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

List:       koffice-devel
Subject:    Re: Review Request: The patch for KCurve and KisPerChannelFilter
From:       "Dmitry Kazakov" <dimula73 () gmail ! com>
Date:       2009-02-20 19:17:50
Message-ID: 20090220191750.14729.32466 () localhost
[Download RAW message or body]



> On 2009-02-20 01:59:07, Johannes Simon wrote:
> > There is a couple of places where two numbers are divided, without checking the \
> > divisor for equality with 0. It might be that it is simply not necessary because \
> > the number will never be 0, but generally, i would discourage you to divide two \
> > numbers without a check. 
> > See for example lines 82, 94 and 95.

Maybe. In KCurve.cc i found _the only_ place where caller can provoke this division: \
in KCurve::io2sp(int x). Other are quite paranoic: widget size should be 1 =) In \
KCurve.h: Tridiagonal system always has an solution in this situation - this is a \
theorem. In splines divisor is m_h[smth] - that is the x-distance between two points \
(it's minimum value is controlled by POINT_AREA)


- Dmitry


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviewboard.kde.org/r/128/#review190
-----------------------------------------------------------


On 2009-02-19 13:30:40, Dmitry Kazakov wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviewboard.kde.org/r/128/
> -----------------------------------------------------------
> 
> (Updated 2009-02-19 13:30:40)
> 
> 
> Review request for KOffice.
> 
> 
> Summary
> -------
> 
> This patch goes together with the next one!
> http://reviewboard.kde.org/r/129/
> 
> What is done:
> KCurve:
> - implemented cubic splines in KCurve widget. Now all the calculations are made \
>                 inside KCubicSpline template.
> - faster calculations of the spline because of caching of the coefficients inside \
>                 the template
> - faster (*much faster*) redrawing of the curve because of caching of the _scaled_ \
>                 @m_pix histogram image in @m_pixmapCache object
> - fixed 'out_of_range' bug when you tried to delete the last point of the curve
> 
> KCurve+KisPerChannelWidget
> - added In/Out spinboxes like in photoshop to allow user to modify the curve with \
>                 keyboard
> - with spinboxes you can move the point to any place of the curve, so you can \
>                 change the order of the points
> - added handy function that creates new point in the middle of the curve and \
>                 focuses on spinboxes to set it
> - and a few small fixes
> 
> 
> The fist patch changes KCurve class. The second one makes KisPerChannelFilter use \
> new features. 
> 
> PS:
> What is planned to be done:
> 1) Look through preview mechanism, because it still slows everything down much.
> 2) Let user unselect the point of the curve by clicking on empty area
> 2,5) Hide In/Out controls when nothing selected
> 3) Far-far-future: color sampler mode for the curve (clicking on the canvas shows \
> the point on the curve) 
> 
> Diffs
> -----
> 
> trunk/koffice/krita/ui/widgets/kcurve.h 928664 
> trunk/koffice/krita/ui/widgets/kcurve.cc 928664 
> 
> Diff: http://reviewboard.kde.org/r/128/diff
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Dmitry
> 
> 

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


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