agg changes

Sean Gillies sgillies at FRII.COM
Mon Aug 27 11:08:27 EDT 2007


Stephen Woodbridge wrote:
> thomas bonfort wrote:
>> hi devs,
>> I've commited a few changes to the agg rendering logic, which greatly 
>> increase performance when drawing many shapes. what's changed is that 
>> the imageObj now holds a pointer to the rendering object and all it's 
>> dependencies, instead of recreating them nearly from scratch each time a 
>> shape is drawn.
>> on the "ntdb_roads_50" from a recent osgeo-discuss post
>> before: 1.1seconds
>> after: 0.5 seconds
>>
>> this is still much more than what gd does, but getting closer (that 
>> particular test file is probably a worst case scenario when comparing 
>> rendering times: many tiny line segments that just show up as a pixel 
>> with gd, but all have to be antialiased in agg)
>>
>> all this to say that I don't think there should be much breakage arising 
>> from this, but I'd appreciate if all of you who are set up to test agg 
>> could give the new version a run and report if anything weird is happening.
>>
>> thanks,
>> thomas
> 
> Thomas,
> 
> This is built with SVN revision 6709.
> 
> http://imaptools.com/cgi-bin/mapserv-4.99?mode=map&map=/u/data/test/spiral.map
> http://imaptools.com/maps/compare-maps2.php?loc=1&ll=42.61935+-71.38918&address=&city=le+paso&state=tx&zipcode=&country=&asrv=1&amf=%2Fu%2Fdata%2Fmaps%2Fgoogle-aa2.map&msa=mapserv-4.10&bsrv=1&bmf=%2Fu%2Fdata%2Fmaps%2Fgoogle-agg.map&msb=mapserv-4.99&submit=Show
> 
> I noticed that the endcaps of fat lines used to be rounded like GD, but 
> now it abruptly ends. look at the ends of the test spiral and the roads 
> on the bottom of the side by side comparison.
> 
> Also, You meantioned a while back that some or all of the cartoline 
> endcap styles worked with AGG, could you review for me and the doc team 
> what the keywords are and how to include them in the AGG layer definition.
> 
> Thanks,
>    -Steve W
> 

Seems to me that AGG would make it possible to do away with the 
cartoline stuff, right? It was only ever a hack to overcome weaknesses 
of GD anyway, and became poorly supported over time.

Sean



More information about the mapserver-dev mailing list