Hello LFS Racers,

A lot of you already know, but many don't, because the information is buried deep inside a thread about a hacker.

In the next few days we'll be able to release a test patch with multiplayer improvements. It's not one that uses the new tyre physics which is still being developed, and does not contain any new content. I had to start working on this because hackers found a way to join a host as an administrator using a false user name, such as a single space. They could then use various cheats and could not be kicked or banned from the host. A dedicated host update 0.6B4 solved this particular problem but one of the hackers demonstrated that he was still able to cause a host to crash. Luckily, this time I was able to work out the probable cause and implement a fix. Just before that, I was already trying out a test of a new way for the multiplayer system to work - instead of waiting for messages from all guests, the system now only waits for messages from the host. So if there is one laggy connection, or someone lags temporarily, it won't have a negative effect on other people (other than seeing a lagging car).

Main improvements due to the new system :

Physics and all replays are still compatible
Game only needs host packets to continue processing
Quicker joining to host - no waiting for other players
No waiting for packets from players with slow connections
Much longer timeouts - lower probability of disconnection

Also there are fixes for various bugs and security issues that have been reported, and some more improvements that I can do because of the new multiplayer system.

EDIT : Test Patch 0.6B7 is now available for testing : http://www.lfsforum.net/showthread.php?t=80725

Great news, take your time.

Thanks for the update

Quote from Scawen :

I had hoped to release a test patch today but some of the bug fixes took half a day and some of the consequences of the new system are leading to various improvement possibilities and code simplifications which are hard to resist. So I expect a few more days on the case then release a patch for public testing.

Its better to finalize it, instead of just release an update with maybe even more bugs!

Great job on these updates. It's going to make the gameplay much faster. No more waitings for players to connect, so we can pit/spec/leave server(and even join server without waiting others now!)

I'm looking forward to this update
Keep up the good work Scawen.

Looking forward to this

Nice news about mp.

amazing news , can't wait to testing this patch.

Quote from DarkKostas :

No more waitings for players to connect, so we can pit/spec/leave server(and even join server without waiting others now!)

Well, this is quite interesting - what you suggest is partly true and partly not.

Yes, you will be able to leave the server without waiting while someone is connecting. This was a real problem in the past but now it should be fine, because the host can continue processing without your packets.

Pit / Spectate / Join - in the first version you will still have to wait while a guest is connecting (to preserve the game state that is being sent to the new player and avoid a Join Out Of Sync) but I'm hoping this will be less annoying already, because the connecting guest will be a bit quicker. His connection should take less time because he doesn't need to wait for a first packet from everyone else connected to the host.

The new systen does allow a possibility of taking away that wait completely, with some work on the way a guest connects. But that isn't really easy so I'd prefer to get the new thing tested. Then we can reassess what are the highest priorities at that point. I do want to get back to the tyre physics but we'll just see what comes up during the patch testing.

Quote from Scawen :

consequences of the new system are leading to various improvement possibilities and code simplifications which are hard to resist.

Quote from Scawen :

I do want to get back to the tyre physics but

I know its some sort of personal goal but don't hurry to go back onto that, that quick. It takes way too much time and I'm afraid not many are specifically waiting on that. (Partly because nobody knows what it is, true). This sim need real tracks and if its with fake cars or not, doesn't matter.

Just wanted to say. Waiting on further updates

Quote from Scawen :

Pit / Spectate / Join - in the first version you will still have to wait while a guest is connecting (to preserve the game state that is being sent to the new player and avoid a Join Out Of Sync) but I'm hoping this will be less annoying already, because the connecting guest will be a bit quicker. His connection should take less time because he doesn't need to wait for a first packet from everyone else connected to the host.

I get your point. I thought it was similar with this one,

Quote from Scawen :

Game only needs host packets to continue processing

but as now connecting is faster it won't cause big problem anymore; especially on (almost)full servers where you had to wait for 32 players to join.

Quote from Scawen :

Then we can reassess what are the highest priorities at that point. I do want to get back to the tyre physics but we'll just see what comes up during the patch testing.

We all want that and after that action of releasing an update, even that small but yet so important, it's making all of us get more hope and patience about the next big update (S3).

Nice Scawen

I may have not understood exactly, but does this update bring an end to the spectating-delay problem? You know, sometimes pressing shift+s does not immediately spec you (mostly when a server is full).

Quote from truepiece :

Nice Scawen

I may have not understood exactly, but does this update bring an end to the spectating-delay problem? You know, sometimes pressing shift+s does not immediately spec you (mostly when a server is full).

Not really, read this post:

http://www.lfsforum.net/showthread.php?p=1741647#post1741647

Quote from truepiece :

I may have not understood exactly, but does this update bring an end to the spectating-delay problem? You know, sometimes pressing shift+s does not immediately spec you (mostly when a server is full).

Flame is right to point you to the other post for more info, but it does depend on which kind of delay you mean.

Sometimes those actions are slow not because a player is connecting, but just because there are so many players, some of them have a high ping and that means that those actions can take a couple of seconds to react, when normally they'd be just half a second or so.

That case should be speeded up so it's no worse if there are many on the server (as long as the server has a good connection). I'd say that the host should be more "responsive" - it should hopefully feel like the server is on Earth and not the Moon.

Quote from Scawen :

Pit / Spectate / Join - in the first version you will still have to wait while a guest is connecting (to preserve the game state that is being sent to the new player and avoid a Join Out Of Sync) but I'm hoping this will be less annoying already, because the connecting guest will be a bit quicker. His connection should take less time because he doesn't need to wait for a first packet from everyone else connected to the host.

The new systen does allow a possibility of taking away that wait completely, with some work on the way a guest connects. But that isn't really easy so I'd prefer to get the new thing tested. Then we can reassess what are the highest priorities at that point. I do want to get back to the tyre physics but we'll just see what comes up during the patch testing.

Will that also cure the shift change bug?

Most spectate delay problems are occurring when somebody else is busy with timing out. As far as I understand this situation will be history.

Quote from BlakjeKaas :

Will that also cure the shift change bug?

This one I think you mean : http://www.lfsforum.net/showthread.php?p=1739042#post1739042
No, that is not affected by this change. I haven't studied that bug in detail yet, though I have been looking at various other ones that resulted in tidying up some cans of worms and leaving the code a lot cleaner and more understandable.

Quote from cargame.nl :

Most spectate delay problems are occurring when somebody else is busy with timing out. As far as I understand this situation will be history.

Thanks, good point and true.

Quote from Scawen :

because there are so many players, some of them have a high ping and that means that those actions can take a couple of seconds to react, when normally they'd be just half a second or so.

Maybe that's the case. The word "spectate" is displayed but the gameplay continues normally for some seconds.

Quote from cargame.nl :

Most spectate delay problems are occurring when somebody else is busy with timing out. As far as I understand this situation will be history.

Good update, thank you!
Quick question: How long will you be able to "time-out" before getting a drop out? I usually every now and then get a 10-15sec mega lag, which is due to troubles on the internet line (package loss).
Would be awesome if it would be increased to 30 secs or so (the time when the server does not get information from you), because then it would not make me drop out, but ofc. I understand if not - my problem should not cause problems for others

Quote from Scawen :

Quicker joining to host - no waiting for other players
No waiting for packets from players with slow connections
Much longer timeouts - lower probability of disconnection

Great stuff

Quote from The Very End :

Good update, thank you!
Quick question: How long will you be able to "time-out" before getting a drop out? I usually every now and then get a 10-15sec mega lag, which is due to troubles on the internet line (package loss).
Would be awesome if it would be increased to 30 secs or so (the time when the server does not get information from you), because then it would not make me drop out, but ofc. I understand if not - my problem should not cause problems for others

If we take that idea to its limits we could reset our router without timing out. That must be a cool thing to do

Quote from Whiskey :

If we take that idea to its limits we could reset our router without timing out. That must be a cool thing to do

Lmao! That would help, but would be a little far fetched
30 seconds is all I need, because usally it's just 10-20 sec delays then it's fine (50% of the time I can manage it without DC)

Allways fun to DC on last lap
*croses fingers*

The title "with same old physics" sounds frustrating . Anyway, nice to see updates that will keep the ragequits down

Great news getting that fixed for the better.
Looking forward to the test-patch! Keep up the good work