Is this project dead?

[fchorney]

It's been a while since I've done Pi development, but it seems to me like this lib is missing code for the zero and the + series? I can't seem to use any of the extended header pins with this library on my zero, and trying to get the info just returns with "KeyError: '900092'"

Just curious if this is still actually being developed, or if I should be looking elsewhere?

[chdsbd]

I gave up on using this project and transferred over to using RPi.GPIO, which provides pretty much the same functionality.

[fchorney]

Yeah that seems to be what I was thinking of doing. It's a shame since this lib has a few things that I liked.

[karel-kahula]

ya me 2

[k4p7]

+1

[alexellis]

@metachris If this was fixed it would serve a lot of people with PI2s wanting to generate PWM signals - it was perfect for robots (servos, motor drivers, LEDs).

As far as I know it still works on ARMv6 - PI A/A+/B/B+/Zero. An alternative may to be to use a tiny 2 dollar Arduino nano as a co-op board over USB with something like PyFirmata.

The merge looks like a nightmare - with so much alterations between RPIO and the original DMA/PWM code.

[chdsbd]

@alexellis RPi.GPIO provides basically all of the functionality of this library (PWM included) while also being maintained and compatible with the latest Raspberry Pis.

[fchorney]

Not necessarily. One example I can think of is that this lib allows you to set up an interrupt on any edge, and the function it calls actually passes in which edge the interrupt fired on, while in RPi.GPIO you can only choose up, down, or both, but on both it doesn't tell you which it triggered on.

[alexellis]

@chdsbd RPi.GPIO really does not match this library's key feature - which I would say is hardware PWM signals through DMA. Even the front page of "RPi.GPIO" states how this is only covered through software - so much less reliable and prone to jitter / higher CPU loads.

If PWM is not important to your projects, then one of the other libraries should be fine for your purposes.

[chdsbd]

@alexellis I have been using RPi.GPIO to trigger servos without issue, so software PWM is at least reliable enough to control servos. Maybe there would be issues controlling LEDs precisely.

[DEvil0000]

RPi.GPIO is fine for simple things but for my use cases it is much to slow.
I really need the DMA stuff so I will stick with this and fix what I need.

@chdsbd It is just not true to say RPi.GPIO provides the same features or performance.

[rdagger]

Agree that this library provides some useful features not found in others. I looked into fixing it a few months ago and it was not trivial. If I recall correctly, it wasn't a matter of just updating addresses. Fundamental changes related to Mailbox property interfaces were needed.

[lurch]

Even without adding PWM support for the Pi2 / Pi3, it would be great if the version of RPIO on pypy could be updated to support B+ / A+ / PiZero 'out of the box' without having to manually patch the code (see #35).

[lurch]

For anybody looking for accurate PWM, there's also the http://abyz.co.uk/rpi/pigpio/ project, which does work on Pi2 / Pi3 / etc.

[JennaSys]

RPIO is a great clean pythonic library, but without it being maintained for the newer versions of the RasPi it doesn't do much good. I've since moved on to using pigpio which is very full featured. That one requires you to run a daemon on the RasPi, but doing so also lets you do remote calls, making it possible to develop code on another dev machine (like a laptop) while still allowing your code to access the GPIO port on a RasPi.