Hello, 2012/4/11 David MENTRE <[email protected]>: > > I currently don't understand why we have this 1/8089 Mapnik scale. The > scale computation seems pretty similar to single_page_renderer. > Multiplying the multi_page_renderer default scale by 1.5 (the > approximate error ratio between 1/12000 and 1/18089) "fixes" the issue > (i.e. we would have 16 pages for Chavagne).
OK, I think I found it. I think we should not correct with current latitude correction factor. My reading of the code is that we don't need it because the we are using Mapnik conversion functions (projection and projection_envelope), contrary to the single page renderer. If Gaël can give his critical view on the attached patch. Best regards, d.
0001-Bug-fix-Don-t-correct-with-latitude-for-multi-page-r.patch
Description: Binary data
