0005053: Crash at the end of worldgen - Dwarf Fortress Bug Tracker

(0019507)
tasteful (reporter)
2012-02-14 05:49
edited on: 2012-02-14 05:53

I'm getting the same thing on the osx version, but creating a five year pocket went fine, it was just for a normal world that i got:

Dump: http://pastebin.com/0tTY0tQn [^]

(0019508)
Footkerchief (manager)
2012-02-14 05:55

Reminder sent to: tasteful

I moved your dump to pastebin so it wouldn't make this report impossible to scroll through.

However, what's more useful than a dump is world gen seeds that reproduce the crash. You can find the seeds in gamelog.txt in the main DF folder.

Also, are you using TrueType? What PRINT_MODE are you using in init.txt?

(0019509)
tasteful (reporter)
2012-02-14 05:58

Thanks.
Here are two failed seeds. http://pastebin.com/NJcHvygn [^]
PRINT_MODE is 2d

(0019565)
nCrazed (reporter)
2012-02-14 16:43

Had the game crashing with this error with any parameters.

Turned out that I extracted the new version into directory of an older install, extracting into a new directory fixed this

(0019585)
Footkerchief (manager)
2012-02-14 18:23

Seeds from 0005083:

*** STARTING NEW GAME ***
Generating world using parameter set CREATE WORLD NOW = 4333333
Seed: woUowokoWk0CqcQA62oY
History Seed: GEei28q4Oo2ASyoAIzn9
Name Seed: LujTdnUn2c72VONe4Nxm
Creature Seed: gjvf9ADHxqAl7I2avIFl

(0019586)
krenshala (reporter)
2012-02-14 18:52

I'm using the linux version, vanilla, 2D without TTF, and have not seen this problem (one world gen'd to year 125).

(0019601)
deekoo (reporter)
2012-02-14 22:14

I've seen this crash twice so far, on Debian AMD64 stable. My 0.34.1 install's in a new directory, not on top of an old one.

GDB backtrace is at:
http://deekoo.net/technocracy/laboratory/df_crashes.html [^]

The seeds were:
*** STARTING NEW GAME ***
Generating world using parameter set CREATE WORLD NOW = 4333333
Seed: 5XPoprJbnyVOl7oZJwA7
History Seed: fx97AbSIGrUg8fIlIR8v
Name Seed: ky5Htle2TatRLNOJicwJ
Creature Seed: HAnFfrql4HWc7PhqwBcm
*** STARTING NEW GAME ***
Generating world using parameter set CREATE WORLD NOW = 4333333
Seed: 6Ixw3E0NzuZ6PvuRXOau
History Seed: De4ZnqN9o6xiAiU1iIAX
Name Seed: bANov55Gs2yrYqEkUPh8
Creature Seed: HUJoybfUIGlhGZLzUnc8

(0019604)
dorf (reporter)
2012-02-14 23:11

On 64bit Win7 as well:
Generating world using parameter set CREATE WORLD NOW = 4333333
Seed: RNK2d8Oi4IFxPdL51Sd4
History Seed: hCar4ofuKfPEFCrAlvxK
Name Seed: pRKHHeV7BY9kypxFphFo
Creature Seed: mFI4wQQQYCiHCHQoMZ5q

(0019645)
oldmansutton (reporter)
2012-02-15 09:13

I had a world fail with:
the terminate called after throwing an instance of 'std::length_error'
what(): vector::_M_fill_insert

using the following seed:

*** STARTING NEW GAME ***
Generating world using parameter set CREATE WORLD NOW = 3233333
Seed: 0kKEMQS0ma4cE8EEYAWE
History Seed: OWSqEkKuKSi6ys6AEUwm
Name Seed: uwAuqKyEKGEiYcKYCQOI
Creature Seed: ewKSkGOkaGIG2I0eSuii

My next world didn't crash, but I aborted it in mid-generation and used it without the full history being generated. Seeds for that one were:

*** STARTING NEW GAME ***
Generating world using parameter set CREATE WORLD NOW = 3333333
Seed: 0kkSUIsQ8eFS4nYTsXy8
History Seed: xEaGTDV0ckmLyWsh2vmJ
Name Seed: rCXAnvRXGBqMDCSrcpf2
Creature Seed: kBDWXt3gzypiyAebvEHv

On a separate computer I have world gens complete successfully, however, they don't write out to disk. However, that may be a fault with my own file system permissions, still researching that one.

Both computers are running Ubuntu Linux 11.10 32bit. Both are running DF 34.01.

(0019646)
oldmansutton (reporter)
2012-02-15 09:25

First computer just completed worldgen successfully using the following:

*** STARTING NEW GAME ***
Generating world using parameter set CREATE WORLD NOW = 4144334
Seed: aeeWiCM2y6uuUuUya8wy
History Seed: QOQWqCkAIOIesm0AowyM
Name Seed: si68OO5HSw28iAO4t9aS
Creature Seed: r7TWu9KJMkk0WNb4KS7c

(0020045)
oldmansutton (reporter)
2012-02-18 09:47
edited on: 2012-02-18 09:52

Sometimes the worldgen completes successfully, but the world itself doesn't get saved. A temporary workaround is to abort the worldgen before the history completes, and then accept it as it is. Actually, I think that is the ONLY way I've gotten a world to save. This leads me to believe the crash is happening at the end of worldgen as it saves the world.

Perhaps the bug could be determined by comparing how the game saves the world at the end of worldgen and how it saves when aborting worldgen and accepting the world as it stands from that point?

(0020852)
Symmetry (reporter)
2012-02-28 09:48
edited on: 2012-02-28 11:47

I always abort in the current version as it's too slow, and I'm still 4/4 crashes on saving. So while that may be a factor, it's not the only cause.
Win7 here.

I'll try with your keys now.
Edit: your key on a medium world saved without crashing.

edit2: Just noticed the date on this. Let's find a better bug for it

Last edit: this is the new .03

(0021335)
Morrad (reporter)
2012-03-10 08:33
edited on: 2012-03-10 09:36

**
EDIT: I was able to fix this issue by removing all my previously saved games / generated worlds. I will leave the rest of my report up, but I have resolved this issue for me.
**

I am still having this happen in 0.34.05, regardless of input parameters as well. Happens on every world gen.

OS: Arch Linux

Terminal output: http://pastebin.com/aHzLv4ij [^]

gdb backtrace: http://pastebin.com/vtiFr9r1 [^]

libs: http://pastebin.com/DvAKYkYW [^]

(0021336)
Kwahn (reporter)
2012-03-10 10:12

I had this issue at first too. Did you put the new version into last year's version's folder? I did, and it caused the crash issues.

I put it in a new folder all it's own and it fixed the issues altogether.

(0021479)
Mazonas (reporter)
2012-03-15 03:49

Confirming another person with this crash issue. Oddly, my game was fine at first but seems to have developed this bug over time. Same whether using basic create new world or custom parameters.

Cannot use the workaround as above (ie, aborting world gen before history completes).

(0021486)
Deathworks (reporter)
2012-03-15 11:53

Dear Mazonas

I have the following suggestion:
(Download if necessary and) Unpack Dwarf Fortress again and overwrite your active version with the freshly unpacked (except for the data/init directory, unless you want extra work :) :) ). Your save files and so on should be fine.
I had a serious problem with save file corruption after trying to start new adventure mode after retiring an adventurer, and after that, I also had consistent world gen crashes in 34.05 until I overwrote the files with freshly unpacked ones.

(0021491)
Footkerchief (manager)
2012-03-15 14:43

Unpack Dwarf Fortress again and overwrite your active version with the freshly unpacked

This is always a bad idea.

(0021496)
Deathworks (reporter)
2012-03-16 00:20

Shouldn't overwriting the files of 34.05 with the files of 34.05 (that is, files of the same version) be safe? I honestly want to know, because in my eyes, it seems a good method of quickly repairing apparently broken binaries/raws.

(0022571)
hyperactiveChipmunk (reporter)
2012-05-17 09:31
edited on: 2012-05-17 09:32

[34.08] Removing previous saves fixes it. Perhaps something to do with the code that tries to determine the region number based on the current contents of the save directory?

(0022857)
Bordet (reporter)
2012-06-05 05:04

[34.10] Removing previous saves fixes it for me.

Issue History
Date Modified	Username	Field	Change
2012-02-14 05:41	Aelig	New Issue
2012-02-14 05:49	tasteful	Note Added: 0019507
2012-02-14 05:53	Footkerchief	Note Edited: 0019507	View Revisions
2012-02-14 05:55	Footkerchief	Issue Monitored: tasteful
2012-02-14 05:55	Footkerchief	Note Added: 0019508
2012-02-14 05:58	tasteful	Note Added: 0019509
2012-02-14 16:38	nCrazed	Issue Monitored: nCrazed
2012-02-14 16:43	nCrazed	Note Added: 0019565
2012-02-14 18:21	Footkerchief	Relationship added	has duplicate 0005073
2012-02-14 18:21	Footkerchief	Issue Monitored: rmric0
2012-02-14 18:22	Footkerchief	Relationship added	has duplicate 0005083
2012-02-14 18:23	Footkerchief	Note Added: 0019585
2012-02-14 18:52	krenshala	Note Added: 0019586
2012-02-14 22:14	deekoo	Note Added: 0019601
2012-02-14 23:11	dorf	Note Added: 0019604
2012-02-15 02:21	wRAR	Issue Monitored: wRAR
2012-02-15 07:09	Footkerchief	Relationship added	related to 0005099
2012-02-15 08:05	hyperactiveChipmunk	Issue Monitored: hyperactiveChipmunk
2012-02-15 09:13	oldmansutton	Note Added: 0019645
2012-02-15 09:25	oldmansutton	Note Added: 0019646
2012-02-15 14:26	Footkerchief	Relationship added	has duplicate 0005131
2012-02-16 07:05	Footkerchief	Relationship added	has duplicate 0005161
2012-02-16 07:05	Footkerchief	Issue Monitored: lightman
2012-02-17 05:22	Logical2u	Relationship added	has duplicate 0005212
2012-02-17 07:03	Footkerchief	Relationship deleted	has duplicate 0005212
2012-02-17 07:03	Footkerchief	Relationship added	has duplicate 0005206
2012-02-18 09:47	oldmansutton	Note Added: 0020045
2012-02-18 09:52	oldmansutton	Note Edited: 0020045	View Revisions
2012-02-18 10:15	Footkerchief	Relationship replaced	related to 0005161
2012-02-22 09:04	Footkerchief	Relationship added	related to 0005283
2012-02-28 09:30	Footkerchief	Tag Attached: Fixed in 0.34.03?
2012-02-28 09:48	Symmetry	Note Added: 0020852
2012-02-28 09:48	Symmetry	Issue Monitored: Symmetry
2012-02-28 09:52	Symmetry	Note Edited: 0020852	View Revisions
2012-02-28 10:01	Symmetry	Note Edited: 0020852	View Revisions
2012-02-28 11:47	Symmetry	Note Edited: 0020852	View Revisions
2012-03-06 09:31	Footkerchief	Relationship added	has duplicate 0003723
2012-03-10 08:33	Morrad	Note Added: 0021335
2012-03-10 08:33	Morrad	Note Edited: 0021335	View Revisions
2012-03-10 09:35	Morrad	Note Edited: 0021335	View Revisions
2012-03-10 09:36	Morrad	Note Edited: 0021335	View Revisions
2012-03-10 10:12	Kwahn	Note Added: 0021336
2012-03-15 03:49	Mazonas	Note Added: 0021479
2012-03-15 11:53	Deathworks	Note Added: 0021486
2012-03-15 14:43	Footkerchief	Note Added: 0021491
2012-03-16 00:20	Deathworks	Note Added: 0021496
2012-03-17 06:29	Footkerchief	Relationship deleted	related to 0005099
2012-05-17 09:31	hyperactiveChipmunk	Note Added: 0022571
2012-05-17 09:32	hyperactiveChipmunk	Note Edited: 0022571	View Revisions
2012-06-05 05:04	Bordet	Note Added: 0022857
2012-06-06 13:34	Footkerchief	Relationship deleted	related to 0005161
2014-01-21 19:48	Footkerchief	Relationship added	has duplicate 0006389
2014-01-21 19:49	Footkerchief	Relationship added	has duplicate 0006314
2014-01-21 19:50	Footkerchief	Relationship added	related to 0005161
2014-01-21 19:51	Footkerchief	Relationship added	related to 0005729
2014-01-26 08:44	Footkerchief	Relationship replaced	duplicate of 0005283
2014-01-26 08:44	Footkerchief	Status	new => resolved
2014-01-26 08:44	Footkerchief	Resolution	open => duplicate
2014-01-26 08:44	Footkerchief	Assigned To	=> Footkerchief
2014-01-26 08:46	Footkerchief	Tag Detached: Fixed in 0.34.03?
2014-07-09 18:12	lightman	Issue End Monitor: lightman

Notes
(0019507) tasteful (reporter) 2012-02-14 05:49 edited on: 2012-02-14 05:53	I'm getting the same thing on the osx version, but creating a five year pocket went fine, it was just for a normal world that i got: Dump: http://pastebin.com/0tTY0tQn [^]

(0019508) Footkerchief (manager) 2012-02-14 05:55	Reminder sent to: tasteful I moved your dump to pastebin so it wouldn't make this report impossible to scroll through. However, what's more useful than a dump is world gen seeds that reproduce the crash. You can find the seeds in gamelog.txt in the main DF folder. Also, are you using TrueType? What PRINT_MODE are you using in init.txt?

(0019509) tasteful (reporter) 2012-02-14 05:58	Thanks. Here are two failed seeds. http://pastebin.com/NJcHvygn [^] PRINT_MODE is 2d

(0019565) nCrazed (reporter) 2012-02-14 16:43	Had the game crashing with this error with any parameters. Turned out that I extracted the new version into directory of an older install, extracting into a new directory fixed this

(0019585) Footkerchief (manager) 2012-02-14 18:23	Seeds from 0005083: * STARTING NEW GAME * Generating world using parameter set CREATE WORLD NOW = 4333333 Seed: woUowokoWk0CqcQA62oY History Seed: GEei28q4Oo2ASyoAIzn9 Name Seed: LujTdnUn2c72VONe4Nxm Creature Seed: gjvf9ADHxqAl7I2avIFl

(0019586) krenshala (reporter) 2012-02-14 18:52	I'm using the linux version, vanilla, 2D without TTF, and have not seen this problem (one world gen'd to year 125).

(0019601) deekoo (reporter) 2012-02-14 22:14	I've seen this crash twice so far, on Debian AMD64 stable. My 0.34.1 install's in a new directory, not on top of an old one. GDB backtrace is at: http://deekoo.net/technocracy/laboratory/df_crashes.html [^] The seeds were: * STARTING NEW GAME * Generating world using parameter set CREATE WORLD NOW = 4333333 Seed: 5XPoprJbnyVOl7oZJwA7 History Seed: fx97AbSIGrUg8fIlIR8v Name Seed: ky5Htle2TatRLNOJicwJ Creature Seed: HAnFfrql4HWc7PhqwBcm * STARTING NEW GAME * Generating world using parameter set CREATE WORLD NOW = 4333333 Seed: 6Ixw3E0NzuZ6PvuRXOau History Seed: De4ZnqN9o6xiAiU1iIAX Name Seed: bANov55Gs2yrYqEkUPh8 Creature Seed: HUJoybfUIGlhGZLzUnc8

(0019604) dorf (reporter) 2012-02-14 23:11	On 64bit Win7 as well: Generating world using parameter set CREATE WORLD NOW = 4333333 Seed: RNK2d8Oi4IFxPdL51Sd4 History Seed: hCar4ofuKfPEFCrAlvxK Name Seed: pRKHHeV7BY9kypxFphFo Creature Seed: mFI4wQQQYCiHCHQoMZ5q

(0019645) oldmansutton (reporter) 2012-02-15 09:13	I had a world fail with: the terminate called after throwing an instance of 'std::length_error' what(): vector::_M_fill_insert using the following seed: * STARTING NEW GAME * Generating world using parameter set CREATE WORLD NOW = 3233333 Seed: 0kKEMQS0ma4cE8EEYAWE History Seed: OWSqEkKuKSi6ys6AEUwm Name Seed: uwAuqKyEKGEiYcKYCQOI Creature Seed: ewKSkGOkaGIG2I0eSuii My next world didn't crash, but I aborted it in mid-generation and used it without the full history being generated. Seeds for that one were: * STARTING NEW GAME * Generating world using parameter set CREATE WORLD NOW = 3333333 Seed: 0kkSUIsQ8eFS4nYTsXy8 History Seed: xEaGTDV0ckmLyWsh2vmJ Name Seed: rCXAnvRXGBqMDCSrcpf2 Creature Seed: kBDWXt3gzypiyAebvEHv On a separate computer I have world gens complete successfully, however, they don't write out to disk. However, that may be a fault with my own file system permissions, still researching that one. Both computers are running Ubuntu Linux 11.10 32bit. Both are running DF 34.01.

(0019646) oldmansutton (reporter) 2012-02-15 09:25	First computer just completed worldgen successfully using the following: * STARTING NEW GAME * Generating world using parameter set CREATE WORLD NOW = 4144334 Seed: aeeWiCM2y6uuUuUya8wy History Seed: QOQWqCkAIOIesm0AowyM Name Seed: si68OO5HSw28iAO4t9aS Creature Seed: r7TWu9KJMkk0WNb4KS7c

(0020045) oldmansutton (reporter) 2012-02-18 09:47 edited on: 2012-02-18 09:52	Sometimes the worldgen completes successfully, but the world itself doesn't get saved. A temporary workaround is to abort the worldgen before the history completes, and then accept it as it is. Actually, I think that is the ONLY way I've gotten a world to save. This leads me to believe the crash is happening at the end of worldgen as it saves the world. Perhaps the bug could be determined by comparing how the game saves the world at the end of worldgen and how it saves when aborting worldgen and accepting the world as it stands from that point?

(0020852) Symmetry (reporter) 2012-02-28 09:48 edited on: 2012-02-28 11:47	I always abort in the current version as it's too slow, and I'm still 4/4 crashes on saving. So while that may be a factor, it's not the only cause. Win7 here. I'll try with your keys now. Edit: your key on a medium world saved without crashing. edit2: Just noticed the date on this. Let's find a better bug for it Last edit: this is the new .03

(0021335) Morrad (reporter) 2012-03-10 08:33 edited on: 2012-03-10 09:36	EDIT: I was able to fix this issue by removing all my previously saved games / generated worlds. I will leave the rest of my report up, but I have resolved this issue for me. I am still having this happen in 0.34.05, regardless of input parameters as well. Happens on every world gen. OS: Arch Linux Terminal output: http://pastebin.com/aHzLv4ij [^] gdb backtrace: http://pastebin.com/vtiFr9r1 [^] libs: http://pastebin.com/DvAKYkYW [^]

(0021336) Kwahn (reporter) 2012-03-10 10:12	I had this issue at first too. Did you put the new version into last year's version's folder? I did, and it caused the crash issues. I put it in a new folder all it's own and it fixed the issues altogether.

(0021479) Mazonas (reporter) 2012-03-15 03:49	Confirming another person with this crash issue. Oddly, my game was fine at first but seems to have developed this bug over time. Same whether using basic create new world or custom parameters. Cannot use the workaround as above (ie, aborting world gen before history completes).

(0021486) Deathworks (reporter) 2012-03-15 11:53	Dear Mazonas I have the following suggestion: (Download if necessary and) Unpack Dwarf Fortress again and overwrite your active version with the freshly unpacked (except for the data/init directory, unless you want extra work :) :) ). Your save files and so on should be fine. I had a serious problem with save file corruption after trying to start new adventure mode after retiring an adventurer, and after that, I also had consistent world gen crashes in 34.05 until I overwrote the files with freshly unpacked ones.

(0021491) Footkerchief (manager) 2012-03-15 14:43	Unpack Dwarf Fortress again and overwrite your active version with the freshly unpacked This is always a bad idea.

(0021496) Deathworks (reporter) 2012-03-16 00:20	Shouldn't overwriting the files of 34.05 with the files of 34.05 (that is, files of the same version) be safe? I honestly want to know, because in my eyes, it seems a good method of quickly repairing apparently broken binaries/raws.

(0022571) hyperactiveChipmunk (reporter) 2012-05-17 09:31 edited on: 2012-05-17 09:32	[34.08] Removing previous saves fixes it. Perhaps something to do with the code that tries to determine the region number based on the current contents of the save directory?

(0022857) Bordet (reporter) 2012-06-05 05:04	[34.10] Removing previous saves fixes it for me.