[Marble-devel] Review Request: GeoData KML and GPX Parser: parse ExtendedData in gx:Track

[Niko Sams]

On Sun, Dec 18, 2011 at 12:23, Thibaut Gridel <tgridel at free.fr> wrote:
> Hi,
>
> replying to list as i couldn't upload patch in the review.
>
> On Saturday 17 December 2011, Niko Sams wrote:
>> > On Dec. 9, 2011, 11:04 p.m., Thibaut Gridel wrote:
>> > > src/lib/geodata/data/GeoDataTrack.cpp, line 184
>> > > <http://git.reviewboard.kde.org/r/103191/diff/5/?file=42953#file42953li
>> > > ne184>
>> > >
>> > >     This implementation does not keep date sorted track, so breaks the
>> > >     display of for instance satellite.
>
>> > Thibaut Gridel wrote:
>> >     Well, this method asks to add one point with a given timestamp
>> >     (unlike the appendxx which barely follow a file parsing convention).
>> >
>> Sorry for the late reply.
>>
>> I understand the problem, but I don't get your proposed solution :/
>> How do you suggest the data should be stored, and when should the sorting
>> happen?
>
> Attached is my patch which keeps the data that can be sorted so.
> It should apply on your proposed one.
>
> As I said, track should keep the idea of consecutive positions of a path.
> Because it is valid to have points without time, file parsing providing the
> context, we can't do much qSort or anything, but keep the QList simply
> following the convention.
>
> addPoint( const QDateTime &when, const GeoDataCoordinates &) permits to add
> points at arbitrary time for example... I think we should support this.
>
I have a very similar patch, but the performance when showing
satellite orbit is very bad.
I will try if you patch is better, maybe due to the simplified remove* methods.

Niko