I heard on another post that fixing the issue that way might remove sensitivity on the triggers.
I don't have any knowledge about SDL so I am not sure about what to do to fix this issue properly...

(Sorry if you get several emails like this, we're marking a bunch of bugs.)

We're hoping to ship SDL 2.0.11 on a much shorter timeframe than we have historically done releases, so I'm starting to tag bugs we hope to have closed in this release cycle.

Note that this tag means we just intend to scrutinize this bug for the 2.0.11 release: we may fix it, reject it, or even push it back to a later release for now, but this helps give us both a goal and a wishlist for the next release.

If this bug has been quiet for a few months and you have new information (such as, "this is definitely still broken" or "this got fixed at some point"), please feel free to retest and/or add more notes to the bug.

--ryan.

We're changing how we do SDL release versions; now releases will be even numbers (2.0.10, 2.0.12, etc), and as soon as we tag a release, we'll move the internal version number to an odd number (2.0.12 ships, we tag the latest in revision control as 2.0.13 immediately, which will become 2.0.14 on release, etc).

As such, I'm moving the bugs tagged with target-2.0.11 to target 2.0.12. Sorry if you get a lot of email from this change!

Thanks,
--ryan.

We're changing how we do SDL release versions; now releases will be even numbers (2.0.10, 2.0.12, etc), and as soon as we tag a release, we'll move the internal version number to an odd number (2.0.12 ships, we tag the latest in revision control as 2.0.13 immediately, which will become 2.0.14 on release, etc).

As such, I'm moving the bugs tagged with target-2.0.11 to target 2.0.12. Sorry if you get a lot of email from this change!

Thanks,
--ryan.

Yeah, we don't want to mark the triggers as buttons (which are considered either pressed or not, and never a matter of degree) unless we're completely desperate to get any input at all.

I suspect the correct thing to do here is mark these as half-axis inputs:

050000004c0500006802000000810000,Sony Computer Entertainment Wireless Controller,a:b0,b:b1,x:b3,y:b2,start:b9,guide:b10,back:b8,leftstick:b11,rightstick:b12,leftshoulder:b4,rightshoulder:b5,dpup:b13,dpleft:b15,dpdown:b14,dpright:b16,leftx:a0,lefty:a1,rightx:a3,righty:a4,lefttrigger:+a2,righttrigger:+a5,platform:Linux

Notice the two "+a" strings in there. Try that. If that doesn't work, make both "+a+ into "-a". Hopefully one of those two will make it work.

(what this says is: Normally axes are meant to go from -32768 to 32767, but for this one, assume its range is from 0 to 32767...or 0 to -32768 if you use -a instead of +a)

--ryan.

Sorry for the late answer, I did not have time to do a test before.

I tried with "lefttrigger:+a2,righttrigger:+a5" and 'lefttrigger:-a2,righttrigger:-a5" but I don't get any input with these settings...

Created attachment 4088 [details]
Updated PS3 mapping

Ryan, I think PS3 controllers don't have real triggers, just button inputs.

I'm attaching a patch for regression testing.

(In reply to Sam Lantinga from comment #7) 
> Ryan, I think PS3 controllers don't have real triggers, just button inputs.

Oh, huh, you're right, they do report as buttons! I had no idea.

I'll test this patch shortly.

--ryan.

I see some other PS3 controller mappings that use axes. I wonder if they should be adjusted as well?

(In reply to Sam Lantinga from comment #9)
> I see some other PS3 controller mappings that use axes. I wonder if they
> should be adjusted as well?

I think yes, but we'd need to make sure they use the same button numbers before we start making changes (also, does this thing show up as a different GUID when connected via Bluetooth vs USB?).

--ryan.

Yes, on Linux controllers connected via Bluetooth have the 05 prefix, and controllers connected via USB have the 03 prefix.

It looks like it's safe to make the same change to the mapping above this one in the patch - it's the same controller with a different version ... on second thought, maybe the firmware version matters here?

Wait, it's actually more weird than this.

On my Mac: an official Sony PS3 controller, plugged into USB, reports those triggers as buttons. I don't know if I ever installed some sort of joystick driver, years ago, but I don't think I did.

BUT: on Linux, plugged into USB, it reports as an axis with full range (resting at -32768, but you can definitely press it gradually through to 32767, it's not like a button press at the hardware level).

The triggers _also_ report as a buttons, which report themselves as pressed whenever you put any pressure on the trigger. So testjoystick will look like this on Linux as you pull and release a trigger:

INFO: Joystick 0 axis 5 value: -32768
INFO: Joystick 0 button 7 down
INFO: Joystick 0 axis 5 value: -22350
INFO: Joystick 0 axis 5 value: -16973
INFO: Joystick 0 axis 5 value: -14284
INFO: Joystick 0 axis 5 value: -12267
INFO: Joystick 0 axis 5 value: -8907
INFO: Joystick 0 axis 5 value: -7226
INFO: Joystick 0 axis 5 value: -6218
INFO: Joystick 0 axis 5 value: -5882
INFO: Joystick 0 axis 5 value: -5546
INFO: Joystick 0 axis 5 value: -4538
INFO: Joystick 0 axis 5 value: -2857
INFO: Joystick 0 axis 5 value: -841
INFO: Joystick 0 axis 5 value: 0
INFO: Joystick 0 axis 5 value: 2856
INFO: Joystick 0 axis 5 value: 5881
INFO: Joystick 0 axis 5 value: 7897
INFO: Joystick 0 axis 5 value: 9578
INFO: Joystick 0 axis 5 value: 10250
INFO: Joystick 0 axis 5 value: 9914
INFO: Joystick 0 axis 5 value: 10250
INFO: Joystick 0 axis 5 value: 11258
INFO: Joystick 0 axis 5 value: 14283
INFO: Joystick 0 axis 5 value: 17644
INFO: Joystick 0 axis 5 value: 19660
INFO: Joystick 0 axis 5 value: 21005
INFO: Joystick 0 axis 5 value: 22349
INFO: Joystick 0 axis 5 value: 24365
INFO: Joystick 0 axis 5 value: 27054
INFO: Joystick 0 axis 5 value: 29071
INFO: Joystick 0 axis 5 value: 30751
INFO: Joystick 0 axis 5 value: 32095
INFO: Joystick 0 axis 5 value: 32767
INFO: Joystick 0 axis 5 value: 28735
INFO: Joystick 0 axis 5 value: 11258
INFO: Joystick 0 axis 5 value: 0
INFO: Joystick 0 axis 5 value: -8571
INFO: Joystick 0 axis 5 value: -20670
INFO: Joystick 0 button 7 up
INFO: Joystick 0 axis 5 value: -32768


I assume we definitely do _not_ want to lose the axis precision on platforms that favor it over the button input.

--ryan.

That's correct. However, that mapping is for the Bluetooth pairing, which I presume doesn't have the full axis range?

I just double-checked and on the Steam Link both the official Sony PS3 controller and the Shanwan controller have full axis range for triggers, both when connected over USB and Bluetooth.

Fabrice, what version of Linux are you running?

I just tested on Linux Ubuntu 18.04 with the PS3 controller connected via Bluetooth and the original mapping works well for me, no issue with the left trigger.

Can you grab the SDL source code and build the testjoystick program in the test directory and run it and post the results of pulling the left trigger and then the right trigger?

Here's the output for me:
saml@saml-linux:~/projects/SDL/test$ ./testjoystick 0
INFO: There are 1 joysticks attached
INFO: Joystick 0: Sony PLAYSTATION(R)3 Controller
INFO:        type: Game Controller
INFO:        axes: 6
INFO:       balls: 0
INFO:        hats: 0
INFO:     buttons: 17
INFO: instance id: 0
INFO:        guid: 050000004c0500006802000000810000
INFO:     VID/PID: 0x054c/0x0268
INFO: Watching joystick 0: (Sony PLAYSTATION(R)3 Controller)
INFO: Joystick has 6 axes, 0 hats, 0 balls, and 17 buttons
INFO: Joystick 0 button 6 down
INFO: Joystick 0 axis 2 value: -32768
INFO: Joystick 0 axis 2 value: -8571
INFO: Joystick 0 axis 2 value: 21677
INFO: Joystick 0 axis 2 value: 32767
INFO: Joystick 0 axis 2 value: 0
INFO: Joystick 0 axis 2 value: -32768
INFO: Joystick 0 button 6 up
INFO: Joystick 0 button 7 down
INFO: Joystick 0 axis 5 value: -32768
INFO: Joystick 0 axis 5 value: -24703
INFO: Joystick 0 axis 5 value: 0
INFO: Joystick 0 axis 5 value: 32767
INFO: Joystick 0 axis 5 value: 0
INFO: Joystick 0 button 7 up
INFO: Joystick 0 axis 5 value: -32768

Here my Linux version by running uname -a:
Linux osmc 4.19.55-6-osmc #1 SMP PREEMPT Sun Nov 3 22:15:28 UTC 2019 armv7l GNU/Linux

I am running on Raspberry Pi 3 Model B (1GB RAM).

I will do the test with testjoystick and let you know.

By the way, in my case, I use SDL through moonlight-embedded. At first I reported the issue on irtimmer github page (https://github.com/irtimmer/moonlight-embedded/issues/747).
And someone mentionned that he had the same issue with a PS4 controller and that he could fix it in the same way I did.
I wanted to mention it just in case it can help :)

When I try to compile the test, I get the following error:

/home/osmc/SDL/SDL2-2.0.10/test/build/../testjoystick.c:95: undefined reference to `SDL_JoystickRumble'
collect2: error: ld returned 1 exit status
Makefile:186: recipe for target 'testjoystick' failed
make: *** [testjoystick] Error 1

Should I "make install" SDL2 so that the test can run?
Or can I specify a path where to find the library I compiled?

I have a PS3 controller and a Raspberry Pi, so I'll take this one.

Fabrice, how did you pair your PS3 controller over Bluetooth? I tried using sixpair and it didn't automatically pair after I unplugged it.

I didn't use sixpair, I wanted to use the default BT stack so that I can connect other devices at the same time.
So I used bluetoothctl to pair instead.
However, I don't remember how I did it, I followed a tutorial but it was a while ago so I don't have any link...

I just checked on a clean install of Raspbian stretch and triggers are working fine with the current mapping... ?

(In reply to fabrice.cros from comment #17)
> When I try to compile the test, I get the following error:
> 
> /home/osmc/SDL/SDL2-2.0.10/test/build/../testjoystick.c:95: undefined
> reference to `SDL_JoystickRumble'
> collect2: error: ld returned 1 exit status
> Makefile:186: recipe for target 'testjoystick' failed
> make: *** [testjoystick] Error 1
> 
> Should I "make install" SDL2 so that the test can run?
> Or can I specify a path where to find the library I compiled?

Yes, try:
sudo apt remove libsdl2-dev
make install

Here is what I get when I run the test and press the triggers:

osmc@osmc:~/SDL/SDL2-2.0.10/test$ ./testjoystick 0
INFO: There are 1 joysticks attached
INFO: Joystick 0: Sony Computer Entertainment Wireless Controller
INFO:        type: Game Controller
INFO:        axes: 6
INFO:       balls: 0
INFO:        hats: 0
INFO:     buttons: 17
INFO: instance id: 0
INFO:        guid: 050000004c0500006802000000810000
INFO:     VID/PID: 0x054c/0x0268
INFO: Watching joystick 0: (Sony Computer Entertainment Wireless Controller)
INFO: Joystick has 6 axes, 0 hats, 0 balls, and 17 buttons
INFO: The key you just pressed is not recognized by SDL. To help get this fixed, please report this to the SDL forums/mailing list <https://discourse.libsdl.org/> EVDEV KeyCode 313
INFO: Joystick 0 button 7 down
INFO: Joystick 0 axis 5 value: -32768
INFO: Joystick 0 axis 5 value: 32767
INFO: The key you just pressed is not recognized by SDL. To help get this fixed, please report this to the SDL forums/mailing list <https://discourse.libsdl.org/> EVDEV KeyCode 313
INFO: Joystick 0 axis 5 value: -13612
INFO: Joystick 0 button 7 up
INFO: Joystick 0 axis 5 value: -32768
INFO: The key you just pressed is not recognized by SDL. To help get this fixed, please report this to the SDL forums/mailing list <https://discourse.libsdl.org/> EVDEV KeyCode 312
INFO: Joystick 0 button 6 down
INFO: Joystick 0 axis 2 value: -32768
INFO: Joystick 0 axis 2 value: -27391
INFO: Joystick 0 axis 2 value: -4874
INFO: Joystick 0 axis 2 value: 32767
INFO: The key you just pressed is not recognized by SDL. To help get this fixed, please report this to the SDL forums/mailing list <https://discourse.libsdl.org/> EVDEV KeyCode 312
INFO: Joystick 0 button 6 up
INFO: Joystick 0 axis 2 value: -32768

It seems that there is one key that is not recognized properly. But you don't have it...
I am not sure if it is relevant or not...

It looks like maybe there was a bug in the version of SDL that shipped on the Raspberry Pi, and that it's working correctly now.

If you run testgamecontroller 0, does it register the trigger axes?

I think there might still be an issue with this key that is not recognized...
In this test and the previous one, I only pressed the triggers one after the other. And no other key. What is the unrecognized input?

Here is what I get with the test testgamecontroller:

osmc@osmc:~/SDL/SDL2-2.0.10/test$ ./testgamecontroller 0
INFO: Controller 0: PS3 Controller (guid 050000004c0500006802000000810000, VID 0x054c, PID 0x0268)
INFO: There are 1 game controller(s) attached (1 joystick(s))
INFO: Attempting to open device 0, guid 050000004c0500006802000000810000
INFO: Watching controller PS3 Controller
INFO: The key you just pressed is not recognized by SDL. To help get this fixed, please report this to the SDL forums/mailing list <https://discourse.libsdl.org/> EVDEV KeyCode 312
INFO: Controller axis lefttrigger changed to 0
INFO: Controller axis lefttrigger changed to 8905
INFO: Controller axis lefttrigger changed to 32767
INFO: The key you just pressed is not recognized by SDL. To help get this fixed, please report this to the SDL forums/mailing list <https://discourse.libsdl.org/> EVDEV KeyCode 312
INFO: Controller axis lefttrigger changed to 3024
INFO: Controller axis lefttrigger changed to 0
INFO: The key you just pressed is not recognized by SDL. To help get this fixed, please report this to the SDL forums/mailing list <https://discourse.libsdl.org/> EVDEV KeyCode 313
INFO: Controller axis righttrigger changed to 0
INFO: Controller axis righttrigger changed to 16383
INFO: Controller axis righttrigger changed to 32767
INFO: Controller axis righttrigger changed to 29238
INFO: Controller axis righttrigger changed to 1679
INFO: The key you just pressed is not recognized by SDL. To help get this fixed, please report this to the SDL forums/mailing list <https://discourse.libsdl.org/> EVDEV KeyCode 313

As you said in your previous message that it seems to work with this version, I tried again with moonlight-embedded now that I have the latest version of the lib SDL from source (and uninstalled the one from apt as you mentioned). 

However the issue is still there with moonlight...
And I compiled moonlgiht-embedded before the attempt, in case there is some dependency atcompilation.

Cameron, is this a Moonlight specific bug?

Any news on your side?
Nobody answered on the github of moonlight-embedded...
If you have any idea, I can test more