I'm glad it works for you and that you like it. 
Okram, you seem to be a magnet for all kinds of weird and rare errors.
But it is good you post them! But the current one, I really do not know what can I do about it. If I understand it correctly, your PC was in a very strange state (disk full), so probably Airio failed to write a small TXT file correctly and somehow it was not released, so subsequent attempts to write to the file ended in sharing error. If it is repeating often, I can try to add some safety code to check on this condition, but it is really so special and ... uhm ... unimportant. Cool, the hacked demo BL1 times are now removed from LFSW, BIG thanks to whoever did that. 
Lets hope at least WR data will stay reliable now.
Lets hope at least WR data will stay reliable now. If you know the IP, insim port, and password, you can try to connect to the server with Airio from home and see what errors it shows. Such (very) remote connection is not good for a long run, but it is great for testing and initial setup, because you have all the files on your PC, can change them any time, apply the changes immediatelly, all this without FTP and such...
I can certainly add something like that, but I do not understand the request completely. Airio knows UTC time and it knows server time. (And these two are reported in !ver, and server time also in custom !now command, if defined, optionally on race start or in other messages.) But how could Airio reliably know time zones of individual clients and show their local time? Also the !time command is already used for something different, so adjustments of command names would be necessary.
What confuses me completely is "a button in the same place found within the start screen between tracks". You mean to show server (or UTC) time on race restart, for a while? I believe that is already possible in Airio FULL/PROS versions, if admins choose to define appropriate race restart message. Or you mean something else?
The most tricky part for a start is to achieve the connection to server. For this three items need to be specified in Airio connection file, usually Airio.con.1.txt which is for server #1. The items are LFS server IP address, LFS server Inim port (not game port!) and the LFS server admin password.
IP address needs to be in the XXX.XXX.XXX.XXX format. If Airio is connected locally (on the same PC where LFS server also runs), general local address 127.0.0.1 may be used. This local run is in fact the preferred mode of operation. You can get the IP address from the server provider otherwise.
InSim port is set using /insim=XXXXX command on the LFS server, usually in setup.cfg server file. LFS provider may require using specific insim port which you cannot change. Otherwise you may choose your own, usually 29999 is used. (Note that you cannot set server insim port through client, this is possible only from server console or server config file.)
The admin password is set on server using /admin=sometext, again in server configuration file, which is setup.cfg by default.
When the three items match in server setup and Airio setup, connection is guaranteed... Uhm...
The error you mention is interesting, but showing AIRIO ERROR lines from Airio log.txt file would be much more helpful. 
An option is to use empty track and set just the 20 cars. But the check precision will go down naturally.
(But see below.)OMG... Well, correct (somehow, plus/minus) time adjustments should help, I think, but I'm not sure. But the idea was to take the WRs of supported restricted cars from AIRW, that's how it should work. But looking closer into this would require deep digging in the code.
Anyway, the cheating concerns mainly demo servers and demo WRs. Only a few licensed people would really try this cheat, because obviously it would result in permanent account ban. At least I hope it is obvious.
Protection from speed hacks
Well, so we have a new era of speed hacks upon us. Last time it was here about 2 years ago. This message gives basic info about what protection you can use in Airio (all versions) to get rid of hackers or at least try to.
1) You may limit maximum allowed speed, per track and car. This is done in Airio.tcd.txt file and it allows (using AllowedSpeed) to say that e.g. on BL1 a FBM can move at speed of 230 kmph max. If higher speed is detected (which may also be result of a crash), the car is spectated for safety. Simple and quite effective, you just need to know what are the realistically possible speeds. (For XFG and XRG on BL1 200 kmph is a safe value.) For the check to actually run and spectate impossibly fast cars also CheckSpeed in appropriate Airio.srv.?.txt file must be set to true.
2) Also possible split/sector times can be checked, comparing driver's data with current WR. This check is activated by setting CheckTime to true in appropriate (or the main) SRV file. Additional items to see are AllowedSectrTime and AllowedSplitTime, but default values of 9800 resp. 9900 are OK (except on oval, where you need lower values). Any car with sector time 2% or split time 1% or more under WR is immediately kicked for safety.
3) Unfortunately the latest development is very sad. Unbelievably, LFS World seems to be currently accepting hacked offline hotlap times as valid ones and presents those ridiculous values as new official world records. This completely screws (sorry) the principle used by point 2 above. Until the troubles are solved on LFS developers side, there's only one solution. Set UpdateWR in CFG file to false, so that new WRs are not read from LFSW. Overwrite the existing Airio.wrs.txt file with the one that is appended, one or two days old with realistic times. Restart Airio, ideally by !ai admin command (note that !rld will NOT work in this case).
Well, especially point 3 is really sad and has profound impact in many Airio areas.
1) You may limit maximum allowed speed, per track and car. This is done in Airio.tcd.txt file and it allows (using AllowedSpeed) to say that e.g. on BL1 a FBM can move at speed of 230 kmph max. If higher speed is detected (which may also be result of a crash), the car is spectated for safety. Simple and quite effective, you just need to know what are the realistically possible speeds. (For XFG and XRG on BL1 200 kmph is a safe value.) For the check to actually run and spectate impossibly fast cars also CheckSpeed in appropriate Airio.srv.?.txt file must be set to true.
2) Also possible split/sector times can be checked, comparing driver's data with current WR. This check is activated by setting CheckTime to true in appropriate (or the main) SRV file. Additional items to see are AllowedSectrTime and AllowedSplitTime, but default values of 9800 resp. 9900 are OK (except on oval, where you need lower values). Any car with sector time 2% or split time 1% or more under WR is immediately kicked for safety.
3) Unfortunately the latest development is very sad. Unbelievably, LFS World seems to be currently accepting hacked offline hotlap times as valid ones and presents those ridiculous values as new official world records. This completely screws (sorry) the principle used by point 2 above. Until the troubles are solved on LFS developers side, there's only one solution. Set UpdateWR in CFG file to false, so that new WRs are not read from LFSW. Overwrite the existing Airio.wrs.txt file with the one that is appended, one or two days old with realistic times. Restart Airio, ideally by !ai admin command (note that !rld will NOT work in this case).
Well, especially point 3 is really sad and has profound impact in many Airio areas.
It is very sad that LFSW accepted the hacked demo WRs as valid ones. I hope Victor will stay up all night making sure this cannot happen ever again. Or am I naive?
Well, it took less then I was thinking, so the FE path files are updated already. They are FE1, FE1R, FE3, FE3R, FE4, and FE4R. You can find the new path files archive here: http://www.airio.eu/files/LFS_PTH_V13.zip. Just upload the files to Airio directory and type !rld, nothing else is necessary. 
PS: Appended is example correction for FE1...
PS: Appended is example correction for FE1...
1) Well, this is kind of intentional, an extra point(s) for being lonely but persistent in your category. Sort of a "support" point.
2) Indeed, good point, I'll try to correct this, but it will take some time (days)...
It is a very strange error, I wonder if some other Linux users see the same. But it obviously happens only when connection to the database is being closed, so I think it is not that important, especially because the error is captured and everything else closes properly. The red line just says that no more database access is supported, when the program closes, just as it says from when the database is available after program start. Also when doing relatively lengthy database operations (think a few seconds max), such as !dbi or !dbe, the database is temporarily deactivated...
Airio 2.4.7
New Airio version hopefully solves especially the recently noted problems concerning SQLite support. Because Airio uses just one connection to the database, shared by all connected servers and multiple threads, under certain conditions collisions occurred leading to various errors. Database connection locking is applied now to make the sharing a safe matter.
The FULL/PROS version also contains options to raise WAVED blue and yellow flags to let people know more cars are approaching from behind or someone slow/crashed is ahead on the race path. Additional options allow to block race join for the whole 1st lap (or just from last split of the 1st lap) to 1st split of the 2nd lap.
The FULL/PROS version also contains options to raise WAVED blue and yellow flags to let people know more cars are approaching from behind or someone slow/crashed is ahead on the race path. Additional options allow to block race join for the whole 1st lap (or just from last split of the 1st lap) to 1st split of the 2nd lap.
Last edited by EQ Worry, .
The problem is if SQLite in 2.4.6 throws errors too often, some routines are not finished correctly and the impact is really unpredictable. I'd suggest setting UseSQLite to false for a day to be sure (within limits) the system works reliably. I'll make a better version available tomorrow, need to think it over a bit... 
Hard to say, a log of some time before the trouble appeared would be most helpful...
I think it is connected with the above problem. I believe there's a SQLite error (random) after people connect, which stops the new connection routine prematurely, leaving rank empty, at 0. I need to solve the SQLite troubles - I know the cause, now only a reliable solution is needed... An option would be to disable SQLite until 2.4.7 is ready. The database can be always imported again later.
Yes, I'm just looking into it (or at least trying to). Also it shows the opposite, cannot commit transaction, because there is none. It seems the transactions are somehow mingled, sometimes one closes another and sometimes one starts before another is closed. Doh...
Due to incorrect char this debug message will disappear only in 2.4.7...
Indeed the !wr command was not working quite correctly with lower/uppercase track characters (uppercase would give you correct results). The bug is corrected now (I think), ready for update in 2.4.7. Crady, I do not see a way this little bug could be connected with track rotations. But when I'm back home I'll be happy to run tests using your configs, which is the best way to unhide problems.
After a closer look I believe these are in fact two separate problems. The first one is that !wr shows incorrect data (at least the name, but sometimes time as well) for some cars on reversed tracks. This hidden bug is now corrected.
The second thing is some impossible custom car lap times stored on AIRW. I see only one way this could happen: 1) A guy joins with FX2, the used handicaps are stored. 2) Later he joins with FXR, but there is an error in the joining routine before the handicaps are updated. 3) He's driving with FXR, but on lap finishes the old handicaps are sent along, and he's seen as using FX2.
I have now updated Airio code, added a check the joining routine was completed and handicaps updated. Only if the check succeeds data are sent to AIRW. This update will be available in Airio 2.4.7 to be released after some testing of newly implemented waved flags and temporarily closed pitlane for joining (see the changelog).
The question is what to do with current AIRW data. For now I guess the best course would be to simply remove the impossible times. If you can help me to identify the items, I'll be grateful. When we're pretty sure the AIRW updates work reliably, an option would be to remove all the custom car lap times and start gathering new, clean data.
Hmmm, I'll test this when I can, which will be in a few more days. But I'm using Aonio all the time, switching between servers often, and I never experienced such kind of troubles - I do not see how replay saving and Aonio could be connected, except maybe UDP or TCP usage? Hmmm...
Then I don't understand how it could be possible... Are you sure the TrackRotate item is empty? (Because it takes precedence...)
If you're using tab-separated stats (which I'd suggest), you may find all names in full Unicode in the STA.UN file, one before last item on each row it is, not the 2nd one which is LFS bytes. (Another option could be to copy the necessary Unicode characters from the LFSEI table?)
Thanks a lot!
HUSSSSSSHHHHHH!
Indeed it is some a bug, two bugs in fact. I have corrected the VRS (virtual records) file generation on the AIRW server, changes will be propagated soon. Until small Airio update unfortunately only 1st part of names with spaces will be shown, but at least the parenthesis will not be empty. And anyway, who cares about incorrect display of usernames for people that use spaces? Uhm...
Seems to me like incorrectly configured LFSW access key. If using paid service, make sure you have credit. If using free service, the key entered in Airio's CFG file must NOT be used anywhere else AND ALSO no other keys under the same account may not be used anywhere else.
Oops, sorry, this one together with button click escaped my attention.
Yup, it is a know inconsistence, Airio runs out of buttons...
Hm, I'm really not sure. I use Aonio all the time and I'm checking replays very often too, and I never had any kind of troubles with this. When I need replay, I press 2 and save it named e.g. _. Then I run another LFS instance and open the replay there, so that I do not need to disconnect.
Airio 2.4.6
New Airio version is available, correcting the very unpleasant car type bug and also adding some required things, such as WR display in !top (and indeed WR lap/splits/sectors included in the output for easy lookup) or hiding commands when someone has blocked messages. Hopefully there are no bugs, uhm...
FGED GREDG RDFGDR GSFDG