Review Request 119621: Support for Geogebra's intersections, constrained points and Loci in kig ( GSoC 14 )

[David Narváez]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/119621/#review64382
-----------------------------------------------------------


Here's the thing: From what I see these objects need some special massaging of the stack. I am thinking we can make that massaging an ObjectType itself and output, e.g., the locus object, in two steps in the filter:

<LocusHelperSomething label="foo">
  <Parent label="bar"/>
  <Parent label="baz"/>
 </LocusHelperSomething>
 <Locus label="foo1">
  <Parent label="foo"/>
 </Locus>
 
Is this reasonable? I am not sure how the code will look at the end, but I think it's worth a try. It may be that using these object types spurs a simplification of the Locus code itself.

- David Narváez


On Aug. 12, 2014, 1:57 a.m., Aniket Anvit wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://git.reviewboard.kde.org/r/119621/
> -----------------------------------------------------------
> 
> (Updated Aug. 12, 2014, 1:57 a.m.)
> 
> 
> Review request for KDE Edu and David Narváez.
> 
> 
> Repository: kig
> 
> 
> Description
> -------
> 
> Line-Line intersection is perfect and simple to implement with the existing model. Other Intersections are a bit challenging since they output more than one object (point) unlike the other commands ( which output just one object).
> LineConic, Circle-Circle intersections seem fine to me but I have some trouble in understanding whether my approach is correctly assigning the output-label to the obtained intersection points ( while filling the m_objectMap). Please check whether the IntImp parameter I am passing to the IntersectionTypes is correct or not. ConicConic Intersections produce 4-points sometimes and they have been implemented as Line-Conic intersections in Kig. So, we have to find the Radical-Lines of the conics first before finding the intersection-points. Currently , there is an issue that the Radical-lines are also added to the document ( while they should not be ), but this should be simple to take care (If you give the go-ahead to this approach). 
> 
> Constrained points are must for Locus. I have implemented them , but the extra-handling required is making things a little ugly.
> Locus is under progress...
> 
> 
> Diffs
> -----
> 
>   geogebra/geogebra.xsl 87ef07b 
>   geogebra/geogebratransformer.h 5f36827 
>   geogebra/geogebratransformer.cpp aee8669 
> 
> Diff: https://git.reviewboard.kde.org/r/119621/diff/
> 
> 
> Testing
> -------
> 
> I ran some basic tests for all the intersection cases which have been implemented. Seem ok. Only issue is with the conic-conic intersection case where we get the additional ConicRadical Lines which should be visible in Kig. However, this should be easy to take care of if this implementation survives.
> 
> 
> File Attachments
> ----------------
> 
> locus_working_1
>   https://git.reviewboard.kde.org/media/uploaded/files/2014/08/10/9559b469-3c24-41d8-b9ba-809e68f86fbe__locus_working_1.ggb
> locus_working_2
>   https://git.reviewboard.kde.org/media/uploaded/files/2014/08/10/fce5bd47-794b-4ae3-aa21-4cd7c688c6fe__locus_working_2.ggb
> locus_working_3
>   https://git.reviewboard.kde.org/media/uploaded/files/2014/08/10/8cc09d1b-ff42-4a99-b2ad-7d7b403c92ad__locus_working_3.ggb
> locus_crashing
>   https://git.reviewboard.kde.org/media/uploaded/files/2014/08/10/c5c142eb-9542-4c2b-b6f8-0250ccdfb3fd__locus_crashing_parabola.ggb
> 
> 
> Thanks,
> 
> Aniket Anvit
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-edu/attachments/20140812/88fc832e/attachment.html>