The problem was a wrong tile extent that was set in our client. Our client
uses OpenLayers (ol5) and a custom tile load function to read MVT data from
a SQLite database. In our custom tile load function we had specified a
fixed extent ([0, 0, 2048, 2048]) for each tile. We have fixed it by
Hi Jody, hi Andrea,
thanks a lot for your feedback.
Possibly our web app doesn't extract vector tiles in the correct winding
order.
I'll inspect our code and let you know the results.
Greetings,
Mauro
Am Fr., 15. Nov. 2019 um 16:33 Uhr schrieb Andrea Aime <
andrea.a...@geo-solutions.it>:
>
The scaling difference is to be expected, if memory serves me right it
switched from a 256px grid to a 4096px one,
to get better accuracy especially on hidpi displays, and when oversampling
the last cached level. It's an improvement, not a bug.
Commits here:
-
Thanks for the detailed report, checking the release notes here:
https://osgeo-org.atlassian.net/jira/secure/ReleaseNote.jspa?projectId=1=16746
Shows a number of map box vector tile problems being fixed that would
result in differences between 2.14 and 2.15.
--
Jody Garnett
On Thu, 14 Nov
Hi,
after upgrading Geoserver from version 2.13 to version 2.15.3, I have
encountered an odd issue concerning GWC and MVT:
Mapbox vector tiles (pbf as well as generated mbtiles) that are generated
by GWC are rendered at the wrong location on the map.
Coordinates, extent, and scaling of vector
