Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[DizzyCrash]

The reason items don't disappear immediately when you destroy them is that the command marks them for destruction, but they're not actually removed until DF resumes. This, in turn, is because you can't just disintegrate items, but have to clean them out properly so ownership, container contents/holder, etc. are removed without leaving references to objects that just vanished.

And the reason you have to unpause is because DFHack doesn't actually know how to do all of that cleanup, so it just sets a special flag and tells Dwarf Fortress itself to do that cleanup, and that only happens when the game is unpaused (once every 100 frames, if I recall correctly).

Can you exile dwarves?

Why do you want to do that?

Cause throwing them off bridges is overloading me with tombstones.

[Rose]

You will be able to exile them in the next version.

In the mean time, you can send them solo to attack the goblin capitol.

[DizzyCrash]

Oh? How do i do that?

[DizzyCrash]

I am just extremely particular who joins my civilization and gets remembered in our halls of heroes. I have spartan levels of high expectations... Except that instead of looking for strong capable warriors i am looking for unique and interesting dwarves with personalities and appearances that i do not currently have.

[Sorg]

Oh? How do i do that?

Just send undesirable dwarves on a mission. You don't need dfhack for that.
http://dwarffortresswiki.org/index.php/DF2014:Mission

[DizzyCrash]

Oh? How do i do that?

Just send undesirable dwarves on a mission. You don't need dfhack for that.
http://dwarffortresswiki.org/index.php/DF2014:Mission

And the children go off the cliff?

[Rose]

Ah yes, sadly, not much you can do with the children besides day care.

[Clément]

unit_soul's unk_4410_1 and unk_4410_2 are focus related. They could be unit_personality's current_focus and undistracted_focus which have moved two field further. I am not sure since I don't really know how they supposed to work. Maybe strainer can confirm.

PS: unit_soul.unk_4410_2 is initially 1 (just after embark) but as soon as a need is changed it is set to 4 times the number of needs, so it really looks like it.

[PatrikLundell]

@DizzyCrash:
You might want to try an old DFHack script called emigration which is intended to make unhappy dorfs leave on their own accord. The version the current DFHack release is broken, but this version ought to work (functionally the same as the one in the next DFHack release):

Spoiler (click to show/hide)

Code: [Select]

[code]
--Allow stressed dwarves to emigrate from the fortress
-- For 34.11 by IndigoFenix; update and cleanup by PeridexisErrant
-- old version:  http://dffd.bay12games.com/file.php?id=8404
--[====[

emigration
==========
Allows dwarves to emigrate from the fortress when stressed,
in proportion to how badly stressed they are and adjusted
for who they would have to leave with - a dwarven merchant
being more attractive than leaving alone (or with an elf).
The check is made monthly.

A happy dwarf (ie with negative stress) will never emigrate.

Usage:  ``emigration enable|disable``

]====]

local args = {...}
if args[1] == "enable" then
    enabled = true
elseif args[1] == "disable" then
    enabled = false
end

function desireToStay(unit,method,civ_id)
    -- on a percentage scale
    local value = 100 - unit.status.current_soul.personality.stress_level / 5000
    if method == 'merchant' or method == 'diplomat' then
        if civ_id ~= unit.civ_id then value = value*2 end end
    if method == 'wild' then
        value = value*5 end
    return value
end

function desert(u,method,civ)
    u.following = nil
    local line = dfhack.TranslateName(dfhack.units.getVisibleName(u)) .. " has "
    if method == 'merchant' then
        line = line.."joined the merchants"
        u.flags1.merchant = true
        u.civ_id = civ
    elseif method == 'diplomat' then
        line = line.."followed the diplomat"
        u.flags1.diplomat = true
        u.civ_id = civ
    else
        line = line.."abandoned the settlement in search of a better life."
        u.civ_id = -1
        u.flags1.forest = true
        u.animal.leave_countdown = 2
    end
    print(line)
    dfhack.gui.showAnnouncement(line, COLOR_WHITE)
end

function canLeave(unit)
    if not unit.status.current_soul then
        return false
    end
   
    for _, skill in pairs(unit.status.current_soul.skills) do
        if skill.rating > 14 then return false end
    end
   
    return dfhack.units.isOwnRace(unit) and  --  Doubtful check. naturalized citizens
           dfhack.units.isOwnCiv(unit) and   --  might also want to leave.
           dfhack.units.isActive(unit) and
           not dfhack.units.isOpposedToLife(unit) and
           not unit.flags1.merchant and
           not unit.flags1.diplomat and
           not unit.flags1.chained and
           dfhack.units.getNoblePositions(unit) == nil and
           unit.military.squad_id == -1 and
           dfhack.units.isCitizen(unit) and
           dfhack.units.isSane(unit) and
           not dfhack.units.isBaby(unit) and
           not dfhack.units.isChild(unit)
end

function checkForDeserters(method,civ_id)
    local allUnits = df.global.world.units.active
    for i=#allUnits-1,0,-1 do   -- search list in reverse
        local u = allUnits[i]
        if canLeave(u) and math.random(100) < desireToStay(u,method,civ_id) then
            desert(u,method,civ_id)
        end
    end
end

function checkmigrationnow()
    local merchant_civ_ids = {}
    local diplomat_civ_ids = {}
    local allUnits = df.global.world.units.active
    for i=0, #allUnits-1 do
        local unit = allUnits[i]
        if dfhack.units.isSane(unit)
        and not unit.flags1.dead
        and not dfhack.units.isOpposedToLife(unit)
        and not unit.flags1.tame
        then
            if unit.flags1.merchant then table.insert(merchant_civ_ids, unit.civ_id) end
            if unit.flags1.diplomat then table.insert(diplomat_civ_ids, unit.civ_id) end
        end
    end

    for _, civ_id in pairs(merchant_civ_ids) do checkForDeserters('merchant', civ_id) end
    for _, civ_id in pairs(diplomat_civ_ids) do checkForDeserters('diplomat', civ_id) end
    checkForDeserters('wild', -1)
end

local function event_loop()
    if enabled then
        checkmigrationnow()
        dfhack.timeout(1, 'months', event_loop)
    end
end

dfhack.onStateChange.loadEmigration = function(code)
    if code==SC_MAP_LOADED then
        if enabled then
            print("Emigration enabled.")
            event_loop()
        else
            print("Emigration disabled.")
        end
    end
end

if dfhack.isMapLoaded() then
    dfhack.onStateChange.loadEmigration(SC_MAP_LOADED)
end
[/code]

It ought to be possible to just use the leaving part of the logic to kick out targeted dorfs.

Edit: The script logic was screwed up (again). Should be fixed now.

[Rose]

Alternately wait till next week, when you'll be able to just flat out tell dwarfs to get lost. Sadly won't work on nobles, which would be the prime use case for such a feature.

[strainer]

unit_soul's unk_4410_1 and unk_4410_2 are focus related. They could be unit_personality's current_focus and undistracted_focus which have moved two field further. I am not sure since I don't really know how they supposed to work. Maybe strainer can confirm.

Thanks for working this out Clément. When I use ..soul->unk_4410_1 as current_focus and ..soul->unk_4410_2 as undistracted_focus my previous rating system works just like before. After embark their working range seems to be around 75 to 125 each. Ill copy this table you posted in Therapists thread for the record here:

""""
The current_focus thresholds for "Overall, <name> is [...] with satisfied needs/by unmet needs, are relative to undistracted_focus:

• very focused: 140% and higher
• quite focused: 120% - 140%
• somewhat focused: 100% - 120%
• untroubled: exactly undistracted_focus
• unfocused: 81% - 100%
• distracted: 61% - 81%
• badly distracted: 61% and below

This is based on only two different values for undistracted_focus.
""""

I also compared the colors and phrases used to describe the focus changes, looking for some difference in 'current_focus' and 'undistracted_focus' changes, but nothing stands out.
"he is (this) after..."

• unfettered
• level-headed
• untroubled
• not distracted
• unfocused
• distracted
• badly distracted

edit - I see there, the yellow is worse than brown ^^

[Clément]

Is there a script or program that would format the data from df-structures so it can be imported in IDA?

[lethosor]

https://github.com/jjyg/df_misc/blob/master/codegen_c_hdr.pl
https://github.com/jjyg/df_misc/blob/master/dump_df_globals.rb also supports an IDC format

[Clément]

Thanks. I fear then generated codegen.h may have some issues with gcc builds. Generated structure for derived classes use

Code: [Select]

struct derived {
    struct base super;
    /* ... */
}; Although, this is usually how you do inheritance in C, it may not use the exact same layout as C++ classes. When deriving non-POD types, GCC optimize for space and and does not count the padding at the end of the base class. Among offsets used by DT, this affects history_event_hist_figure_diedst and item_crafted.

It should work fine with MSVC, where, IIRC, the tail padding is kept.

[mifki]

Thanks. I fear then generated codegen.h may have some issues with gcc builds. Generated structure for derived classes use

Code: [Select]

struct derived {
    struct base super;
    /* ... */
}; Although, this is usually how you do inheritance in C, it may not use the exact same layout as C++ classes. When deriving non-POD types, GCC optimize for space and and does not count the padding at the end of the base class. Among offsets used by DT, this affects history_event_hist_figure_diedst and item_crafted.

It should work fine with MSVC, where, IIRC, the tail padding is kept.

Yes. When I was recently using codegen_c_hdr to analyse osx binary, I modified it to just output all the fields from base class directly, and then all the layouts were correct.