Small bug in the appearance of the "Lost connection nickname (username)" message. The '(username)' part is colored the same as the last character in the nickname. Only spotted it once and it was red in that case. Figured it's worth a report because if it's not a bug and it's intended to be red it wont cost you much time.
After that i took a quick look at other similar messages that could be affected but found nothing out of the ordinary.
Maybe it isn't a big problem, but anyway: sometimes my game is crashing when i'm close it. I don't see error or something, just hear sound and game doesn't close while I don't press Enter or Esc. Same here with changing textures quality (low to high and back), but it happens when some track is loaded.
I also have problems with crashing on loading tracks aswell as exiting game. It happens every now and then, and is not uncommon. The thing in common is that it is stucked in an infitive loop and I canot get up an error table / info.
Nvidia Gforce GT420M Cuda 1GB
Intel Core i5-460M processor
DDR3 memory (3GB)
Don't know if this can be at any help, and it's worth mention that it has happend before the latest version too. Cross checked with a computer of a little older age, XP and dual core 6850 something if I remember correct.
Let me know if there is anything you can work out of that info.
I can reply on this as well because my IP changes every night. Normally I get disconnected but this night it stayed stuck in the "no cars in race" message and a lag bar which was raised in red all the way to the top of the screen. I had to [esc], click button [leave] and rejoin. Everything back to normal after that.
Audi TT timed out during track change;
This happened also quite often in the old version with various people. So people loosing connection during track changes is not new for me.
Didn't get any message, just stayed there like there still was a server connection. Only indication was the gigantic red lag bar. Didn't try to actually join or type something because I knew what happened. Wasn't to be bothered reporting it too because it is just so normal to me, loosing connection because of this IP change.
Hhmmm I think it's quite good in general. Seen some time outs in the log but nothing unusual in my eyes. It's too bad that status full server cannot be reached anymore, not enough people installed the patch.
Even have a hard time promoting it because there are more people then I thought having difficulties to just install this simple patch manually.
I experienced this in 0.6B, this bug is here for longer time
edit: Scawen, can you please add arrow when going out of pits on open configs, like on standard tracks. I know that cruisers wont like it, so maybe you can make it appear only if /cruise=no. Often there is real mess on open config tracks on cargame, people don't know which way they should go...
There is no 'real mess'. Airio auto spectates when somebody drives the wrong direction in pit. Sure its nice if these little things will be improved someday but that is not what this patch is about. Hmm
Btw.. On topic...
Didn't time out like time out with a message from server again this night. Same big red bar reaching top of my screen. No message nothing. Just "no cars in race".
If you had tried, I'm sure nothing would have happened. Only a packet from the host can make you join the track, after you've sent a request to join. But possibly that could have made the socket report an error...
The red bar to the top of the screen should be easy to fix. It was on my notes to do but I wanted to get the version out and tested. The obvious missing thing is a time-out on the guest's side. I should make it disconnect if it hasn't received a packet from the host after 40 seconds or so.
Still I find it a bit mysterious that Windows is under the impression that the TCP connection is alive and healthy even after your IP address has changed and there is no more communication between your computer and the host. It doesn't seem to be reporting any kind of error on the socket, otherwise LFS would disconnect.
Maybe that is normal for TCP, I don't know. Anyway I'm sure it could only happen because in this new version, if you don't have a car on track, the guest doesn't try to send any packets at all unless it receives something from the host or you click something in the interface (like join, change name, etc.)
EDIT : Further reading suggests this is not mysterious, just normal. The sockets were not set with the SO_KEEPALIVE option, so when no data is sent there is simply no checking (by the operating system) to see if the connection still really exists. I'll just implement a simple guest-side time-out as planned.
I think quite a lot of people actually did test this patch and it looks like its stable. I know roadmaps are a bit sensitive with LFS development but is there a general outlook for an official patch update? Is there still a lot on the list besides the time out routine and some other smaller fixes?
The lag overview which talked about earlier and being said that it requires another incompatibility test patch, is that planned to be taken care of.. Or?
What I'm trying to figure out is, is it better to downgrade back now to 0.6B4 or are we near an official (intermediate) release?
FIX : Two instances of LFS no longer write to the same MPR / SPR
Connection now closes if no packets are received for 40 seconds
SHIFT + N : sound on / off now works in multiplayer screens
Pit exit direction arrow now works in open configurations
Updates to Latvian, Serbian, Indonesian, Romanian
Latvian training lessons now included
I was thinking the same thing, and the pit direction arrow is now implemented (in B8) based on the direction of the start grid, for open configs. The path direction is still used as before for normal configs. Let's see if it does work in all places. I guess it will, unless there is a pit lane whose direction is very different from the grid direction. Can't think of any though...
My plan is to release that incompatible version with the ping and lag info next week. It's not a really big deal, should be easy enough to code and will just be a very slight bandwidth increase to send the info to the guests. Do tell me though if it doesn't really seem needed. I know chucknorris feels it is important, though I'm still convinced it is less important now than it was before this multiplayer update.
I understand your uneasy feeling about the divided community. It's not good for players on either side of the test patch wall. But looking ahead, a proper test patch period creates reliable LFS versions, so we have to be patient. I hope you will not revert to B4 because that would reduce the testing.
Which operating system is that? I think you should take a look at the cause of that timer bounding...
It is indeed less important as far as checking the overall server health is concerned.
However, IMO it is still very important to be able to easily monitor actual latency of individual players somehow, especially when there is close quarters racing going on as a laggy (although I suppose jitter is just as important) player can easily cause havoc.
It's quite surprising how much useful connection info the dancing of the old lag bars could show to the experienced eye.