[Kstars-devel] Re: KDE/kdeedu/kstars/kstars/projections

[jsid at emor3j.fr.eu.org]

The only thing I have noticed about the horizon is that it looks a bit
strange since it's not drawn over the entire width of the map and I think
some stars below the horizon are drawn.
I agree with you for the QRect thing, so I will restore the line
*onVisibleHemisphere = dX*dX < M_PI*M_PI/4.;

Thanks,

-- 
Jérôme SONRIER


> Sorry for the delayed response. I feel sort of stupid for putting in the
comment that I did, since it doesn't really give helpful
> information. As I recall, there was a very good reason why that was like
that, but I can't remember what it was.
>
> Do you notice any change with the horizon (testing in outline and filled
mode) with and without that line?
> Also, constellations. Unfortunately I can't test stuff at the moment
since I don't have a working trunk and unfortunately don't have a lot of
time to build one at the moment.
>
> As for the QRect thing, I am biased against that method, because it will
create a new QRect object (which needs to malloc because it uses a
d-pointer) every single time we need to project something, and then
destruct it when we're done. So if we do 100 000 projections, we have
just malloc'd and free'd memory 100000 times, which is needless.
>
> Cheers,
> Henry de Valence
>
> On 13 February 2011 09:32, Jérôme SONRIER <jsid at emor3j.fr.eu.org>
wrote:
>> Ho thanks, I have changed this only for testing and forgot to correct
the code  before to commit it.
>> But now I have a question : should we use the actual code :
>> *onVisibleHemisphere = dX*dX < M_PI*M_PI/4.;
>> or something like
>> *onVisibleHemisphere = QRect( 0, 0, int(m_vp.width), int(m_vp.height)
).contains( p[0], p[1] ) ;
>> which show a larger part of  the sky ?
>> --
>> Jérôme SONRIER
>> Le samedi 12 février 2011, Henry de Valence a écrit :
>>> Is there a reason why the onVisibleHemisphere code was changed?