I want to use the Bullet physics engine for a project of mine which is in D.  I have found many a bunch of different posts in the Forum (the most recent was in 2016).  The primary problem seemed to be dealing with the fact the Bullet is made in C++.  (It has been 4 years, has D & C++ binding gotten easier now?)

Bullet now has a first-party C API (which they use for PyBullet (Python bindings for Bullet)).  It should then theoretically be easy to make bindings!

I would really like to see bindings for Bullet created.  And I will be putting effort to do it, but I know there are others in the D community who would far better at managing it than myself.

@BLM768 @Dechcaudron @vuaru

On Thursday, 30 January 2020 at 18:52:36 UTC, Jonathan Levi wrote:
> I want to use the Bullet physics engine for a project of mine which is in D.  I have found many a bunch of different posts in the Forum (the most recent was in 2016).  The primary problem seemed to be dealing with the fact the Bullet is made in C++.  (It has been 4 years, has D & C++ binding gotten easier now?)
>
> Bullet now has a first-party C API (which they use for PyBullet (Python bindings for Bullet)).  It should then theoretically be easy to make bindings!
>
> I would really like to see bindings for Bullet created.  And I will be putting effort to do it, but I know there are others in the D community who would far better at managing it than myself.
>
> @BLM768 @Dechcaudron @vuaru

You don't need a binding to start using it. You can easily call C functions from D as long as you provide a proper linkage.

On Thursday, 30 January 2020 at 19:40:24 UTC, Ferhat Kurtulmuş wrote:
> On Thursday, 30 January 2020 at 18:52:36 UTC, Jonathan Levi wrote:
>> [...]
>
> You don't need a binding to start using it. You can easily call C functions from D as long as you provide a proper linkage.

Writing a binding is a good learning practice also. I was trying to call some opencv functions, and I ended up with an opencv binding.

On Thursday, 30 January 2020 at 18:52:36 UTC, Jonathan Levi wrote:
> I want to use the Bullet physics engine for a project of mine which is in D.  I have found many a bunch of different posts in the Forum (the most recent was in 2016).  The primary problem seemed to be dealing with the fact the Bullet is made in C++.  (It has been 4 years, has D & C++ binding gotten easier now?)
>
> Bullet now has a first-party C API (which they use for PyBullet (Python bindings for Bullet)).  It should then theoretically be easy to make bindings!
>
> I would really like to see bindings for Bullet created.  And I will be putting effort to do it, but I know there are others in the D community who would far better at managing it than myself.
>
> @BLM768 @Dechcaudron @vuaru

Do you have a link to the C API documentation/examples? Using dpp, it should be nothing more than adding an #include statement at the top of your D file and calling the functions. Or use dstep to generate the bindings.

On Thursday, 30 January 2020 at 20:53:39 UTC, bachmeier wrote:
> Do you have a link to the C API documentation/examples? Using dpp, it should be nothing more than adding an #include statement at the top of your D file and calling the functions. Or use dstep to generate the bindings.

This piqued my interest, too, but I don't see any documentation. The only C wrapper that I found is part of the shared memory client/server example. Pybullet piggybacks on that (an example program including code from a separate example). At first glance, all of this looks pretty weird and I can't make out how stable, complete or efficient it actually is.

On Friday, 31 January 2020 at 00:37:27 UTC, Gregor Mückl wrote:
> On Thursday, 30 January 2020 at 20:53:39 UTC, bachmeier wrote:
>> Do you have a link to the C API documentation/examples? Using dpp, it should be nothing more than adding an #include statement at the top of your D file and calling the functions. Or use dstep to generate the bindings.
>
> This piqued my interest, too, but I don't see any documentation. The only C wrapper that I found is part of the shared memory client/server example. Pybullet piggybacks on that (an example program including code from a separate example). At first glance, all of this looks pretty weird and I can't make out how stable, complete or efficient it actually is.

https://github.com/bulletphysics/bullet3/blob/master/examples/pybullet/pybullet.c

This sounds like a c binding

On Friday, 31 January 2020 at 09:00:45 UTC, Andrea Fontana wrote:
> On Friday, 31 January 2020 at 00:37:27 UTC, Gregor Mückl wrote:
>>
>> This piqued my interest, too, but I don't see any documentation. The only C wrapper that I found is part of the shared memory client/server example. Pybullet piggybacks on that (an example program including code from a separate example). At first glance, all of this looks pretty weird and I can't make out how stable, complete or efficient it actually is.
>
> https://github.com/bulletphysics/bullet3/blob/master/examples/pybullet/pybullet.c
>
> This sounds like a c binding

This is what I was referring to. The actual C wrapper layer is in the SharedMemory example code:

https://github.com/bulletphysics/bullet3/blob/master/examples/SharedMemory/PhysicsClientC_API.h

The file you linked is the python binding around that.

Bullet currently consists of two nearly entirely independent implementations in src/. This C wrapper seems to wrap only the newer bullet3 implementation, which looks noticeably less complete to me.

On Friday, 31 January 2020 at 10:02:24 UTC, Gregor Mückl wrote:
> On Friday, 31 January 2020 at 09:00:45 UTC, Andrea Fontana wrote:
>> On Friday, 31 January 2020 at 00:37:27 UTC, Gregor Mückl wrote:
>>>
>>> This piqued my interest, too, but I don't see any documentation. The only C wrapper that I found is part of the shared memory client/server example. Pybullet piggybacks on that (an example program including code from a separate example). At first glance, all of this looks pretty weird and I can't make out how stable, complete or efficient it actually is.
>>
>> https://github.com/bulletphysics/bullet3/blob/master/examples/pybullet/pybullet.c
>>
>> This sounds like a c binding
>
> This is what I was referring to. The actual C wrapper layer is in the SharedMemory example code:
>
> https://github.com/bulletphysics/bullet3/blob/master/examples/SharedMemory/PhysicsClientC_API.h
>
> The file you linked is the python binding around that.
>
> Bullet currently consists of two nearly entirely independent implementations in src/. This C wrapper seems to wrap only the newer bullet3 implementation, which looks noticeably less complete to me.

It must have started out as a Java implementation. Just look at the verbosity of the names:

B3_SHARED_API void b3CalculateInverseKinematicsAddTargetPositionWithOrientation(b3SharedMemoryCommandHandle commandHandle, int endEffectorLinkIndex, const double targetPosition[/*3*/], const double targetOrientation[/*4*/]);

On Friday, 31 January 2020 at 14:01:22 UTC, bachmeier wrote:
> It must have started out as a Java implementation. Just look at the verbosity of the names:
>
> B3_SHARED_API void b3CalculateInverseKinematicsAddTargetPositionWithOrientation(b3SharedMemoryCommandHandle commandHandle, int endEffectorLinkIndex, const double targetPosition[/*3*/], const double targetOrientation[/*4*/]);

This doesn't strike me as very verbose. I actually like verbose names because they explain what the method does well. And in a proper ide you can type CIKAT, press ctrl+space and the entire name will get autocompleted.