunk/KDE/kdelibs/plasma/private/nepomukactivitieshandler.h PRE-CREATION 
> /trunk/KDE/kdelibs/plasma/private/nepomukactivitieshandler.cpp PRE-CREATION 
> 
> Diff: http://reviewboard.kde.org/r/1699/diff
> 
> 
> Testing
> -------
> 
> Needs more testing
> 
> 
> Thanks,
> 
> Ivan
> 
> 

_______________________________________________
Plasma-devel mailing list
Plasma-devel@kde.org
https://mail.kde.org/mailman/listinfo/plasma-devel

[prev in list] [next in list] [prev in thread] [next in thread] 