|
|
|
|
|
|
|
|
This is the entire save folder. Save '2.sav' is the most recent, made right before we had to close do to stalled 'background worker thread' issue which also happens in single player. The main purpose of this post though is to check out the thousands of exceptions which should appear in the client-side log (cyborgs). |
|
|
There are 2 issues here then. stalled 'background worker thread' and these numerous exceptions. |
|
|
On the topic of missing notifications/alerts for the client side player, we found this time, with me hosting and the connection seeming more stable in general that we always both saw them. However, there was a noticeable delay in the 5 second or so range between it appearing for me and him and i assume what happened before is with a worse connection they were getting lost or timing out entirely before they appeared. |
|
|
4/7/2022 9:05:10 PM 4.014 CLIENT DELAYED11944 TID15 Error! Called FromActualInputEventOfPlayerID on game command 18585, which was already back in the pool! at System.Environment.get_StackTrace () [0x00000] in <695d1cc93cca45069c528c15c9fdd749>:0 at Arcen.Universal.ArcenDebugging.ArcenDebugLog (System.String Message, Arcen.Universal.Verbosity Verbosity) [0x00000] in <4384e505222f4699bade4a0ab6ffe4ff>:0 at Arcen.AIW2.Core.GameCommand.get_FromActualInputEventOfPlayerID () [0x00000] in <71252d42b88d4a6181b670e17cee4da8>:0 at |
|
|
|
|
|
notifier bug found again. I don't see Iego. Some of the numbers even look different. notifier.JPG (39,341 bytes) |
|
|
Here is my side and the save at that point. |
|
|
|
|
|
The error message part is fixed! * Fixed a one-line error that was causing multiplayer clients to eventually get a flood of errors relating to GameCommands already being back in the pool. ** What was happening was that it was being executed twice on the client if the command was a frame late, rather than only executing once. Mostly this would only show up after about 45 minutes of play, evidently. |
|
|
To my knowledge, all of the versions of these are now fixed. The stalled background thread thing may or may not be fixed by the next version, but there's a new protocol for finding those, which you can see here: * In the internal debug menu (the one from the escape menu), there are now "Log All Threads" and "Abort All Threads" buttons. ** If the game is getting stuck in a background deadlock, this will give us some information on it. The tooltips have more information on it. ** If you press the Abort All Threads button, it may crash the game (in theory -- so far it has not in testing), but it gives us exact stack traces (within reason) for all the threads it is aborting. Assuming that there is a stuck thread, and not a stuck Interlocked value, this will identify any deadlock for us to within the method that it exists in, at least. ** If we wind up with still having deadlocks and no interesting thread outputs here, then I'll instrument the Interlocked values to keep an eye on those. If the game gets stuck again, I'd appreciate a followup on a ticket like 0026729, since that's all related now, I think. |
- Anonymous
