0004287: Inexplicable, yet reproducible, crash

(0016938)
danaris (reporter)
2011-04-01 08:57

I have been able to reproduce this in version 0.31.25, and isolate the crash more precisely in time.

I have uploaded a save to http://dffd.wimbli.com/file.php?id=4107 [^] which should, if loaded in .25, crash within about 30 seconds with no user input.

Unfortunately, I have not been able to isolate the cause of the crash any better.

~ TGC

(0016969)
Dwarfu (manager)
2011-04-01 21:40

I'm not getting the crash on VISTA.

How did you move over your save to the newer versions?

What other changes have you made to your raw files?
Your ducks/geese are missing their BILL definition in body_default.txt.

What utility are you using to map out those unrevealed ore veins?

(0016970)
Dwarfu (manager)
2011-04-01 21:42

Reminder sent to: danaris

I posted a note instead of sending it to you, please check the bug report and respond when you are able.

(0016977)
danaris (reporter)
2011-04-02 06:11

- To move the save over, I had to make some compromises in the raw files, as a simple copy of the latest version's raws produced several errors. I'm afraid I don't recall exactly what I had to change; however, I was experiencing the crash before I ever transferred it to a version with different raws (as I said in the initial description, it started in .18 where the fort originated).

- I do not have access to Vista; the only Windows version I presently have access to is XP.

- I use dfhack (dfvdig and occasionally dfreveal) to designate the veins for digging. I have never had problems with it, either in other forts in earlier versions, or in the couple of forts I have started in more current versions (.23, .25).

(0016984)
Kumquat (reporter)
2011-04-02 10:57

Meddling with raws in a savegame makes this 0001271.

(0016987)
danaris (reporter)
2011-04-02 12:31

Yes, good job, you have completely misidentified the issue.

As I said twice, the crash appeared before I touched the raws. Additionally, none of the changes I made to make the new version's raws compatible with my save had anything to do with SHELL—that much, I do remember.

(0016988)
Footkerchief (manager)
2011-04-02 13:11
edited on: 2011-04-02 13:12

No need to get snippy. Do you have a version of the save that was never loaded with post-world-gen raws changes?

(0016996)
danaris (reporter)
2011-04-02 20:34

I believe so; however, I'll need some time to dig it up.

However, I can tell you for certain that the raws for that save would be the raws from the then-current version of Ironhand's graphics set, with no further modifications beyond what Ironhand does. I think that would have been 0.51...I'll see if I can figure it out Monday when I'm at the computer it would be on.

(0017015)
danaris (reporter)
2011-04-03 11:58

Haven't had chance to check the raws yet, but something occurred to me that may or may not be relevant.

This fort was able to obtain the King, but he showed up with no preferences. Up to this point, he has made no mandates or demands. It seems plausible to me that if the King decided to make a demand at this point, but had no preferences on which to base it, that could cause a crash.

It might be something totally different, but I figure it's worth mentioning.

(0017016)
Footkerchief (manager)
2011-04-03 12:34

That part is an old bug, 0003705. As far as anyone can tell, their lack of preferences causes them to simply not issue any mandates.

(0017021)
danaris (reporter)
2011-04-03 13:06

Ah, OK. Hadn't seen that one before.

(0017051)
danaris (reporter)
2011-04-04 06:01

I have tested it on the Mac OS X version, and it crashes the same as on Windows XP.

It does provide more crash info: http://topazgryphon.org/~tcollett/df/Ancientcrypts_crashes_osx.txt [^]

(0017085)
danaris (reporter)
2011-04-05 10:36

I must apologize most humbly to Kumquat. Having done some additional research, it appears that I had indeed meddled with the raws before I moved the save out of .18. Restoring the raws to their original state avoided the crash, at least so far.

I have moved the game to .25, and am testing it out there. This did, however, necessitate some changes to the raws once again (including removing the GRAZER tag, since there are no grasses upon which they can graze). I do not expect to see a repeat of this crash (though I did get a reproducible instance of the TTF crash bug...), given how things have gone so far.

(0017092)
Footkerchief (manager)
2011-04-05 11:41
edited on: 2011-04-05 11:41

Thanks for the update. Please reopen this if the crash shows up again.

Notes
(0016938) danaris (reporter) 2011-04-01 08:57	I have been able to reproduce this in version 0.31.25, and isolate the crash more precisely in time. I have uploaded a save to http://dffd.wimbli.com/file.php?id=4107 [^] which should, if loaded in .25, crash within about 30 seconds with no user input. Unfortunately, I have not been able to isolate the cause of the crash any better. ~ TGC

(0016969) Dwarfu (manager) 2011-04-01 21:40	I'm not getting the crash on VISTA. How did you move over your save to the newer versions? What other changes have you made to your raw files? Your ducks/geese are missing their BILL definition in body_default.txt. What utility are you using to map out those unrevealed ore veins?

(0016970) Dwarfu (manager) 2011-04-01 21:42	Reminder sent to: danaris I posted a note instead of sending it to you, please check the bug report and respond when you are able.

(0016977) danaris (reporter) 2011-04-02 06:11	- To move the save over, I had to make some compromises in the raw files, as a simple copy of the latest version's raws produced several errors. I'm afraid I don't recall exactly what I had to change; however, I was experiencing the crash before I ever transferred it to a version with different raws (as I said in the initial description, it started in .18 where the fort originated). - I do not have access to Vista; the only Windows version I presently have access to is XP. - I use dfhack (dfvdig and occasionally dfreveal) to designate the veins for digging. I have never had problems with it, either in other forts in earlier versions, or in the couple of forts I have started in more current versions (.23, .25).

(0016984) Kumquat (reporter) 2011-04-02 10:57	Meddling with raws in a savegame makes this 0001271.

(0016987) danaris (reporter) 2011-04-02 12:31	Yes, good job, you have completely misidentified the issue. As I said twice, the crash appeared before I touched the raws. Additionally, none of the changes I made to make the new version's raws compatible with my save had anything to do with SHELL—that much, I do remember.

(0016988) Footkerchief (manager) 2011-04-02 13:11 edited on: 2011-04-02 13:12	No need to get snippy. Do you have a version of the save that was never loaded with post-world-gen raws changes?

(0016996) danaris (reporter) 2011-04-02 20:34	I believe so; however, I'll need some time to dig it up. However, I can tell you for certain that the raws for that save would be the raws from the then-current version of Ironhand's graphics set, with no further modifications beyond what Ironhand does. I think that would have been 0.51...I'll see if I can figure it out Monday when I'm at the computer it would be on.

(0017015) danaris (reporter) 2011-04-03 11:58	Haven't had chance to check the raws yet, but something occurred to me that may or may not be relevant. This fort was able to obtain the King, but he showed up with no preferences. Up to this point, he has made no mandates or demands. It seems plausible to me that if the King decided to make a demand at this point, but had no preferences on which to base it, that could cause a crash. It might be something totally different, but I figure it's worth mentioning.

(0017016) Footkerchief (manager) 2011-04-03 12:34	That part is an old bug, 0003705. As far as anyone can tell, their lack of preferences causes them to simply not issue any mandates.

(0017021) danaris (reporter) 2011-04-03 13:06	Ah, OK. Hadn't seen that one before.

(0017051) danaris (reporter) 2011-04-04 06:01	I have tested it on the Mac OS X version, and it crashes the same as on Windows XP. It does provide more crash info: http://topazgryphon.org/~tcollett/df/Ancientcrypts_crashes_osx.txt [^]

(0017085) danaris (reporter) 2011-04-05 10:36	I must apologize most humbly to Kumquat. Having done some additional research, it appears that I had indeed meddled with the raws before I moved the save out of .18. Restoring the raws to their original state avoided the crash, at least so far. I have moved the game to .25, and am testing it out there. This did, however, necessitate some changes to the raws once again (including removing the GRAZER tag, since there are no grasses upon which they can graze). I do not expect to see a repeat of this crash (though I did get a reproducible instance of the TTF crash bug...), given how things have gone so far.

(0017092) Footkerchief (manager) 2011-04-05 11:41 edited on: 2011-04-05 11:41	Thanks for the update. Please reopen this if the crash shows up again.

Issue History
Date Modified	Username	Field	Change
2011-03-21 08:30	danaris	New Issue
2011-04-01 08:57	danaris	Note Added: 0016938
2011-04-01 08:57	danaris	Tag Attached: Save Included
2011-04-01 21:40	Dwarfu	Note Added: 0016969
2011-04-01 21:41	Dwarfu	Tag Attached: AWAITING UPDATE
2011-04-01 21:42	Dwarfu	Note Added: 0016970
2011-04-02 06:11	danaris	Note Added: 0016977
2011-04-02 10:57	Kumquat	Note Added: 0016984
2011-04-02 12:31	danaris	Note Added: 0016987
2011-04-02 13:11	Footkerchief	Note Added: 0016988
2011-04-02 13:12	Footkerchief	Note Edited: 0016988	View Revisions
2011-04-02 20:34	danaris	Note Added: 0016996
2011-04-03 11:58	danaris	Note Added: 0017015
2011-04-03 12:34	Footkerchief	Note Added: 0017016
2011-04-03 13:06	danaris	Note Added: 0017021
2011-04-04 06:01	danaris	Note Added: 0017051
2011-04-05 10:36	danaris	Note Added: 0017085
2011-04-05 11:41	Footkerchief	Note Added: 0017092
2011-04-05 11:41	Footkerchief	Status	new => resolved
2011-04-05 11:41	Footkerchief	Resolution	open => no change required
2011-04-05 11:41	Footkerchief	Assigned To	=> Footkerchief
2011-04-05 11:41	Footkerchief	Note Edited: 0017092	View Revisions
2011-04-05 11:41	Footkerchief	Relationship added	duplicate of 0001271
2011-04-06 05:24	Dwarfu	Tag Detached: AWAITING UPDATE