workflow: some resumed jobs are ignored by dwarves

[kimitsu]

There seem to be a situation for some jobs: a job that gets resumed (after item constraint check) is ignored by dwarfes - it shows up as resumed in workshop, but never gets active. Manual suspending/resuming does not resolve the issue, using "now!" or moving jobs around in the list does nothing too. Marking workshop for removal and stopping removal, after which jobs get restored in suspended state, and then resuming them manually doesn't seem to reliably solve the issue. Canceling repeat, cancelling the job and then adding the job back manually solves the problem for time being.
Example save (dfhack 0.40.23-r1): http://kimitsudesu.net/files/df/region1.zip
On level -4 there is a clothier's workshop with bunch of resumed jobs that are ignored even though most of the dwarfs have that labor active. Note that there is another clothier's workshop to the left of it. It had the same set of jobs and they were supposed to work in parallel, but in fact only one of them worked and the other got stuck. To isolate the issue I cleared all jobs from the one workshop that worked.

[acwatkins]

Also having this issue on dfhack develop branch, with 0.40.24, all over the place. Removing repeat, canceling, re-adding, and re-repeating the top stuck item seems to fix it temporarily for the entire list. Just canceling and letting repeat readd it doesn't work. Removing some other item doesn't fix the issue. So something seems to be up with the job that would be active itself.

I initially wondered if it was the newish priority code, somehow the unsuspended job has the lowest priority or something Disabling all other labors and hauling, leaving the dwarf with nothing to do doesn't help.

[melkor217]

Setting job field unk_v4020_1 to -1 fixes the job. Or causes game crash sometimes.

[acwatkins]

Odd since most items are -1 and that is its init value. I've investigated some, and the items that are stuck have different values for the unknown 40.20 field.

Below are two jobs that are stuck, note their odd unk_v4020_1 value.

Job 15876: MakeArmor (repeat)
material: any (cloth)
item: vest ()
Input Item 1: cloth; quantity=10000; min_dimension=10000
flags2: plant
unk_v4020_1: 6
Constraint ARMOR:ITEM_ARMOR_VEST amount 10 (gap 5)
items: amount 0; 0 stacks available, 10 in use.
Job 16490: MakeArmor (repeat)
material: any (yarn)
item: dress ()
Input Item 1: cloth; quantity=10000; min_dimension=10000
flags2: yarn
unk_v4020_1: 162

[acwatkins]

I've tested for a few days, it seems solid except for making bone bolts still getting occasionally stuck on me, not sure if it is related. Please test for stability, as setting to -1 in the past caused a crash. The if check to see if it needs to be suspended seems to have fixed that issue, though it would be good to get more people testing it. (Hopefully it isn't just less likely to occur)

Also note, existing stuck jobs wont be unblocked by this change. This change is only for newly resumed jobs. So, for existing stuck jobs, you must cancel them as before.

[melkor217]

Yaaay, such fix!

[argh226]

I do have the same issue with loom, leatherwork, smelter, etc.

Is there a fix for this? I do not understand the "unk_v4020_1 value"

:(

[melkor217]

@argh226, do you use master branch?

[lethosor]

The master branch is identical to the latest release; the develop branch contains that change, but you'll have to compile it from source for now.

[argh226]

No I'm using the one with df2014 with newbpack.

[lethosor]

It should be fixed in DFHack 0.40.24-r4 (presumably), but isn't fixed in the most recent DFHack release.
Just so you know, "the one with df2014 with newbpack" isn't enough for us to be able to tell what you're using - your platform and DFHack version (which is displayed when running help in the DFHack console) are more helpful.

[argh226]

I'm using newbpack which is 40.24-r3.
That would certainly mean that I'm 1 version behind :)

[lethosor]

r4 is the next version (i.e. not yet released) - you're using the latest version.

[argh226]

Btw, you guys did a really nice job with these hacks. They are extremely useful.
👍

[argh226]

Maybe its obvious, is there any release date for the R4 ?

[PeridexisErrant]

If 772ad03 fixes the problem, can the issue be closed?