Factorio Forums

Posted: **Wed Aug 07, 2019 6:58 pm**

Ref:
- 74252 - uses similar custom fluid fuzzer in control.lua
- 74116 - this issue may be reproduced by this script

Steps to reproduce:

Load "fluid_fuzzer-crash.zip" (more waiting, in case of early save corruption) or "fluid_fuzzer-crash-b.zip" (less waiting)

fluid_fuzzer-crash.zip

(169.13 KiB) Downloaded 130 times

fluid_fuzzer-crash-b.zip

(170.61 KiB) Downloaded 138 times
Increase speed to x64
Wait some time (65s for first save, 5s for second save)

Crash

log

Code: Select all

   0.000 2019-08-07 20:48:37; Factorio 0.17.63 (build 46739, win64, steam)

Code: Select all

Factorio crashed. Generating symbolized stacktrace, please wait ...
c:\cygwin64\tmp\factorio-build-x93bju\src\fluid\fluidsystem.cpp (171): FluidSystem::addConnection
c:\cygwin64\tmp\factorio-build-x93bju\src\fluid\fluidbox.cpp (1613): FluidBox::updateBlockedUndergroundConnections
c:\cygwin64\tmp\factorio-build-x93bju\src\fluid\fluidsystem.cpp (433): FluidSystem::updateBlockedUndergroundConnections
c:\cygwin64\tmp\factorio-build-x93bju\src\fluid\fluidsystem.cpp (307): FluidSystem::updateFluidIDAndFilters
c:\cygwin64\tmp\factorio-build-x93bju\src\fluid\fluidmanager.cpp (286): FluidManager::moveOut
c:\cygwin64\tmp\factorio-build-x93bju\src\entity\pipe.cpp (85): Pipe::onDestroy
c:\cygwin64\tmp\factorio-build-x93bju\src\entity\entity.cpp (302): Entity::forceDestroy
c:\cygwin64\tmp\factorio-build-x93bju\src\script\luaentity.cpp (683): LuaEntity::luaDestroy
c:\cygwin64\tmp\factorio-build-x93bju\src\script\luabinder.hpp (326): LuaBinder<LuaLogisticContainerControlBehavior>::callWrapper
c:\cygwin64\tmp\factorio-build-x93bju\libraries\lua\ldo.c (350): luaD_precall
c:\cygwin64\tmp\factorio-build-x93bju\libraries\lua\lvm.c (710): luaV_execute
c:\cygwin64\tmp\factorio-build-x93bju\libraries\lua\lapi.c (1027): f_call
c:\cygwin64\tmp\factorio-build-x93bju\libraries\lua\ldo.c (137): luaD_rawrunprotected
c:\cygwin64\tmp\factorio-build-x93bju\libraries\lua\ldo.c (623): luaD_pcall
c:\cygwin64\tmp\factorio-build-x93bju\libraries\lua\lapi.c (1053): lua_pcallk
c:\cygwin64\tmp\factorio-build-x93bju\src\util\luahelper.cpp (3520): LuaHelper::pCallWithStackTrace
c:\cygwin64\tmp\factorio-build-x93bju\src\script\luagamescript.cpp (979): LuaGameScript::runNthTickHandler
c:\cygwin64\tmp\factorio-build-x93bju\src\script\luaeventdispatcher.cpp (120): LuaEventDispatcher::dispatch
c:\cygwin64\tmp\factorio-build-x93bju\src\scenario\scenario.cpp (1210): Scenario::update
c:\cygwin64\tmp\factorio-build-x93bju\src\mainloop.cpp (1191): MainLoop::gameUpdateStep
c:\cygwin64\tmp\factorio-build-x93bju\src\mainloop.cpp (1060): MainLoop::gameUpdateLoop
c:\cygwin64\tmp\factorio-build-x93bju\src\util\workerthread.cpp (42): WorkerThread::loop

Code: Select all

  84.723 Error FluidSystem.cpp:171: connection.isConnected() was not true

Code: Select all

  84.723 Error CrashHandler.cpp:189: Map tick at moment of crash: 314372

factorio-current.log: (13 KiB) Downloaded 147 times

Posted: **Tue Aug 13, 2019 3:07 pm**

[0.17.64]

: 74269-manual-reproduction.gif (493.63 KiB) Viewed 3803 times

-- edit:
issue looks to be around building pipe-to-ground ghost (may be similar to 74252), where unconnected pipe-to-ground pair is now broken: left one connects to ghost but right one connects(not exactly because of fluid mixing) to left one. This state is not stable save-loadable, because after save-load, right pipe-to-ground will disconnect.

This also produces some heavy-mode desync when right fluid system is changed somehow (part of gif where i build additional pipe above existing one). With heavy-mode enabled, this may crash after building pipe above existing pipe with water, without heavy-mode it will crash on removing pipe with water on assembler output.

log of heavy mode crashing game

-- edit:
Another crash when testing pipe-to-ground ghost:

log of crash in entity get selection box

-- edit:
And another:

crash in FluidBox notifyBlockedUndergroundConnectionsOfDestroy

-- edit:
And yet another one:

reproduction

crash in consistency

-- edit:
Factorio Pandora's Box:

reproduction

crash compatible connection is disconnected

Posted: **Tue Aug 13, 2019 6:26 pm**

Just LOL: new vanilla feature: underground connection from pipe-to-ground directly to pipe, and diagonal!

: 74269-dangling-pointer.gif (519.44 KiB) Viewed 3768 times

Most likely after removal of pipe-to-ground in center, right pipe-to-ground points to "related pipe-to-ground" that is already removed. New pipe that is placed when created gets same address in memory as removed pipe-to-ground and so connection is drawn diagonal to regular pipe

Posted: **Tue Aug 13, 2019 6:29 pm**

Can you please stop breaking the game

Posted: **Tue Aug 13, 2019 6:33 pm**

steinio wrote: ↑Tue Aug 13, 2019 6:29 pm Can you please stop breaking the game

I will when this bug report gets assigned to someone (@Dominik?) - then i will know devs admit it is a bug with "enough reproduction" (and laugh) (and huge spike on crash tracker)

Posted: **Thu Aug 15, 2019 1:19 pm**

Yes, this is very closely related to the 74252 but still a separate issue. I should have thought of these ghost things.

I was on a holiday. I come back and find this mayhem. Want to go back.

Posted: **Thu Aug 15, 2019 7:25 pm**

Dominik wrote: ↑Thu Aug 15, 2019 1:19 pm I was on a holiday. I come back and find this mayhem. Want to go back.

This whole fluid system rework has been way more fun than you expected, hasn't it?

Factorio Forums

[Dominik][0.17.63] FluidSystem.cpp:171: connection.isConnected() was not true

[Dominik][0.17.63] FluidSystem.cpp:171: connection.isConnected() was not true

Re: [0.17.63] FluidSystem.cpp:171: connection.isConnected() was not true

Re: [0.17.63] FluidSystem.cpp:171: connection.isConnected() was not true

Re: [0.17.63] FluidSystem.cpp:171: connection.isConnected() was not true

Re: [0.17.63] FluidSystem.cpp:171: connection.isConnected() was not true

Re: [Dominik][0.17.63] FluidSystem.cpp:171: connection.isConnected() was not true

Re: [Dominik][0.17.63] FluidSystem.cpp:171: connection.isConnected() was not true