Closed xwerswoodx closed 11 months ago
I push another build, it should be fixed for the location on the picture, I am still doing a check if it causes any other issue
Much better fix, Xwerswoodx found two spots with +1 and -1. I wasnt able to find any. For me is enought accurate to be implemented. Thoose spots Xwerswoodx found are just very very rare cases and just +1 or -1. We should check if runuo or servuo have the same problem there just for curiosity. 5895 1330 12
I ended my method to fix height, I think it does not worth at least how I did it. I think it increase cpu and memory usage a lot.
Yeah a bit silly, I would like to do some more checks to make it more accurate but it uses so much cpu if I do it, so I think with this way it can be best to stop, I am not sure what else I can do anymore honestly.
(Issue: #983, #617, #877) (Issue: #779)(Issue: #1112) (Issue: #720)
Tested with 7.0.97.25 -It is hard to find a spot where sphere gets different Z than UOClient. In thoose cases, diference is -1 or +1, much better than before. -This fix many terrain issues where Uoclient allow u to climb a hill but sphere dosnt. It is almost impossible to find one of this cases now. -I just find a case where heights get mess but I think it is very rare case. I upload a ScreenShot At screenshoot u can see diference betwen sphere(blue) and uoclient(orange) heights.
The arrangement is not perfect but offers a substantially significant improvement
Got to say there are few more points along code where terrain Z is taken. I dont know if fix needs to be applied on all thoose. Link to some of thoose points (I think there are more than thoose showed on the link, for example at GetPointFix() )