CTD after hyperjump

[dhouseholder]

Hello, the attached save file consistently produces a CTD when hyperjumping. Simply load the save, unpause the game, and click hyperjump to jump to the preselected system. Or try another system. The game will crash within 10 seconds after arriving. I have tested this savefile on 20230203 on Windows and Steamdeck.

vert.zip

[dhouseholder]

output.txt
including output.txt

[impaktor]

Info: Created shader geosphere_star (address=0x55556392ece0)

Thread 1 "pioneer" received signal SIGSEGV, Segmentation fault.
Frame::GetPositionRelTo (this=this@entry=0x55fc934b1570, relToId=...) at /home/me/usr/src/pioneer/src/Frame.cpp:352
352		if (m_thisId == relToId) return vector3d(0, 0, 0);
(gdb) bt
#0  Frame::GetPositionRelTo (this=this@entry=0x55fc934b1570, relToId=...) at /home/me/usr/src/pioneer/src/Frame.cpp:352
#1  0x000055555577dc3f in Body::GetPositionRelTo (this=0x555560398d10, relToId=...) at /home/me/usr/src/pioneer/src/Body.cpp:205
#2  0x000055555577de63 in Body::GetPositionRelTo (this=<optimized out>, relTo=<optimized out>) at /home/me/usr/src/pioneer/src/Body.cpp:221
#3  0x000055555567b645 in AICmdFlyAround::TimeStepUpdate (this=0x555560d66ea0) at /home/me/usr/src/pioneer/src/ShipAICmd.cpp:1601
#4  0x00005555557ecb62 in Ship::AITimeStep (this=0x555560dbe0d0, timeStep=timeStep@entry=0.0166666675) at /home/me/usr/src/pioneer/src/Ship-AI.cpp:37
#5  0x00005555557612e2 in ShipController::StaticUpdate (this=0x555560df0dc0, timeStep=0.0166666675) at /home/me/usr/src/pioneer/src/ship/ShipController.cpp:11
#6  0x000055555566fff7 in Ship::StaticUpdate (this=0x555560dbe0d0, timeStep=0.0166666675) at /home/me/usr/src/pioneer/src/Ship.cpp:1215
#7  0x000055555568f5b3 in Space::TimeStep (this=0x555560d0e3e0, step=step@entry=0.0166666675) at /home/me/usr/src/pioneer/src/Space.cpp:1042
#8  0x00005555557a4669 in Game::TimeStep (this=0x55555d898e80, step=0.0166666675) at /usr/include/c++/12.2.1/bits/unique_ptr.h:191
#9  0x000055555563e812 in GameLoop::Update (this=0x555555be56c0, deltaTime=0.0165201146) at /home/me/usr/src/pioneer/src/Pi.cpp:944
#10 0x00005555555f3a87 in Application::Run (this=0x555555bcec00) at /home/me/usr/src/pioneer/src/core/Application.cpp:192
#11 0x00005555555d5102 in main (argc=<optimized out>, argv=<optimized out>) at /home/me/usr/src/pioneer/src/Pi.h:145

[dhouseholder]

is there something wrong with the cargo capacity used/available on my ship?
When I first started the game I sold some equipment off for more trade cash. Not sure if that could be related.

[bszlrd]

No, that's intentional. The ship hase more overall capacity (equipment+cargo) than cargo hold. Simplistic way to differentiate ships. But that's subject to change in the future to a more elaborate way of handling

[dhouseholder]

OK, it's just very confusing. I have 14 tons battle weapons and 2 tons hydrogen for a total of 16 tons of cargo. The ship info page says I have 8 tons cargo capacity and 0 tons used.

[bszlrd]

That's indeed odd. Could you open a separate issue for it?

[dhouseholder]

@nozmajner opened as #5557

[Gliese852]

Perfectly reproducible, very nice! Including in the current master.
An attempt is made to get an object with an index of ~500000 from a static array of size ~100, which, of course, leads to an access to someone else's memory.

[Gliese852]

This is an interesting case - the NPC ship jumped into the same system as you, at the time of your jump it was already in hyperspace. Therefore, it appeared in the new system, but in its autopilot was a pointer to the body from the previous system. This will eventually crash the game.

[Gliese852]

The fix turned out to be quite simple.
@dhouseholder In your save, there are several dozen ships in hyperclouds on exit from the system, with "bad" autopilots.
In the fixed version of the game, if you wait ~4 days, they will all disappear, and the new ships will have clean autopilots, and will not cause a crash if you jump after them. Or you can immediately jump into e.g. Gatildress, it seems no one is going there, and the game does not crash.

[zonkmachine]

Duplicates?
#5242
#5258

[Gliese852]

Indeed, these too.