Trouble understanding mSkipAxis

[hzulla]

I'm having trouble understanding mSkipAxis in filterTrigger().

The variable isn't setup and just assumed to be false.

Once it's true, it stays true, even if the device configuration later reaches a non-Trigger inputID and the device wants to use Axis 2 or 5 there.

So there can be edge cases where a device won't configure properly.

Would it be okay to just remove mSkipAxis, even if it causes more calls to strstr during the configuration run?

[psyke83]

The purpose of mSkipAxis was to work around an issue I encountered during configuration with my DS3 controller: the left and right trigger would map correctly, but the next item after the right trigger - left thumb - would be erroneously skipped over.

Can you check and see if the same issue occurs with your BigBen controller when mSkipAxis is removed? Thanks.

[hzulla]

Ah, that makes sense. Thanks for explaining, that makes my original interpretation of how it works wrong. The edge case remains, though, so will look into this.

[psyke83]

Something like this might improve mSkipAxis behaviour: https://github.com/RetroPie/EmulationStation/compare/master...psyke83:mSkipAxis_fix

If you think this implementation is incorrect or may still cause edge cases, let me know... otherwise I can submit as-is.

[hzulla]

Thanks. Yes, I do think that's the possible edge case, but let me test this first. (After the weekend, though.)

[hzulla]

Maybe just leave it as it is for now. The edge case I described is just that - an edge case. I will try to offer a proper fix for #568 some time later, but until then the existing behaviour of EmulationStation seems fine. Thanks.