Something happened during server track change from WE1 => RO9 on a 0.6K7 server.
Some people experienced a crashed LFS. I didn't, I was minimized.. But when I switched back I was missing almost all textures, could see 'through the ground' most was gray... Should have taken a screenshot but I thought it was me.
---
Ah somebody did make a screenshot, thanks Joona ... Too bad we don't have a crash address... Maybe it can be recreated, will try later.
Yeah, added a [LEFT 0 - MID 1 - RIGHT 1] monitor config in View options, added the correct widths (1440, 1920) and it seemed great (first picture), except the 2-3 pixel wide line on the right monitor while in pits/main menu (second picture), which is probably the scaling issue. Part of the UI on the left screen is also out (upwards).
Again, I just test it, I don't have any real goals with these two monitors together for LFS.
[I try to retrieve/ask that crash address from somebody's log] ...
Ehh something else, can something be done about those white things here on the spectators area? I don't know if it's me but they appear and disappear all the time. It's like they are pillars but they are not. It's "flickering" when I drive across the straight.. (sry about the 1.2 Mb PNG, I am in PNG mode).
Thanks for the info. I'll look into this in the morning.
After the track change, before entering the game, were people on the grid? I'm asking because only see one car in the race at first, and it seems to have started at the origin.
Does this replay start from when you first entered the game from the lobby, or is it after a quick restart?
I don't know because lobby isn't included in replays. But at least someone needs to be at the grid to click ready. Can I check if there was more then one person on the grid formation.. I don't think it is in any log... Hmm..
There is no quick restart anymore like Airio used to do that to sort the grid (IS_REO).. Been too lazy to re-implement that @PRISM so between lobby and the start of this replay nothing special happened. Well.. Four, I guess, S2 licensed racers got kicked according to the deb_log... Maybe one or all of them where on the grid formation, no idea. Maybe someone can remember, difficult to question them
I think I've figured out this bug now. It's really an old bug, nothing to do with Rockingham. I can reproduce some of the symptoms by allowing a guest to set /ready when the host changes track. The bug is that there is no protection on the receive of the ready packet. The protection is only on the send. If there is only one player in the race and he clicks "ready" within a few ms of the host sending the message to change track, his computer allows him to send the "ready" packet. The "change track" packet is the first to arrive, and the player in race is set "not ready". Then his "ready" packet arrives, LFS concludes "everyone in the race is ready" and enters the game, even though the track is not fully loaded. The track can be is various states of loadedness on all the guests, so some crash, others don't see objects, etc. In this case the "ready" player was jHnZ1997 who was continually trying to join the race and click ready, according to deb.log.
EDIT: Thanks for all the evidence. I used the screenshots, the deb.log, the crash address, the MPR and descriptions to get to the answer.
Ah wow! I've seen that behavior over and over again but it never leaded to this massive crash, very rare bug then.. If it is too much work to fix this I would say just skip it..
Rule #1 in multiplayer games: Never trust the client.
Nice found both of you, funny that LFS is so stable that you have to fix bugs like this while other devs would have said "Will not be fixed now, too much bugs to fix before.". Keep it up!
NOTE! You only need the small K8 patch if you already have K7.
Misc :
Moved ABS option in car setup above the Traction Control option
TCP position packets option now also sends TCP packets to host
Borderless window system :
Cleaner switch between (borderless) windowed and full screen modes
Full screen vertical sync option now affects borderless window mode
VR :
Warning if Rift headset is opened using the OpenVR system
Text command /rift changed to /vr (but /rift still works)
New parameters for the /vr command : /vr=rift and /vr=openvr
Fixes :
User interface elements should now be perfectly aligned with screen
Rare MP bug could cause LFS to enter world before track was loaded
AI path drive limits were wrong at one corner of Kyoto National
I don't have this problem all time with all servers, it's like the UDP packets are being blocked/unblocked depending on day/hour randomly, some are working fine and some not ( depending on days ), so I need a quite large servers list to test with .