Print

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

[Clément]

No it is hard-coded. You can modify the values here and recompile. I am hesitating between doing the quick patch and moving the threshold (and strings) in game_data.ini (it would allow to keep DT working with older DF versions).

[Saiko Kila]

The limits for stress are:
10000 (and above) stressed
25000 (and above) haggard
50000  (and above) harrowed
Negatives don't have particular checks, and thus associated descriptions, as far as I can see. The max is 100000 and the min is -100000 (DF uses simple max() and min() routines to keep them in these margins).

[strainer]

Good to know the bounds for sure. I also ran the save linked earlier and got a few stressed to 1000k with maniacal mistreatment. The red arrows appear solidly at 10k.

The 'Focus' counters have changed in 44.10 I dont know if DT is interested in that.

[Clément]

DT does not currently show the focus, but I am interested in what you know. I think I understand the needs but I don't understand the global focus values.

https://github.com/DFHack/df-structures/blob/c3025feb80c6f8e7441ea5dcf4f463a9cf89cbbd/df.units.xml#L1817-L1818

Code: [Select]

<int32_t name="current_focus" comment='weighted sum of needs focus_level-s'/>
<int32_t name="undistracted_focus" comment='usually number of needs multiplied by 4'/>
If current_focus is a weighted sum, I don't know what are the weights (not the need_levels apparently) and undistracted_focus is not the number of needs multiplied by 4. DFHack's comments may need to be updated.

[Clément]

New version released: 40.0.1

Minor bug fixes release.

Changelog:

• added "see dead body" thought
• fixed happiness thresholds for 0.44.10 (thresholds can now be configured in game_data.ini for older versions)

This release fixes the happiness thresholds for 0.44.10 but break them for older versions. When playing with DF 0.44.09 or older, you need to add a modified game_data.ini in DT's data folder (use "Open data directory" in the "File" menu, if you don't know where it is). Replace the happiness_levels section with (add a zero in each threshold):

Code: [Select]

[happiness_levels]
miserable/threshold = 500000
miserable/desc = "is utterly harrowed by the nightmare that is their tragic life"
very_unhappy/trhreshold = 250000
very_unhappy/desc = "is haggard and drawn due to the tremendous stresses placed on them"
unhappy/threshold = 100000
unhappy/desc = "is under a great deal of stress"
fine/threshold = -100000
content/threshold = -250000
happy/threshold = -500000
ecstatic/threshold = -1000000Remember to delete your modified game_data.ini, when you want to go back to the latest version.

Windows builds are also available on DFFD (win32, win64).

PS: (edit: fixed now) macOS deployment failed for unknown reason although it was not changed since the last release. Is there a macOS expert that can check the deploy script?

Code: [Select]

$ .travis/$TRAVIS_OS_NAME/deploy.sh
Reading whole disk (Apple_HFS : 0)…
Elapsed Time:  39.304ms
Speed: 40.8Mbytes/sec
Savings: 95.9%
created: /Users/travis/build/Dwarf-Therapist/Dwarf-Therapist/DwarfTherapist.sparseimage
/dev/disk1                                          /Volumes/Dwarf Therapist
hdiutil: couldn't unmount "disk1" - Resource busy
hdiutil: convert failed - Resource temporarily unavailable
The command ".travis/$TRAVIS_OS_NAME/deploy.sh" failed and exited with 1 during .

[strainer]

Just a comment typo? "Haggard and dr(a)wn"

[lethosor]

Is it possible that something else was writing to the disk image? If not, maybe macdeployqt finished before the disk write was actually finished. I would just add a retry loop of some kind around the "hdiutil eject" step and see if that helps.

[Clément]

OSX deployment fixed itself when I restarted the job today. I don't know if a retry loop will help. I still added it just in case.

Sorry for the typo, but with osx deployment fixed, I don't think worth making a new release right now. Your dwarves will be drowning in unhappiness for a little longer.

[strainer]

DT does not currently show the focus, but I am interested in what you know.

In 44.10 the old counters are both set to zero, Im guessing the structure finder has just lost them as the description text for focus and distraction doesn't seem to have changed.

`current_focus` seemed to track focus level (higher=more focused) and `undistracted_focus` seemed to track "current_distraction" (higher=more distracted). Keeper displayed a 1 digit signed score based on current_focus-current_distraction that indicated the jist of the description text for units focus and distraction. When I made the score I thought they were components of the same 'focus' thing, but maybe focus affects job quality, and distraction affects punctuality.

A signed percentage might work for happiness/gladness/unstress now that its bounds are known, then percentage for focus and undistraction could be shown along side. I never resolved the proper bounds of current_focus and undistracted_focus|current_distraction as I calibrated adhoc the combined logarithmic score and it worked alright, they were around 0 to 100 each I think.

[Clément]

They seem to have moved (unit_soul.unk_4410_1 and unit_soul.unk_4410_2 in DFHack 0.44.10-r1).

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.

[Clément]

Here is how need columns could look like:


The number is the need level (if I understand correctly, the speed at which focus goes down), the color is the need focus. There is an issue with the "Pray/Meditate" column: a dwarf can have several of these kind of need for different deities. I am currently displaying the sum of need levels and the worst focus (for example "Kol Athelineth" is actually 10 + 5).

[Clément]

Experimental memory layouts for 0.44.11:

• Linux 32 bits
• Linux 64 bits
• OSX 32 bits
• OSX 64 bits
• Windows 32 bits
• Windows 64 bits

Edit: They should be working now.

[hagger]

Experimental memory layouts for 0.44.11:

• Linux 32 bits
• Linux 64 bits
• OSX 32 bits
• OSX 64 bits
• Windows 32 bits
• Windows 64 bits

Currently unable to get this working - Logs are giving a few warnings of this form:
2018-Jun-24 14:54:30.241 WARNING   core   string at 1695402092016 is length 1599361861 which is larger than cap 18 [E:\projects\Dwarf-Therapist\src\dfinstancewindows.cpp:100] (DFInstanceWindows::read_string)

Then a whole lot of errors about pointers that are as follows:
2018-Jun-24 14:54:30.244 ERROR   core   POINTER VECTOR SIZE IS NOT A MULTIPLE OF POINTER SIZE [E:\projects\Dwarf-Therapist\src\dfinstance.cpp:268] (DFInstance::enum_vec)

It seems like I'm doing something stupid, like saving the new memory layout file incorrectly or missing out part of the file, but I've tried a few times and with a few different methods (creating file using Atom and notepad, and ANSI/UTF-8 encoding). It's definitely saved as an ini file.

Anyone got any tips?

windows 10, running against 44.11 64bit, with the new R40.0.1 DT version. Don't have DFHack installed or any other utilities

[lethosor]

It's possible that some structure layouts changed. As far as I know, nobody has done much work on df-structures/DFHack yet, so those wouldn't have been identified and corrected for the layouts if that's the case. There's a reason they're experimental - sometimes they just don't work.

[Clément]

I generated those from BenLubar's automatic update, just before going to sleep. I did not test anything. It can work if the structures (at least the ones used by DT) did not change, this time it did not. Just wait for the fixed version.