[OSM-dev] Blank Tiles at z23 and Deeper

[jpk]

On Wed, Oct 10, 2012 at 4:15 PM, Jon Burgess <jburgess777 at gmail.com> wrote:

> On Wed, 2012-10-10 at 11:08 -0400, jpk wrote:
> > The hash directory thing is good, and I can get that sorted (yay,
> > progress!), but it doesn't explain all the sea-colored tiles at >=z23
> > for tiles that aren't yet cached.  From what I gather looking at
> > mod_tile/renderd source, I see mapnik render the tile first, based on
> > the z/x/y coordinates, and then the hashing and caching stuff happens
> > after.  Is that right?  If so, it would seem that the problem lies
> > somewhere in there.  Any pointers as I start looking in that
> > direction?
> >
>
> At zoom 23 the world dimensions are 2^23 * 256 pixels, or 2^31. Any
> calculation done on these pixel values with signed 32 bit arithmetic has
> a risk of overflowing or changing sign by accident.
>
> I cannot tell you where this limitation occurs. If you really want to
> fix it then you could use a debugger to see where things go wrong. Or
> you could try compiling mod_tile and Mapnik with the GCC the -ftrapv
> option which should trap any signed integer overflow.
>
>     Jon
>

Ah, the number 23 seemed odd to me until now, but this makes sense.

Relatedly, I've been playing with Mapnik and so far haven't run into
problems getting toy stuff to render down to z30, so I suspect the
interesting stuff related to my blank tiles is happening in
mod_tile/renderd.  Anyway, I plan to dig into that next chance I get.

If I get something working and want to share patches, should I post them
here on dev@ or is there a better place?

-jpk


-- 
john p. kiffmeyer
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20121016/2dad0096/attachment.html>