[OpenLayers-Users] limitation of fix #487 to wrap the dateline

[Ryan Hofschneider]

On Nov 30, 2007, at 1:32 PM, pdecker wrote:

> Could someone explain the limitation listed below?
>
> Limitations/expectations of this patch:
> ...
> Requests which cross the date line are *not* wrapped, so they will  
> request
> content outside the world if it does not divide evenly.

Although I do not know what the above limitation means, I can report  
that I have had much success with the wrapDateLine option.

If you haven't seen the live example of this feature already, it is at:
http://openlayers.org/dev/examples/wrapDateLine.html

Some trickery will often have to occur at the WMS service provider  
level to support features that span the anti-meridian however. I am  
using GeoServer with PostGIS+PostgreSQL, neither of which provide  
support for wrapped features. I must split linestrings and polygons at  
the anti-meridian, transforming them into multilinestrings and  
multipolygons. The features I ultimately serve up from the database  
are wholly contained in the [-180,180] longitude range.

If this is all that you do for polygons that span, then you end up  
with a vertical line where the two parts of the split polygon meet.  
Depending on how you are trying to render your polygons (e.g., filled  
or not) you can do some additional trickery with your Style Layer  
Descriptor (SLD) and additional linestrings of the polygon's outer  
boundary to eliminate that vertical line.

I have also experimented with various hacks to properly display of  
vectors and markers from a WFS layer on a wrapped display. I have  
accomplished this by brute force -- the solution is quite inelegant,  
inefficient, and breaks some core functionality. First, the entire  
data set from a WFS layer is loaded all at once. Second, each feature  
is displayed three times -- once at its original location, at a -360  
degree offset, and at a +360 degree offset.

Cheers,
Ryan