Closed simonerossoni1996 closed 2 years ago
Hi @simonerossoni1996
The PR looks fine to me.
To keep a uniform code style, please follow the conventions from: https://github.com/CRPropa/CRPropa3/blob/master/CONTRIBUTING.md
This essentially means: add white spaces around the division:
Vector3<T> u = axis / axis.getR();
One more thing: make sure that testVector.cpp
is still passing after this change.
Hi @rafaelab
thanks for your comments. I modified the style so that it is uniform now. I included two more lines in the code since there was a problem with the test rotation around the zero axis. Now all the tests are passed.
Thank you again.
Hi @simonerossoni1996,
the PR looks fine for me. I have only one small comment: Can you add a short sentence to the Changelog https://github.com/CRPropa/CRPropa3/blob/master/CHANGELOG.md. If it is done the PR can be merged
HI @JulienDoerner ,
thanks for your comment. I added a new line in the Changelog file.
Dear all,
I have fixed a bug present in the function getRotated() of Vector3.h. All the details and tests can be found in the presentation
ConeEmission_Targeting_issues.pdf
I tried to included this modification in the function randVectorAroundMean() of Random.cpp instead of Vector3.h. I ran a speed test to check the impact of these two different modifications on code performance. Here the result where Method 1 corresponds to the change in Vector3.h and Method 2 corresponds to the change in Random.cpp.
speed-test.pdf
For a large number of simulated particles the two method do not differ greatly. Therefore I suggest to modify the function getRotated() in Vector3.h in order to make this function mathematically correct and usable without further modifications.
Thank you.