[Rseding91] [0.16.30] Crash: "Surface::getAdvancedTile"

[Allaizn]

Hi everyone!

Since about 2-3 weeks ago I started getting random crashes on Factorio (which never happened before btw). Since I'm currently heavily working with cars, I suspect that its something with them, although I can't yet confirm to crash with no cars present. I hope the logfile and save will help, but please don't nerf cars (at least not yet ).
I'm running some mods, which should be their most current version, but I attached them anyway.

Have fun hunting bugs!

Edit: changed title version to 16.30 (logs for that version are in the later posts)

[Rseding91]

Thanks for the report. Do you have the minidump file that's created during crashing still? If so, can you please upload that?

Also, is the issue reproducible or does it happen randomly?

[Allaizn]

I tried reproducing, but sadly to no avail. It happened to me once again since posting, and I think the dump in the attachment is from the second time. I didn't even realize that the dump existed I added the log of the second crash as well.
It's somewhat mysterious to me what could cause them (I also have experience in debugging since I learned programming long ago). This new crash happened within 15sec of me loading the save attached in the first post. I simply started running to the top of the contraption, and then crashed while running .
The next crash will definitely have its dump saved, and thank you very much for looking into it!

[Allaizn]

And I just crashed again, this time while moving the camera through map view to place some circuit connections. I attached (the slighty further built) save, (the now indeed most recent versions of) all mods, and the dump as well as the log
While checking the times I noticed that the auto-save should have been very near: last auto-save file was written at 21:41, while the crash dump at 21:55. My auto-save interval is set at 10min, which was stretched by me hanging around <=40 FPS/ UPS. Maybe some access rights conflict? This would at least explain the random part about the crashes... I will try to let Factorio run on that map for a few auto-save cycles to test that hypothesis, and then report here.

[Allaizn]

Well, it doesn't seem to correlate with the auto-save to strongly. Setting the saving interval to 1min and then afking for about 30min didn't crash. But after setting it back up to 10 min and playing for a while I again crashed, this time immediately after an autosave. I simply ran upwards and removed the lower last stack inserter from each row (the unnecessary ones) to then crash after about 15 of them.

[Allaizn]

After doing other stuff for an hour, I decided to finish up for tonight with a little more cablework... Only to crash 1min after loading the autosave2 attached in my last post Log and Dump are attached, and I am deeply puzzled as to what may cause these crashes since I didn't do more than removing a few inserters and placing a blueprint twice.

[Allaizn]

I got a few more crashes and crash logs:
- First a crash with the same setup as before :

factorio-current.log

Log1

(10.02 KiB) Downloaded 145 times

&

factorio-dump-current.dmp

Dump1

(521.14 KiB) Downloaded 147 times

- After that I tried removing all unused mods by toggling all mods but creative mode and progressive running, to crash again :

factorio-current.log

Log2

(8.41 KiB) Downloaded 146 times

,

factorio-dump-current.dmp

Dump2

(514.43 KiB) Downloaded 153 times

as well as

mods.zip

Mods

(3.79 MiB) Downloaded 154 times

- I then kept on trying for good measure to crash again, this time while having the furnace array running:

factorio-current.log

Log3

(9.76 KiB) Downloaded 150 times

,

factorio-dump-current.dmp

Dump3

(514.56 KiB) Downloaded 156 times

and

SmelterTest2.zip

save

(4.24 MiB) Downloaded 155 times

Having done all this made me suspicious of the progressive running mod, since I mostly (if not everytime?) crashed while running. This would explain why I didn't crash while afking for the auto-save test, as well as the randomness part of the crashes. It would also fit the time frame of my crashes, since I only got that particular mod quite recently (with the mod file currently in my directory last "edited" on February the 22nd). My only guess would be that the mod either doesn't like the sheer amount of cars (I think its about 3.5k of them), or that it has problems with my 10-exoframes power armor made possible by the creative mode fusion reactor I disabled it for now and will report once I have played a few hours, either to confirm the cause, or with more crash reports
Happy weekend for everyone reading this!

[Allaizn]

A day later, and my suspicion is somewhat confirmed: I didn't crash anymore after removing progressive running from my mods list. I then tried adding the mod again, and crashed just now, again while running. This crash happened on an entirely new map, which at least confirms that the crashes are not map bound. I played for quite some time before crashing, and only about an half an hour ago started to place down some cars. I therefore think that its indeed something between progressive running and cars that freaks out the game. It's a shame though that I'm not able to figure out how to reproduce this behavoir
I'll definitely keep on playing without progressive running anymore, and will report here if our beloved game crashes once again I leave it up to you to decide whether or not you want to mark this thread as fixed, or try and find out how to prevent such crashes in the future (maybe some easy check allows for a definitive error message)

[Allaizn]

Well, blame on me, I crashed again Seems like progressive running was not at fault! I'm definitely confused now. Mods active are only creative mode and bottleneck, which are both quite popular (and it would therefore be quite the surprise to be the only one with crashes). My only other idea are the cars, but I don't know how to test that idea, since I *really* want to work with them .
Log, Dump and map, as well as the current mod folder are attached as always.

[Allaizn]

Progress Update: Progressive Running doesn't seem to be the culprit, since I crashed two more times. Once more with 0.16.28, and now once with 0.16.29. Should I edit the title to reflect the fact that the crash persisted into the newest version?
The first crash happened to me about 15 secs after I loaded the save, seemingly triggered by me right clicking onto the blank deconstruction planner in my upper left hotbar (I think slot 2 or 3?) to create its filter. I didn't see the filter menu before crashing.
The second crash happened while I was placing items into cars, by holding the left mouse button and running to the left (1 stack of iron ore, created by the creative mode requesting feature, per car). Once again I noticed that the auto-save should have been very near, since the last auto-save was at 18:31 and the crash was at 18:44 with an interval of 10 min and about 40 UPS.
Any ideas on what I could try in an effort to reproduce the crashing/ workaround? I have Visual Studio installed and have no problems following complex instructions, so feel free to order me around
I will get my new desktop PC on Saturday, which will replace my current laptop as my gaming rig, allowing me to test my builds on a far more powerful PC (i7 8700K & GTX 1080 Ti). I hope that the crashes won't happen there, but it would be nice to at least know why I keep getting them

[Allaizn]

Progress Update: The culprit seems to be either creative mode or tons of cars in vanilla.
The steps taken to reach this conclusion:
1. Toggle all mods but creative mode off
2. Create brand new map, size 150x150, grass only (everything else set to "None"), no ore, no enemys, no pollution
3. Pave everything with refined concrete for faster walking
4. Place one giant loop of blue belt
5. Place just shy of 10k cars on said belt
6. Simulate factory by using a few stack inserters on top feeding from and to infinity chests
7. Wait a while and crash (no movement necessary)
All this was done in 0.16.30, crash log, dump and map file are attached. I crashed once so far, but I will continue to stress test this thing, until either you or me finds out whats happening.
My current suspicion is that the inserters don't perfectly check for the inventory they're interacting with, leading to read and writes to invalid memory. That's purely speculation on my part though.
I also attached a crash on the smelter map for good measure (Happened shortly after I placed a belt so that cars started moving again)

[Allaizn]

It seems that my crashing is vanilla behavoir: Disabling all mods seemed to solve my problem, but I just crashed again. The auto-save was about 1 sec before, which leads me to believe that the saving doesn't correlate to the crashing, which in turn rules out causation.
Its not pure vanilla though, since the map was created when creative mode was installed, and therefore things like reach distance and no need for crafting materials didn't revert back after disabling the mod.
map, log and dump are attached as always, and I am as clueless as ever with regards to the source of these crashes
Edit: As to what I was doing as I crashed: I built a single furnace and some inserters feeding and emptying it to test some timings. After that I literally just stood there with my mouse not hovering over anything (I think) and observed the furnace while thinking what to do next, just to be interrupted by the crash notification.

[posila]

Hi, what kind of computer do you have? Based on your low-voltage model of CPU I assume you have a laptop. My bet is on overheating. Does the laptop have any fans? Can you install some software monitoring CPU temperature (for example http://openhardwaremonitor.org/)?

[Allaizn]

Sorry for the late answer, but stuff was happening IRL, causing me to forget about the bug report sorry for the delay.
I ultimately wasn't able to figure out what the core problem was, so overheating could be possible, though unlikely, since the laptop was comfortably sitting on my lap.
I now upgraded to a much more beastly desktop, and not once in all of my car development has factorio crashed again, so I'd say it's safe to assume that there was some kind of non obvious hardware problem, and hence case closed for this bug report.