Print

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

[Clément]

This is not a known bug and copying a file with a new name each time is not trivial enough to happen by accident. If you cannot remember what you did on December 18th, we won't know. They may have been created by another application.

[feelotraveller]

Looking at the dates DT39.1 was released on December 18th (or 19th... time zones) so there is a good guess about one thing that was done.

The install path does however look quite non-standard.

(My windows days are barely memorable but is that what you would expect the file names to look like when a file of that name 'README.rst' already exists in that location?  E.g. 'README.rst (1)' if it gets written for a second time.)  Why it would get written a million times though, I have no idea... 

p.s. Clément, don't hate me forever... 

[lethosor]

Not that this helps, but it looks to me like that list is sorted alphabetically (999,999 comes between 9,999,989 and 9,999,990), so it's possible that there are between 1 and 9 million of those. If it's a power of two (e.g. 1,048,576), it could be due to something that duplicated everything matching some pattern too many times (20 in that case). I still have no idea what that could be, though.

[AfterShave]

Lethosor, the total number of files is listed at the bottom though 840798. So it does cap out at one million. File 0 without the appended (#) and File(999999).

Feelotraveller. Nah I just gave it L: because it's my largest disk. It could be literally any other folder name and be the same as far as I can see it. But yeah, copies in windows count up between the (). However, a plain copying and pasting will make it Filename - Copy (#). So it does look a bit weird. There's no Copy in the names. So I guess something tried to make new files somehow?

[PatrikLundell]

I get numbers without a copy indication when downloading files that have the same name as a previous download. Could it be something similar at play here?

[Clément]

New version released: 41.0.3

Changelog:

• added can_assign_military method for scripts
• changed military status grouping (remove outdated "Champions" and added "Mercenaries")
• added expulsion related thoughts
• fixed writing nicknames on some Linux distributions
• fixed trying to read itemdef flags for objects not from DF memory

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

[thistleknot]

May I so kindly ask where the wiki is for qt compilation instructions?

I'm not partial to QT.  I'm just trying to get my build environment up.  I was able to install & run make/cmake and derived a cmakelist.txt file that opens in QT (side note: unfortunately running make after cmake doesn't seem to find a makefile @ CLI).  I have msys which has g++.  I had to rename sh.exe to sh.bak to get the (qt) configuration to initially run through but receive the following

Spoiler (click to show/hide)

Code: [Select]

CMake Error at CMakeLists.txt:16 (find_package):
  By not providing "FindQt5.cmake" in CMAKE_MODULE_PATH this project has
  asked CMake to find a package configuration file provided by "Qt5", but
  CMake did not find one.

  Could not find a package configuration file provided by "Qt5" with any of
  the following names:

    Qt5Config.cmake
    qt5-config.cmake

  Add the installation prefix of "Qt5" to CMAKE_PREFIX_PATH or set "Qt5_DIR"
  to a directory containing one of the above files.  If "Qt5" provides a
  separate development package or SDK, be sure it has been installed.

[Clément]

Qt should be in the msys repository (mingw-w64-x86_64-qt5 for 64 bits), you do not have to compile it yourself. pacman will install it in a standard location that cmake should find without extra help. Otherwise, as the error message says: "Add the installation prefix of "Qt5" to CMAKE_PREFIX_PATH", that means adding "-DCMAKE_PREFIX_PATH=/path/to/qt5" to cmake options.

If you are not used to command line, you might prefer using Qt Creator with one of the official Qt builds.

[Clément]

I should stop postponing writing the windows build instructions, so I'll try a draft here.

I know several way to do it, but I don't know which one is the best. The best choice may depend on if you need a 32 bits or 64 bits build, if you want to develop (and your favorite development tools) or if you simply want to have a usable build.

• Qt Creator with built-in MinGW: less installaters but 32 bits only.
• Qt Creator with MSVC: more installers, 32/64 bits.
• Visual Studio: I don't know it, so I'll skip it.
• MSYS2: linux-style command line, easier if you used to it but difficult otherwise, more difficult to deploy.
• Build script with MSVC: what I use for release build currently, but my script is not really portable. A better script would still require some environment setup.

Qt Creator

Download and install Visual Studio 2017 or Visual Studio 2015 (choose C++ support during installation? I don't remember if it is always installed), if you want MSVC builds.

Download Qt Online Installer and install the components:

• Qt > Tools > Qt Creator
• Qt > Qt 5.x > MSVC 201x/MinGW 5.x, your choice of version
• Qt > Tools > MinGW 5.x if using MinGW, choose the same version as above

Download and install CMake.

Setup Qt Creator for using CMake.

Open Dwarf-Therapist project by selecting CMakeLists.txt. Select a kit and you are ready to build.

MSYS2

Download and install MSYS2. Run a MinGW console (32 or 64 bits, your choice).

Install dependencies with pacman (I am not sure about the package lists, I may have missed some).
For 32 bits:

Code: [Select]

pacman -S mingw-w64-i686-qt5 mingw-w64-i686-gcc mingw-w64-i686-cmake mingw-w64-i686-ninjaFor 64 bits:

Code: [Select]

pacman -S mingw-w64-x86_64-qt5 mingw-w64-x86_64-gcc mingw-w64-x86_64-cmake mingw-w64-x86_64-ninja
Then from Dwarf Therapist source directory:

• Create a build directory, and switch to it.

Code: [Select]

mkdir build
cd build

• Configure CMake

Code: [Select]

cmake .. -G Ninja

• Compile with ninja

Code: [Select]

ninja

On deployment

If you run DwarfTherapist.exe from the file explorer, it will fail to find dlls (for Qt or GNU libraries for MinGW builds). There is a windeployqt tool shipped with Qt that will copy the required dlls next to the executable. IIRC it does not copy enough dlls with MSYS2.

You also need to copy the content of the "share" directory in a "data" directory next to the executable.

[thistleknot]

I would have never have figured out msys2.  I thought that was the same as my msys

When attempting the Ninja step (64 set)

Spoiler (click to show/hide)

Code: [Select]

C:\Users\user\Documents\dt\Dwarf-Therapist\build>cmake .. -G Ninja
CMake Error: CMake was unable to find a build program corresponding to "Ninja".  CMAKE_MAKE_PROGRAM is not set.  You probably need to select a different build tool.
CMake Error: CMAKE_C_COMPILER not set, after EnableLanguage
CMake Error: CMAKE_CXX_COMPILER not set, after EnableLanguage
-- Configuring incomplete, errors occurred!
See also "C:/Users/user/Documents/dt/Dwarf-Therapist/build/CMakeFiles/CMakeOutput.log".

I also tried

Spoiler (click to show/hide)

Code: [Select]

"c:\Program Files\CMake\bin\cmake.exe" -G .. Ninja
C:\Users\user\Documents\dt\Dwarf-Therapist\build>"c:\Program Files\CMake\bin\cmake.exe" -G .. Ninja
CMake Error: Could not create named generator ..

Generators
  Visual Studio 15 2017 [arch] = Generates Visual Studio 2017 project files.
...



I also tried (reverting changes to original commit each attempt)

Spoiler (click to show/hide)

Code: [Select]

C:\Users\user\Documents\dt\Dwarf-Therapist\build>cmake .. -G "MSYS Makefiles"
-- The C compiler identification is GNU 4.9.2
-- The CXX compiler identification is GNU 4.9.2
-- Check for working C compiler: C:/Strawberry/c/bin/gcc.exe
-- Check for working C compiler: C:/Strawberry/c/bin/gcc.exe -- broken
CMake Error at C:/Program Files/CMake/share/cmake-3.13/Modules/CMakeTestCCompiler.cmake:52 (message):
  The C compiler

    "C:/Strawberry/c/bin/gcc.exe"

  is not able to compile a simple test program.

  It fails with the following output:

    Change Dir: C:/Users/user/Documents/dt/Dwarf-Therapist/build/CMakeFiles/CMakeTmp

    Run Build Command:"C:/PROGRA~2/GnuWin32/bin/make.exe" "cmTC_fed14/fast"
    C:/PROGRA~2/GnuWin32/bin/make.exe -f CMakeFiles/cmTC_fed14.dir/build.make CMakeFiles/cmTC_fed14.dir/build
    make.exe[1]: Entering directory `C:/Users/user/Documents/dt/Dwarf-Therapist/build/CMakeFiles/CMakeTmp'
    The system cannot find the path specified.
    make.exe[1]: *** [CMakeFiles/cmTC_fed14.dir/testCCompiler.c.obj] Error 1
    make.exe[1]: Leaving directory `C:/Users/user/Documents/dt/Dwarf-Therapist/build/CMakeFiles/CMakeTmp'
    make.exe: *** [cmTC_fed14/fast] Error 2

[Clément]

Your environment does not look right. Are you sure you are using a MinGW shell?

[thistleknot]

Your environment does not look right. Are you sure you are using a MinGW shell?

you mean a shell with mingw in my path?

I read I didn't have the correct mingw32, so I ran mingw-get and pulled the mingw32-base and it populated that folder (c:\mingw), but I already had mingw path from the qt ver of mingw that was installed.

Yes.

[Clément]

MSYS2 installer should have created shortcuts for three kind of shell: MSYS, MINGW32, and MINGW64, they setup the proper environment so you don't have to worry about it. For compiling you will need to use a MINGW one, for package management any of them will do fine.

Don't use mingw-get (it's from your outdated mingw), use pacman from MSYS2. And don't forget to check for updates after the installation as instructed on the msys2 website (step 5 and 6).

[thistleknot]

what a fun filled time that was

I have the columns sorted by role %, then based on attribute % (all other role dimension weights set to 0, i.e. disabled).    The one that's 0% that's the role next to it's [raw] value reported as a %.

I suppose this was an issue with the older version as well, and is something that I think as a statistician should be able to solve (the way that analytical ability > agility in all %'s.  In other words, unequal means/variances (ANOVA). Different co-efficients yet normalizing them the same.  I'm not sure if there is an easy way around it.  Maybe a tri-mean vs median?  Maybe normalize each domain separately? 

I don't like that idea, leave as is?  I mean I see one agility score > .5.

I also think highlighting only the values that are either the top 10% or bottom 10% of the displayed roles (this is not the same as roles > 90% unless the screen is redrawn using recdf).  It means what is drawn, from all those available squares, select the highest 10% of those squares regardless of their value (i.e. grab the top 10% ranked).  In other words, a rank ecdf of the final drawing.

Spoiler (click to show/hide)

biggest mistake (on my part) was the need to be running from the

there is a shortcut under

start - 'msys2 - 64-bit' - 'MSYS2 MinGW 64-bit'

which maps to

C:\msys64\msys2_shell.cmd -mingw64

which is where the aforementioned commands need to be run under, it's not your standard cmd prompt.  Easy to make considering that pacman commands work inside cmd.exe

once past that it compile (along with the relevant github commands to change forks).

here are my high level notes from that session

Spoiler (click to show/hide)

Code: [Select]

#git clone https://github.com/cvuchener/Dwarf-Therapist.git

#from cmd prompt from git[hub (desktop)] project
git remote add cvuchener https://github.com/cvuchener/Dwarf-Therapist.git
git fetch cvuchener
git checkout cvuchener/role-needs

git commit
#git apply

confirm branch with

git status


#c:\msys64\msys2

C:\msys64\msys2_shell.cmd -mingw64

ninja

pacman -Syu

cd /c/Users/user/Documents/dt/Dwarf-Therapist

CMakeLists.txt
target_link_libraries(${EXE_NAME} Qt5::Widgets Qt5::Qml Qt5::Concurrent Qt5::Core Qt5::Gui Qt5::Network ${LIBS})

cp -R ../share data

./DwarfTherapist.exe --trace


there was no need for me to even open it in qt creator.  It built it from the ninja statement.

thank you clement for the support

update

already confirmed the median is reported correct (1091) and at 50%!

I'm not sure how the other values are calculated.  Eyeballing it it appears the edge cases are done correctly as well as the center, trying to confirm individual #s I find them to be within 5%, so I think there is some sort of post stretching being done.

Don't recall how to use --stats

Ex transform check

Spoiler (click to show/hide)

[soul4hdwn]

Hi, how do I put my custom/super labor to appear on the main grid/column assignation part of screen?

Do I have to make a role instead?  I couldn't find a way to edit grids (the compact labor in specific) or make a new grid view tab.

I apologize if was already asked, I see progress on customizing and custom features but I didn't see my specific and hopefully minor issue recently via the search.  I'm not quite a power user, sorry.

edit: "role" isn't related but issue still stands on using/seeing the custom/super labor on grid and or activating it on a dwarf.