Loïc Guilloux
Loïc Guilloux
I don't really understand the problem (except the visual issue for shores), but the fix seems to be a hack to me.
After more reading of the diff, I think the problem is "edge tiles are supposed to self flood but the checks are too strict and most of them don't flood"....
Single core speed is what matters for OpenTTD. And if possible try to put the server on the less powerful machine. Enabling "pause on join" may help too.
OpenTTD doesn't use steam network feature at all. Non-Steam and Steam executable are exactly the same.
Likely duplicate of #9956 and #9876 (same trace)
There's a runtime cpu check for this case via `ViewportSortParentSpritesSSE41Checker`, could it be this function triggering illegal instruction ?
`ViewportSortParentSpritesSSE41Checker` is called way before, and seems to return `true`, as it really looks like `ViewportSortParentSpritesSSE41` is called here (at least that's a guess from the call stack). But that...
I made a quick (and maybe dirty) test on https://github.com/OpenTTD/OpenTTD/compare/master...glx22:OpenTTD:sse
I can PR my test branch.
I think it should be possible to do something similar to `STR_ORDER_GO_TO_STATION_CAN_T_USE_STATION` to prevent string duplication. The strings would be like `{PUSH_COLOUR}{YELLOW}Description{POP_COLOUR}Cost` and you can change text colour for cost...