View Issue Details

IDProjectCategoryView StatusLast Update
0009015Dwarf FortressDwarf Mode -- Jobs, Cancellation and Suspensionpublic2015-07-12 04:39
ReporterPatrikLundell Assigned To 
PrioritynormalSeverityminorReproducibilityhave not tried
Status newResolutionopen 
PlatformPCOSWindowsOS Version8.1
Product Version0.40.24 
Summary0009015: "Item lost or destroyed" removes repeat jobs
DescriptionAfter finding my web collection loom web collection repeat job missing several times I added 10 instances of the repeat job, and they slowly decrease in number.
I suspect the appearance of the "item lost or destroyed" message indicates the disappearance of a job, but that's just the trigger to check the number of remaining jobs.

I suspect the failure to perform the job because the web selected to pick up then gets destroyed by someone passing over it causes DF to incorrectly cancel the repeat job, rather than just the particular job instance.
Additional InformationIt ought to be quite easy to check the code to see if repeat jobs are removed rather than just the repeat job instance to confirm/reject the suspicion. Also, it is also quite possible the same logic is applied to other repeat workshop jobs, as the logic is probably the same throughout (I've had carpentry repeat jobs missing even though I thought I'd set them up, for instance).

I'm not using auto web collection since I want a dedicated web collection loom separated from the actual work loom (similar to clay and sand collection separated from actual work). Also, I don't want civilians to rush out into the caverns in an uncontrolled fashion.

I'm using the Lazy Newb Pack with Dwarf Therapist and Announcement Filter, plus DFHack Performance Tweaks and Pure Bugfixes.
TagsNo tags attached.

Activities

lethosor

2015-07-08 07:11

manager   ~0032859

Job cancellations are supposed to cancel repeating jobs as well (e.g. when materials are no longer available). The cancellations themselves could be related to 0001254, although that's supposed to be fixed - are there areas of the caverns that are unrevealed?

PatrikLundell

2015-07-12 04:39

reporter   ~0032887

Last edited: 2015-08-12 03:03

I agree that repeating jobs should be cancelled when out of job materials, but not when the individual item selected by DF is destroyed while there are hundreds of other legal items readily available. In that case only that particular instance of the job should be cancelled. This should result in the repeat job being moved to the end of the job queue as if the job had been completed successfully.
Earlier a totally separate part of the second cavern was undiscovered, but not when the observations in the bug report were made. I can not guarantee there are not any additional undiscovered totally separated areas, but I doubt it: the forum helped me find the undiscovered part after I'd observed a lack of cavern critters in the second cavern. Excessive tree grows blocking paths is not to blame either, since I have aggressively paved over most of the first cavern and removed all accessible trees in it to fight the tree FPS hit.
This situation is sort of a reverse of one I had in an earlier play, when I got spamming that buckets were lost or destroyed. Fire snakes destroyed my wooden buckets, but somehow DF had decided those particular buckets would be used, so the pond filling jobs were rescheduled and recancelled rapidly until I removed and recreated the pond designation(s).

Edit:
I've just experienced another variation of the theme. My mechanic has been ordered to produce quartzite mechanisms on repeat (using DFHack to select the material, but I strongly suspect the behavior would be the same without selection of a specific material) to use up the stones brought on embark. However, the wagon has been deconstructed, and the surf keeps pushing items around, so I've had the repeating job cancelled and removed twice so far, without having used up the stones.

Add Note

Note

Issue History

Date Modified Username Field Change
2015-07-08 01:46 PatrikLundell New Issue
2015-07-08 07:11 lethosor Note Added: 0032859
2015-07-12 04:39 PatrikLundell Note Added: 0032887
2015-08-12 03:03 PatrikLundell Note Edited: 0032887