Re: [deegree-users] Regression between 3.2-pre2 and 3.2-pre7?

[Andreas S. <sc...@oc...>]

Michiel J. van Heek wrote:

Hi,

> > Michiel J. van Heek wrote:
> >
> >> > Michiel J. van Heek wrote:
> >> >
> >> >> > Michiel J. van Heek wrote:
> >> >> >
> >> >> >> Quoting Andreas Schmitz <sc...@oc...>:
> >> >> >>
> >> >> >> > Michiel J. van Heek wrote:
> >> >> >> >
> >> >> >> >> Or a threshold for the number of vertices? Because the number of
> >> >> >> >> features in my map is not that big (all countries in the  
> >> world), but
> >> >> >> >> the number of vertices is huge (e.g. very detailed coastlines).
> >> >> >> >
> >> >> >> > my guess is that this is related to a recent improvement regarding
> >> >> >> > exactly that use case (rendering a small portion of geometries
> >> >> with many
> >> >> >> > vertices), some tests currently fail for pre7 snapshots. Can you try
> >> >> >> > with the pre6 release and see if the problem still occurs?
> >> >> >> >
> >> >> >> Indeed, the problem does not occur in the pre6 release. However, in
> >> >> >> the pre7 snapshot (built last night) the problem persists. So this is
> >> >> >> something you are already working on to solve?
> >> >> >
> >> >> > yes, expect a fix in a couple of days (I'll keep the list posted).
> >> >> >
> >> >> Great to hear!
> >> >
> >> > this should be fixed in the latest SNAPSHOT, although we're still trying
> >> > to figure out a test failure. Can you verify that your issue is fixed?
> >>
> >> Built my webapp with the latest 3.2-pre7-SNAPSHOT. Unfortunately, same
> >> result: no map, no errors in log, when too many features / vertices. I
> >> will try to solve this problem by simplifying my geometries on
> >> small-scale zoom levels.
> >
> > still that's strange. Can you provide us with the data so we can analyse
> > the issue further?
> 
> Sure! I will send a PostGIS dump of my countries table and the  
> relevant deegree config files to you off-list. Thanks for looking into  
> this.

thanks. There was still a bug in the clipping optimization, which I
fixed just now. Tomorrows nightly should contain the fix.

Best regards,
Andreas Schmitz
-- 
Occam Labs UG (haftungsbeschränkt)
Godesberger Allee 139
53175 Bonn, Germany
+49 228 937 988 65
http://www.occamlabs.de