[Mapnik-users] Difference output of labels from wms and command line

[Dane Springmeyer]

Jay,

On Oct 31, 2008, at 12:24 PM, Jay Douillard wrote:

> Hey Dane,
> Thanks! I'm  using the BC Albers projection, which makes BC look  
> more normal than the Mercator, which makes the top of BC look HUGE!
>

Ya, I agree - mercator is terrible for BC and Alaska. I've done a lot  
of watershed mapping of those areas (the units of which look a lot  
like your educational regions), and Albers IS excellent.

I've recently tweaked the BC albers projection to slant for the North  
Pacific region, which is fun. A sample image is here:

http://mapnik-utils.googlecode.com/svn/trunk/tutorials/unique_values/map/unique_values.png

Ironically, I used a python script to build those styles for that  
example and have yet to fix up the labeling priority, so it suffers in  
the same way as the current OGCServer patch (ie a few labels are  
underneath shapes).


> Yes I am using the 129 patch ( http://trac.mapnik.org/attachment/ticket/129/wms_theofilu.diff 
>  )
>

Okay, welcome to the testing group!

> #!/usr/bin/env python
> from mapnik.ogcserver.WMS import BaseWMSFactory
> from mapnik import *
> mapfile = 'map.xml'
> class WMSFactory(BaseWMSFactory):
>   def __init__(self):
>     BaseWMSFactory.__init__(self)
>     self.loadXML('/home/sysadmin/map/map.xml')
>     self.finalize()
>
> Which I cribbed from your example http://mapnik-utils.googlecode.com/svn/trunk/tutorials/wms/
>
> (Which just changed I noticed, the xml loading has moved to a new  
> file...)
>

Geez you are sharp :)

> I  tried your suggestion of putting the styles into difference  
> instances, and that seems to cause problems. I can still load the  
> "edi" shape layer but if I add in the second label layer (in my  
> OpenLayers WMS request) it causes a error and I get image tiles with  
> the error message written on them!
>


Okay, what is that error?  If it seems useful at all to debugging the  
patch, please post the details to the ticket:  http://trac.mapnik.org/ticket/129

> Once again this same XML works from the CLI.
>

Right, makes sense. The patch need improvement... maybe you're the man?


> Thanks for looking into this Dane!
>
> Jay
>

No problem. Looking into it is the easy part :)

The OGCServer code is, honestly, not the easiest to make sense of due  
in part to its attempt to meet the OGC spec.

And I have not had time or a paid-work need for OGC WMS recently.

So Jay, do you need to set up a truly public-facing OGC-spec WMS server?

Or do you just need to dynamically serve map tiles to your OpenLayers  
app without worrying about GetCapibilities and OGC compliance?

Dane



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.berlios.de/pipermail/mapnik-users/attachments/20081031/c9aeabf6/attachment.html>