Print

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

[leafar]

well i ran it with -g option:

Spoiler (click to show/hide)

Reading symbols from /home/rafael/DwarfFortress/MWDF-3a/Dwarf Fortress/libs/Dwarf_Fortress...(no debugging symbols found)...done.
(gdb) start
Temporärer Haltepunkt 1 at 0x804bdcc
Starting program: /home/rafael/DwarfFortress/MWDF-3a/Dwarf Fortress/libs/Dwarf_Fortress
ERROR: ld.so: object './hack/libdfhack.so' from LD_PRELOAD cannot be preloaded: ignored.
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".

Temporary breakpoint 1, 0xf75bdd13 in main () from /home/rafael/DwarfFortress/MWDF-3a/Dwarf Fortress/libs/libgraphics.so
(gdb) continue
Continuing.
Gtk-Message: Failed to load module "overlay-scrollbar"
[New Thread 0xf25b9b40 (LWP 5086)]
[New Thread 0xf1db8b40 (LWP 5087)]
Loading bindings from data/init/interface.txt
New window size: 1280x400
Font size: 16x16
Resizing grid to 80x25
Resizing font to 16x16

Resetting textures
[New Thread 0xf0b75b40 (LWP 5088)]
[New Thread 0xf01bfb40 (LWP 5089)]
[New Thread 0xef9beb40 (LWP 5090)]
[New Thread 0xeefffb40 (LWP 5091)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xf1db8b40 (LWP 5087)]
0xf7bfa3bd in DFHack::Core::doUpdate(DFHack::color_ostream&, bool) () from ./hack/libdfhack.so
(gdb) continue
Continuing.
[Thread 0xf25b9b40 (LWP 5086) exited]

Program received signal SIGSEGV, Segmentation fault.
0xf7bfa3bd in DFHack::Core::doUpdate(DFHack::color_ostream&, bool) () from ./hack/libdfhack.so
(gdb) continue
Continuing.
[Thread 0xef9beb40 (LWP 5090) exited]
[Thread 0xeefffb40 (LWP 5091) exited]
[Thread 0xf0b75b40 (LWP 5088) exited]
[Thread 0xf1db8b40 (LWP 5087) exited]
[Thread 0xf29ce880 (LWP 5082) exited]

Program terminated with signal SIGSEGV, Segmentation fault.
The program no longer exists.

not quite sure what this tells me, though

Out of curiosity, what hex editor do you use?

i normally use GHex, but for this i did write a python script

[ag]

Code: [Select]

0xf7bfa3bd in DFHack::Core::doUpdate(DFHack::color_ostream&, bool) () from ./hack/libdfhack.so
not quite sure what this tells me, though

Well, it says that it crashes somewhere here:

https://github.com/peterix/dfhack/blob/master/library/Core.cpp#L1149

Which is rather hard to explain unless perhaps you added that md5 tag to the data for the wrong version of df - there are several in symbols.xml.

[lethosor]

I tried editing it again and didn't get a crash... weird.
Those errors definitely seem to be a dfhack problem, like ag pointed out, rather than a problem with the executable. Maybe reinstalling dfhack would help, but I can't be certain

On a related note, have you tried running this specific dfhack installation with the vanilla executable? If that fails as well, it's likely a dfhack-related problem and not a problem with your patch.

[leafar]

Which is rather hard to explain unless perhaps you added that md5 tag to the data for the wrong version of df - there are several in symbols.xml.

that was it! thank you!
i didn't notice there were different entries...

[ORCACommander]

does anyone have a problem with revflood crashing dwarf fortress? i tried to use it to reveal the complete lay out of my caverns on a 5x5 embark using ironhand graphics. Only thing i can think of is it is somehow triggering the ram limit.

[Hommit]

hmm... is plugin for sorting gem list by material value possible?

Look at the gem stockpile list. Then open up inorganic_stone_gem.txt. What you should notice is that the list is arranged in the exact same order that they are in the file. Just shuffle them around to your liking and voilà.

so, i did just that
but for whatever reason it completely wrecked worldgen.
no ores and aquifers at all in whole world, and when test embark, some tiles is flashing. on mouse hover, they show very fast cycling of all gems
file: http://rghost.ru/private/46189159/5c61bfde8908d607643faadc228b5270
any idea wtf?
(probably belongs to modding, but...)
up: hm, if backup file was in the same folder, will it conflict with original one?

[drayath]

Yes, duplicate raws (file duplicates or single entry duplicates) are the modders version of !FUN! 

[ash88]

tl;dr problem: Queen isn't a dwarf - anything I can do?

Hey guys - I'm bummed out because I made my best fortress ever, and the queen showed up, and it is a drow elf.  I'm using masterworkDF.  Anyways, as you know I can't assign them rooms if they aren't dwarves (bug #4611 - in vanilla it's the same thing with Goblins ), and they don't show up in Dwarf Therapist... etc.  They are just running around outside... fortunately the siege-ers haven't found them yet...

Is there some way to change the race of a unit?  If I did this would my noble become room-assignable and DwarfTherapist viewable?  I'm willing to try something messy and extreme.

I'm hoping I don't have to abandon for such a un-satisfying reason.  This fortress really is my best ever.  Might have to start randomly digging down and find a FUN-NER conclusion to Mountainhome Gloombitch...

Cheers.

[ORCACommander]

there is an lua script that can designate monarchs:

hack\scripts\make-monarch.lua

I think to work it just hit k select a dorf and run the script

[Rumrusher]

tl;dr problem: Queen isn't a dwarf - anything I can do?

Hey guys - I'm bummed out because I made my best fortress ever, and the queen showed up, and it is a drow elf.  I'm using masterworkDF.  Anyways, as you know I can't assign them rooms if they aren't dwarves (bug #4611 - in vanilla it's the same thing with Goblins ), and they don't show up in Dwarf Therapist... etc.  They are just running around outside... fortunately the siege-ers haven't found them yet...

Is there some way to change the race of a unit?  If I did this would my noble become room-assignable and DwarfTherapist viewable?  I'm willing to try something messy and extreme.

I'm hoping I don't have to abandon for such a un-satisfying reason.  This fortress really is my best ever.  Might have to start randomly digging down and find a FUN-NER conclusion to Mountainhome Gloombitch...

Cheers.

you got yourself a DROW Cacame.

[Meph]

Reading all this about adding caravan guards to your fort, and changing the race of nobles got me thinking:

Animal men can learn.
Animal men can be bought in caravans.
Animal men can't do labor... could this be fixed?

It would be nice if races could buy other minor races (goblin buy trolls for example, or undeads buy skeletons) that can do labor in the fort. Any sort of automated script possible, to make them civ-members?

[Warmist]

Reading all this about adding caravan guards to your fort, and changing the race of nobles got me thinking:

Animal men can learn.
Animal men can be bought in caravans.
Animal men can't do labor... could this be fixed?

It would be nice if races could buy other minor races (goblin buy trolls for example, or undeads buy skeletons) that can do labor in the fort. Any sort of automated script possible, to make them civ-members?

There are two ways to "fix" them not doing labors: 1) -not recommended- dfusion friendship. This hacks the code to fool df into thinking that your main race is multiple races and thus citizens can be anything. 2) start a script that assigns jobs (or replaces dwarf jos) to animal men. Example here: https://gist.github.com/warmist/5636653 Best used with onReactionComplete + custom reaction (e.g. immigration office?). Note: script is not perfect. It does not check burrows and general reachability, but the general idea is simple.

[Tierre]

what is the essential list of commands you always start then playing? In init and manually? I add
binpatch armorstand-capacity
binpatch custom-reagent-size
binpatch deconstruct-heapfall
binpatch deconstruct-teleport
binpatch hospital-overstocking
binpatch training-ammo
binpatch weaponrack-unassign

keybinding add Alt-V digv
keybinding add Alt-O "liquids-here"
keybinding add Alt-Q "changeitem q 5"
keybinding add Alt-C "clean all"
keybinding add Alt-S "sort-units"
to Init file and use

fixmerchants
fixdiplomats

on embark.

I was thinking of making a script to turn on after loading a game with all the tweak fixes and all but i wanted input on those commands as i am afraid i don;t know them all.

[peterix]

I've got a dfhack-internals question for peterix or anyone else who knows.

The sample plugin skeleton.cpp has a comment that dfhack console commands are called from a different thread than the main DF thread.

Is this only true of the console commands?  Or do plugin_* run in a different thread?

I assume that vtable hooks must run in DF's main thread, at least.


(I know, I know, I should write test code and figure it out myself.  DFHack is dauntingly complex, and I'm trying to get a handle on how it all fits together.)

I guess you already found out yourself, but I'll reply anyway, just in case

There are several threads in DF with DFHack installed. The main simulation thread, the graphics thread, DFHack console thread and DFHack hotkey thread (just waits for hotkey events). A plugin can potentially have code that runs in any and all of them.

Commands need to acquire a lock to work with the game data, which is unlocked every time the game updates (see CoreSuspender and Core::Suspend). The lock has a queue/stack of condition variables that keeps track of all the plugins/threads which requested a suspend. Those are woken up as soon as the simulation thread is in a safe spot (onUpdate()).

Plugins can have callback functions that get called when a particular event happens/is detected. This is called from onUpdate and in turn, from the simulation thread.

So, you could easily have any number of threads and synchronize with the game by using Core::Suspend or the much nicer, CoreSuspender object

[ag]

Plugins can have callback functions that get called when a particular event happens/is detected. This is called from onUpdate and in turn, from the simulation thread.

So, you could easily have any number of threads and synchronize with the game by using Core::Suspend or the much nicer, CoreSuspender object

One additional detail related to vmethod hooks: Normally the CoreSuspender lock is recursive, i.e. suspending again from a nested function call is safe. This applies to methods like plugin_onupdate, etc. However, since vmethod hooks are usually called directly by DF code, they implicitly have full right to access game data because of that, but the bookkeeping for the lock isn't properly updated, and CoreSuspender would deadlock. Therefore, if a vmethod hook needs to call something that uses CoreSuspender, it should use a special CoreSuspendClaimer object, which updates the bookkeeping info to officially declare that implicit lock ownership.