View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0005576 | Spring engine | General | public | 2017-05-27 15:43 | 2017-06-19 21:06 |
| Reporter | Google_Frog | Assigned To | |||
| Priority | normal | Severity | major | Reproducibility | sometimes |
| Status | closed | Resolution | unable to reproduce | ||
| Product Version | 103.0 +git | ||||
| Summary | 0005576: (Circuit AI?) 103.0.1-939 Desync | ||||
| Description | This game desynced: http://zero-k.info/Battles/Detail/457999 Here are the infologs from both players: * https://github.com/ZeroK-RTS/CrashReports/issues/1165 * https://github.com/ZeroK-RTS/CrashReports/issues/1166 | ||||
| Tags | No tags attached. | ||||
| Attached Files | |||||
| Checked infolog.txt for Errors | |||||
|
|
desync with AI present, so not necessarily engine. |
|
|
circuitai isn't tested well, so memleaks / memory corruption, etc is very likely. several test games with spring + ai compiled fsanitize=address should be made. reported upstream: https://github.com/rlcevg/CircuitAI/issues/45 |
|
|
Circuit's "Warning: AI for team 1 (ID: 0) failed handling event with topic 9, error: 209" messages are engine errors, not AI. topic 9 is EVENT_UNIT_DAMAGED. Engine loves to send EVENT_UNIT_DESTROYED prior to EVENT_UNIT_DAMAGED. Hence warning. WRT fsanitize=address: I'll make test-runs with Circuit 0.9.15.a. (0.9.15 was thoroughly tested, 0.9.15.a is basically recompilation of 0.9.15) |
|
|
"Engine loves to send EVENT_UNIT_DESTROYED prior to EVENT_UNIT_DAMAGED." citation needed. CUnit::DoDamage checks isDead before sending the UnitDamaged event, isDead is set by CUnit::ForcedKillUnit at the very top before UnitDestroyed is sent. |
|
|
> citation needed. Test subject is 103.0.1-939-g2b3f946 and ZK-1.5.5.1 Compile&Run this CppTestAI: http://paste.springfiles.com/view/025233e1 with this startscript: http://paste.springfiles.com/view/077bc842 It will produce next output: [f=0009930] Skirmish AI <Test AI using the new C++ wrapper-0.1>: EVENT_UNIT_DAMAGED: 12319 [f=0009930] Skirmish AI <Test AI using the new C++ wrapper-0.1>: EVENT_UNIT_DESTROYED: 12319 [f=0009930] Skirmish AI <Test AI using the new C++ wrapper-0.1>: EVENT_UNIT_DAMAGED: 12319 [f=0009930] Skirmish AI <Test AI using the new C++ wrapper-0.1>: ERROR :Unknown Unit DAMAGED: 12319 "loves" was an exaggeration, but it happens. And I can't edit own messages... |
|
|
> circuitai isn't tested well, so memleaks / memory corruption, etc is very likely. sorry, that was very incorrect, didn't watch progress of CircuitAI for a long time. sync error is at frame 57481, so a lot of other stuff could went wrong. |
|
|
not effectively investigable. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2017-05-27 15:43 | Google_Frog | New Issue | |
| 2017-05-27 15:43 | Google_Frog | File Added: 20170525_145508_IsisDelta_v02_103.0.1-939-g2b3f946 develop.sdfz | |
| 2017-05-27 15:49 | Kloot | Note Added: 0017703 | |
| 2017-05-27 16:35 | abma | Note Added: 0017707 | |
| 2017-05-27 16:36 | abma | Summary | 103.0.1-939 Desync => (AI?) 103.0.1-939 Desync |
| 2017-05-27 16:36 | abma | Summary | (AI?) 103.0.1-939 Desync => (Circuit AI?) 103.0.1-939 Desync |
| 2017-05-27 16:40 | abma | Note Edited: 0017707 | |
| 2017-05-27 16:55 | lamer | Note Added: 0017709 | |
| 2017-05-27 17:12 | Kloot | Note Added: 0017710 | |
| 2017-05-27 17:12 | Kloot | Note Edited: 0017710 | |
| 2017-05-27 18:57 | lamer | Note Added: 0017713 | |
| 2017-05-27 20:11 | abma | Note Added: 0017714 | |
| 2017-06-19 21:06 | Kloot | Status | new => closed |
| 2017-06-19 21:06 | Kloot | Resolution | open => unable to reproduce |
| 2017-06-19 21:06 | Kloot | Note Added: 0017851 |
