Drag is not fluid enough

[virshlee]

I compared flutter_map to the normal google maps. Don't get me wrong, flutter_map's drag works fine, but it's as not smooth as the one in google maps. Do you have any idea what's the difference?

[kleeb]

Google Maps SDK has prebuilt vectorisation. Thats why this zooming in/out feels so smooth when in the meantime its downloading all relevant tiles. Not sure what would need to be done to make the UX better, maybe some rasterisation of the tiles before we get a new one from the network.

[johnpryan]

also related to flinging behavior: #271

[virshlee]

I compared flutter_map to the normal google maps. Don't get me wrong, flutter_map's drag works fine, but it's as not smooth as the one in google maps. Do you have any idea what's the difference?

I see it even when all the needed tiles are loaded, also I checked the javascript web implementation of leaflet, and that was also smoother than the flutter version, I haven't checked the full code there though, do you have any idea about the main differences in the dragging? I think that the move method itself is the problem, as the bit jerky movement is even seen when I make an animated move to a certain location, not even dragging.

[johnpryan]

Right now what we have is a rough attempt to allow for two-finger dragging and zooming. Pull requests are accepted :)

[virshlee]

Right now what we have is a rough attempt to allow for two-finger dragging and zooming. Pull requests are accepted :)

I will take a look at it, also it is possible that it would be better if I could run it in 60fps, I can't go above about 40 fps even in performance mode.

[virshlee]

Right now what we have is a rough attempt to allow for two-finger dragging and zooming. Pull requests are accepted :)

Do you think the whole plugin could be made more fluid by multi threading some back-end work? Ex the gesture could be calculated on a different thread or isolate as it's called in flutter I think.

[MihaMarkic]

I feel that putting gesture calculation in isolate would yield worse performance. Or is that calculation that much demanding? Perhaps some perf measurement could help.

[maRci002]

Hi, @virshlee could please take a test ride with #572 PR, maybe it resolves this issue.