GSoC Ideas page - XInput

[Juan Jose Gonzalez]

Hi,

Thank you for your answer. As soon as I have a little time I'll take a
look at the HID infrastructure and try to create a backend for my code
with it. Right now I'm pretty busy, but in a month or so I should have a
little more free time.

I'll also post a patch for the "front-end" of my xinput implementation
as soon as I find a little time to look over it, since it's been a while
since I last touched it and I don't know if it's in a good enough state
to be published.


Regards,
Juan


On 02/11/2016 02:07 PM, Aric Stewart wrote:
> On 2/11/16 1:51 AM, Juan Jose Gonzalez wrote:
>> Hi,
> Hello!
>
>> I saw the XInput (Xbox controller input) entry on the SoC site and was
>> wondering what the requirements are.
>>
>> Context: I implemented XInput (Input + Rumble) a few months ago and have
>> been using it to play XInput-only games since. My implementation allows
>> for different backends, but I've only implemented an evdev backend for
>> linux. I also started working on editable controller mappings through
>> registry entries. I didn't try to submit any patches or finish the
>> mapping config, since by the time I had polished my code enough, the
>> hidclass development had gained steam and it seemed that that was what
>> wine was supposed to use for XInput.
>>
>> Another option, and in my opinion the cleanest, would be to directly
>> access XBox controllers through USB, possibly using code from the
>> xboxdrv project.
>>
>> What are your thoughts on this? Should the XInput driver be based on
>> hidclass, xboxdrv, or should it be able to use different backends? If
>> it's just supposed to use hidclass, my implementation is pretty much
>> useless, since splitting backend and driver seems like unnecessary
>> overhead when hidclass already uses different backends.
>>
> As the primary author working on the HID support I feel pretty strongly that we should be using the HID back-end for Xinput. Xinput support is actually a large part of why I have been working on it. I feel like yet another controller implementation that has its own back ends is really not what wine needs.
>
> I doubt that all your work would be useless. There is still a lot of Xinput work that happens above the backend. I am sure there is a lot there you have done that would be extremely valuable. Maybe you could start by using your existing code and think if HID as another back end to your code. I would recommend looking that the hid.dll functions for Xinput. It is likely there are hid.dll functions that are missing that you need. Those are additional opportunities for things to implement. 
>
> If we work together on this then it also gives some more weight to the HID infrastructure.
>
> Your back end work is also probably still useful also. I assume it was using linux input, as opposed to hidraw or something, We still need someone to help with the back ends for HID and HidClass around linux input. 
>
> Both of that expertise would be extremely valuable!
>
> -aric
>
>
>> On 02/10/2016 10:04 PM, André Hentschel wrote:
>>> Hi,
>>>
>>> please update https://wiki.winehq.org/Summer_Of_Code with your Ideas and mention yourself as possible mentor where appropriate.
>>> Some things can be dropped I think, OpenMP maybe?
>>>
>>> Thanks!
>>>
>>>
>>
>>