I confirm this. Sorry about that! It's an easy fix, just an oversight. Today I'll be looking at what can be done for a fairly quick update. Not sure which day it'll be yet.
I think I've come across another bug related to the ISP_MCI packet.
It seems like my Insim isn't receiving CompCars for all players. We have ~16 players online, and there are always 1-2 players for which I am not receiving CompCars within the ISP_MCI packet.
I'm on Insim Ver. 9 and I have the feeling that it has something to do with the latest change to the packet length. I feel like there is one person per 8 player missing. Maybe an index or counting error when trying to split all the players into ISP_MCI packets?
It's really hard to test, but when myself was affected, I asked someone to go to spec mode, and suddenly I got position updates for myself again, and the other person was affected after they've joined the track again.
I've looked in the code and it's really exactly the same code, only using 16 cars as the maximum per pack instead of 8. So I can't really see how it would go wrong.
Have you definitely got the number of cars per IS_MCI correct?
something like:
int pack_size = pack->Size * 4;
int num_cars = (pack_size - 4) / 28; // max 16
Hi I can confirm that there's something funky with IS_MCI
I've set up 17 AIs to a race and connected insim - i've got 2 MCI packets with 8 CompCars each:
I've tested some other combinations of cars:
3 AI cars - 1 MCI packet with NumC = 3
9 AI cars - 1 MCI packet with NumC = 8
15 AI cars - 1 MCI packet with NumC = 8
16 AI cars - 1 MCI packet with NumC = 8
17 AI cars - 2 MCI packets with NumC = 8 each (contains first 16 PLIDs)
25 AI cars - 2 MCI packets with NumC = 8 each (contains first 16 PLIDs)
32 AI cars - 2 MCI packets with NumC = 8 each (contains first 16 PLIDs)
couldn't test with more because couldn't think of a way to have more AI in track
All right, I've got it now. Thank you very much for the test. It really is a stupid bug of mine. That function was TOO similar to the old one. Although it knows the correct number of cars per packet, at *one point* it refers to the number 8, instead of that.
So what it does: sends the correct number of packets as if it would be adding 16 cars in each packet, but only fills 8 cars into each one.
I'm sorry about the confusion. There was a long big push where I was converting a lot of packets in LFS and actually could not test every one. I shouldn't have missed this though. Anyway it's fixed in my code.
I don't think I can do an update tomorrow but I hope to get an version done on Monday afternoon.
Cheers for the testing. I've reverted to protocol version 8 for now, but glad you found the cause. Looking forward to activate version 9 again whenever you say.
Noticed a potential bug today using W45 (sorry if already mentioned)
1. Joined an online server with a limit on which mod could be used.
2. Left the server and entered single player mode.
3. Opened the mod menu.
4. In the browse section the only visible mod was the mod that was allowed in the online server previously. The remaining mods could still be selected via the installed list, but they weren't visible in the browse menu. Tried different filter settings with no luck.
Looks like some of the mod images should take vertical size into consideration when shown on the Garage and/or Vehicle Mods screen, as they're being sized to fit horizontal width, but are too high to fit on screen, with bits cropped off.
It's the numberplate of the first person who joins with that mod, I think, rather than a timer. The other day everyone joining with mods had my numberplate on their car, after I had joined first. (IIRC correctly this was a bug that happened before mods and was fixed)