[Kstars-devel] Review Request: Fix dissapearing Milky Way at high zoomlevel on QPainter

[Rishab Arora]

You're right, there are issues when the points leave the FOV in the
Gnomonic and Equirectangular projections. I was using the patched version
of Kstars committing it and discovered I hadn't tested it properly
(embarrassing moment). I reported this to Rafal and this won't be shipped.
I have made this new patch (attached: p4.patch ) which works for QPainter
but due to issues in the two mentioned projections it falls back to manual
clipping for them. What's your view on this?


--
Rishab Arora
@spacetime29 <https://twitter.com/#%21/spacetime29>
Find me on Google+ <https://plus.google.com/118197889888669218025/posts>



On Sat, Mar 24, 2012 at 3:03 AM, Henry de Valence <hdevalence at gmail.com>wrote:

> FYI, the reason that the clipping was done manually was because there are
> issues with some projections, and the fact that QPainter's clipping is
> meant for flat objects, while this is doing clipping in spherical
> coordinates, clipping along great-circle arcs instead of straight lines.
>
> When I was refactoring the code into the two painting backends, I tried
> just using QPainter's clipping, but there were regressions. Maybe it works
> now -- but did you check all of the projections?
>
> Cheers,
> Henry de Valence
>
> P.S. I realize it would be better if I commented on the reviewboard, but I
> don't have an account, so I'm mailing here.
>
> On Sun, Mar 18, 2012 at 03:43, Rishab Arora <ra.rishab at gmail.com> wrote:
>
>>    This is an automatically generated e-mail. To reply, visit:
>> http://git.reviewboard.kde.org/r/104320/
>>
>> On March 18th, 2012, 12:11 a.m., *Rafal Kulaga* wrote:
>>
>> This patch indeed fixes the problem, but I wonder if the clipping here is needed at all? It seems that we can get rid of it.
>>
>>  I think so too! And that is what I wanted to confirm if we get any advantage at all from ignoring points manually instead of painters perform the clipping themselves. Attached p3.patch which works on both backends on the same principle.
>>
>>
>> - Rishab
>>
>> On March 17th, 2012, 11:12 p.m., Rishab Arora wrote:
>>   Review request for KStars, Rafal Kulaga and Akarsh Simha.
>> By Rishab Arora.
>>
>> *Updated March 17, 2012, 11:12 p.m.*
>> Description
>>
>> Fixes bug https://bugs.kde.org/show_bug.cgi?id=266929 for QPainter backend only.
>> At high zoomlevel, the points leave the FOV and are ignored completely, whereas their fill should not be.
>> Used clipping in QPainter to fix bug.
>>
>>   Testing
>>
>> Tested at various zoomlevels and resolutions. No issues found.
>>
>>   *Bugs: * https://bugs.kde.org/show_bug.cgi?id=266929<http://bugs.kde.org/show_bug.cgi?id=https://bugs.kde.org/show_bug.cgi?id=266929>
>> Diffs
>>
>>    - kstars/projections/projector.h (966d002)
>>    - kstars/projections/projector.cpp (e8c1aff)
>>    - kstars/skyqpainter.cpp (43f06e8)
>>
>> View Diff <http://git.reviewboard.kde.org/r/104320/diff/>
>>
>> _______________________________________________
>> Kstars-devel mailing list
>> Kstars-devel at kde.org
>> https://mail.kde.org/mailman/listinfo/kstars-devel
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kstars-devel/attachments/20120324/6a82189d/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: p4.patch
Type: text/x-patch
Size: 4764 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/kstars-devel/attachments/20120324/6a82189d/attachment-0001.patch>