Posts made by David Ehnebuske

So, after playing with the onionGpio module a bit more, I've reluctantly concluded that I can't use it to read gpio input that's reliable enough to do things like read a rotary encoder, even after fixing the bug in what it returns for getValue(). The trouble is partly that reading a pin is really slow because the implementation uses file manipulation and file I/O to read the bit in a hardware register somewhere that represents the pin. And partly it's because Python, being a garbage-collecting language, might take a break from executing your code to do a bunch of housekeeping between one of line of your code and the next. (Maybe it's possible to tell the Python environment to hold off on housekeeping stuff for a section of code. Some languages that garbage collect let you do that. I wouldn't know. But doing a getValue() executes a lot of code, as I said.)

That means, for example, that if you're trying to decide which pin went HIGH first, pin1 or pin2, you can't do it reliably at the millisecond level by doing

currentPin1 = pin1.getValue()
currentPin2 = pin2.getValue()

because too much time may have passed between the execution of the two lines of code.

onionGpio (once it's fixed) should be okay for things that aren't time critical, but I think you're better off, even for only moderately time-sensitive stuff, to go for a different environment.

Has anyone had success using the onionGpio python package? I managed to make it work, sort of, but not without problems. Here's what I've noticed.

First, out of the box, it only works with Python 2.7, not 3.6. You can easily port it to Python 3.6 by copying onionGpio.py from /usr/lib/Python2.7 to /usr/lib/Python3.6 and then editing the file to make the print statements in it into Python3-compatible calls to print(). There are only a few and they's debugging stuff.

Second, the documentation says the getValue() method returns 0 or 1. Actually, it returns 0, '0', or '1'. Mostly the latter two except in the case of an error. To me it looks like a bug in the implementation. Also, it seems to me that it would be much better if getValue() returned True or False since it's reporting on whether a digital pin is on or off.

Third, the implementation is a bit fragile in the sense that interrupting a program using it (for example, by typing ^C) is fairly likely to leave the GPIO system in a state that causes the next program that uses the module to fail when it runs. The reason for this is that the module got interrupted before it cleans up after itself.

You can manually clean things up by going to /sys/class/gpio and doing an ls. Look for a directory named gpio? where ? is a number. then type the command:

echo "?" > unexport

where ? is the number at the end of the gpio? you found. E. g., if you find gpio2 listed, you'd type:

echo "2" > unexport

to clean it up.

I guess not. I reluctantly conclude this stack of stuff doesn't work. Too bad. Time to see if I can find something else to do with an Omega 2+.

Fwiw, I've tried to do this, too, and get the same non-results you did.

I've been trying to get the Onion BLE dongle to work as a connection to some bluetooth speakers. Using opkg, I've installed bluez-libs, bluez-utils, pulseaudio-daemon and pulseaudio-tools. But I have not been able to make it all work.

I can get the BLE dongle up and running (using hciconfig hci0 up). I can then use the various subcommands of bluetoothctl to locate my BT speakers and pair with them. But I cannot connect to them. It always just reports a generic failure.

As for pulseaudio, I can get the daemon up (using pulseaudio --start -D) but it crashes after responding to a query with pacmd list. (The list shows no inputs, no outputs and no sound cards.)

Has anyone gotten farther?

@Leonardo-Motyczka Hmmm. Just to be sure, did you upgrade to the latest release of the firmware? My Omega2+ is running Ω-ware: 0.1.9 b150. I needed to upgrade using a USB flash drive as detailed above since I couldn't connect via Wifi.

The only other thing I can think of is that people have reported the Omega2+ won't connect to Wifi APs whose SSID contains special characters. I don't think anyone's tracked down which characters don't work, though.

Whatever the causes, it is certainly annoying and disheartening to open the box, turn it on and have essentially nothing work and get little or no feedback about what has gone wrong. I feel your pain. I hope you find a solution soon.

The answer to my question is, "Yes, it can." It appears that my Omega2+ can't connect to an open Wifi AP.

I first updated the firmware from a USB drive and tried again. Same story.

Next I added WPA2PSK security to the Wifi AP I was trying to connect to (leaving the SSID and everything else unchanged). I tried configuring the Omega2+ again, and all was well. Who'd have thought.

So, I found the -v option on the "wifimanager" command. When I use that, I see it doesn't get very far before the command hangs:

root@Omega-6E87:/# wifimanager -v
radio0 is up
Reading configured networks in station mode...
"Ehnebuske"

Scanning nearby networks...
found"Ehnebuske"



Current Count: 1
Number of configured nets: 1

Then nothing. Is it possible that the Omega2+ can't connect to an open Wifi AP?

After doing a "factory reset" on my Omega2+ it boots up nicely and its AP shows up. But if I connect to it, all I get is an "Index of /console/" page in the web browser. Thinking I just need to oupgrade, I thought I'd use the commands "wifisetup" and then "wifimanager" to connect to the internet. After running multiple experiments, here's where I ended up.

The "wifisetup" command looks like it works okay. It scans for and reports the available Wifi networks, and lets me choose the one I want. After doing that, if I do a "wifisetup list" command, I see a JSON structure showing the AP I chose. It's an open network (no security) with a simple SSID -- nine letters, all lowercase except for the first one. But if I then run the wifimanager command, it never completes -- the command prompt never comes back, though I can bring it back by typing ^C.

If I reboot, the amber LED on Omega2+ never changes from flashing to steady.

Anyone have an idea what might be going wrong, or what I should do next?