godot/modules/mono
Ignacio Roldán Etcheverry 0daa86d70e
Merge pull request #70511 from raulsntos/dotnet/fix-update-script-class-info
C#: Skip getting class info for unbound generics
2022-12-30 01:29:53 +01:00
..
build_scripts Rename float=64 SCons option to precision=double 2022-12-10 16:43:45 +01:00
doc_classes
editor C#: Avoid generic types in the script path attribute generator 2022-12-28 23:20:02 +01:00
glue Merge pull request #70511 from raulsntos/dotnet/fix-update-script-class-info 2022-12-30 01:29:53 +01:00
icons
mono_gd C#: Remove need for reflection to invoking callable delegates 2022-10-30 01:24:15 +02:00
thirdparty C#: Replace libnethost dependency to find hostfxr 2022-09-07 16:36:36 +02:00
utils Style: Cleanup header guards for consistency 2022-09-26 13:51:17 +02:00
.editorconfig
.gitignore .NET: Generate SdkPackageVersions.props from version.py 2022-10-05 20:31:26 +02:00
__init__.py
class_db_api_json.cpp Use JSON::stringify where possible 2022-10-11 00:27:23 +02:00
class_db_api_json.h
config.py SCons: Unify tools/target build type configuration 2022-09-26 16:31:46 +02:00
csharp_script.cpp C#: Rename ConvertToX methods 2022-12-23 19:40:32 +01:00
csharp_script.h Rename all gdnative occurences to gdextension 2022-12-12 11:04:57 +01:00
Directory.Build.props
Directory.Build.targets C#: Generate symbols packages 2022-10-08 13:59:06 +02:00
godotsharp_defs.h
godotsharp_dirs.cpp Unify usage of GLOBAL/EDITOR_GET 2022-10-18 19:01:48 +02:00
godotsharp_dirs.h [.NET Export] Improve .NET export process. 2022-10-07 13:33:06 +03:00
interop_types.h
managed_callable.cpp C#: Remove need for reflection to invoking callable delegates 2022-10-30 01:24:15 +02:00
managed_callable.h C#: Remove need for reflection to invoking callable delegates 2022-10-30 01:24:15 +02:00
mono_gc_handle.cpp
mono_gc_handle.h
README.md Rename float=64 SCons option to precision=double 2022-12-10 16:43:45 +01:00
register_types.cpp
register_types.h
SCsub SCons: Unify tools/target build type configuration 2022-09-26 16:31:46 +02:00
signal_awaiter_utils.cpp Rename CONNECT_ONESHOT TO CONNECT_ONE_SHOT 2022-09-06 19:00:33 +02:00
signal_awaiter_utils.h

How to build and run

  1. Build Godot with the module enabled: module_mono_enabled=yes.
  2. After building Godot, use it to generate the C# glue code:
    <godot_binary> --generate-mono-glue ./modules/mono/glue
    
  3. Build the C# solutions:
    ./modules/mono/build_scripts/build_assemblies.py --godot-output-dir ./bin
    

The paths specified in these examples assume the command is being run from the Godot source root.

How to deal with NuGet packages

We distribute the API assemblies, our source generators, and our custom MSBuild project SDK as NuGet packages. This is all transparent to the user, but it can make things complicated during development.

In order to use Godot with a development of those packages, we must create a local NuGet source where MSBuild can find them. This can be done with the .NET CLI:

dotnet nuget add source ~/MyLocalNugetSource --name MyLocalNugetSource

The Godot NuGet packages must be added to that local source. Additionally, we must make sure there are no other versions of the package in the NuGet cache, as MSBuild may pick one of those instead.

In order to simplify this process, the build_assemblies.py script provides the following --push-nupkgs-local option:

./modules/mono/build_scripts/build_assemblies.py --godot-output-dir ./bin \
    --push-nupkgs-local ~/MyLocalNugetSource

This option ensures the packages will be added to the specified local NuGet source and that conflicting versions of the package are removed from the NuGet cache. It's recommended to always use this option when building the C# solutions during development to avoid mistakes.

Double Precision Support (REAL_T_IS_DOUBLE)

Follow the above instructions but build Godot with the precision=double argument to scons

When building the NuGet packages, specify --precision=double - for example:

./modules/mono/build_scripts/build_assemblies.py --godot-output-dir ./bin \
    --push-nupkgs-local ~/MyLocalNugetSource --precision=double