[MapProxy] 2 Problems with meta_size in 0.9

[Holger Schiebold]

Thanks a lot Oliver,

I've just tried out and both problems are fixed. I have got another 
little point / question  - everytime I upgrade mapproxy I have to 
install the fastpng PIL version again. With the PIL version that comes 
with mapproxy at the moment I get wrong outputs from mapproxy. It seems 
to me that real black pixels (0 0 0) are changed to transparent. That 
problem especially appears  with labels and symbols. With the fastpng 
version of  PIL everything is fine.

Regards,
Holger



Am 15.10.2010 14:43, schrieb Oliver Tonnhofer:
> On 15.10.2010, at 10:56, Holger Schiebold wrote:
>    
>> First Problem: I try to setup in 0.9rc1 meta_size to a value like [2, 2] in the cache section of mapproxy.yaml but that has not the expected affect in smaller requests to the original wms. It seems to me that everytime [4, 4] as the default value is used. So i have to set the tile_size to 192, 192 in my case because this special wms we try to cache is only able to deliver up to 1600x1200.
>>      
> Thanks for pointing me to that issue. I was unsure where to put the the options for meta_size/meta_buffer and I put them to the grids section at first, but changed them later to the cache section. What I missed was to move the global options from grid to cache. I fixed that now and updated the snapshot at http://mapproxy.org/static/rel/MapProxy-0.9.0.dev-20101015.tar.gz
>
> So you can now configure the meta_size/meta_buffer for each cache and also in the globals.cache section::
>
>
> cache:
>    mycache:
>       meta_size: [4, 4]
>       sources: [xxx]
>
> globals:
>    cache:
>       # set new default
>       meta_size: [6, 6]
>       meta_buffer: 200
>
>    
>> Second : I can not seed with the original grid.py. I get following error :
>>   File "/opt/janus/pyvenv/mapproxy_dbwk2/lib/python2.6/site-packages/mapproxy/grid.py", line 779, in _meta_size
>>     return min(self.meta_size[0], grid_size[0]), min(self.meta_size[1], grid_size[1])
>> TypeError: 'instancemethod' object is unsubscriptable
>>      
>
> That's the issue I mentioned to Andreas. It is already fixed.
>
> Regards,
> Oliver
>
>    


-- 
Holger Schiebold


____________________
codematix GmbH
Felsbachstrasse 5/7
D-07745 Jena
Tel. +49 (3641) 3038-34
www.codematix.de