Using FarmbotOS beta updates

Maybe a PinBinding / CircuitsGPIO issue ( else, if Pi GPIO 9 is "special" should it be selectable for binding ? )

Setup a binding Pi GPIO 9 -> Action -> Read Status and then SYNC NOW and got

2019-08-26T14:00:00Z
Ha! I think I see the problem /boot/config.txt says dtparam=spi=on so the WebApp front-end should not be offering pins Pi GPIO 7,8,9,10,11 for Pin Binding :frowning:

Have an enormous ElixirConf US 2019 ! . . . blow their socks off with FarmBot OS v8 :slight_smile:

2019-09-05T14:00:00Z
Hi @connor, do you need me to open Issues for the 3 remaining problems in 8.0.0-rc64 ?

Also, can you tell me where the following [debug] logger messages originate ?
I want to try fixing them too.

2019-09-12T14:00:00Z _ Hi @connor

Seems you’ve ‘nailed’ ( fixed **even more** :slight_smile: ) the problem above of full recovery after an “Internet” outage, so thanks for that.

Testing today 2019-09-12T14:00:00Z on 8.0.0-rc80 and just noticed that the package includes all the artefacts for the farmbot 8.0.0-rc68 release also !?

Apologies! I actually didn’t see this thread get created for some reason. I’m currently working through these today, and am hoping to get a recover deploy out early this week with fixes for this stuff. Will keep you updated!

Thanks @connor !

I’ve discovered that the :mdns_lite app is emitting those [debug] IGNORING log records.
I’ll see if I can fix it and post something over in their repo.

The WebApp PIN BINDING list is much shorter now :slight_smile: But I suspect even little old Pi GPIO 4 should also be off limits because I get this after Binding it first time :

Update ! @connor, the :mdns_lite logger noise has been squelched by author . . just update the dep version.

2 Likes

Testing on v8.2.0-rc2

Hi @connor

Seems a config adjustment is required to support Telemetry over AMQP :slight_smile:

@jsimmonds The fix for this was deployed just now. Sorry for the inconvenience.

1 Like

@RickCarlino, the only “inconvenience” was having to stop my ad-hoc, unofficial QA activities :slight_smile:
Thanks !

2 Likes

Some quick feedback on Pre-Release v8.2.4-rc3

  • Farmwares fail, e.g.
02:07:46.101 [info]  [take-photo] OpenCV import error.
  • espeak “kind `a works” ™ ( very chopped-up audio and only on HDMI )

2019-12-12T13:00:00Z

Testing on Pre-Release v8.2.4-rc7

Looking a heap better in many ways ( Thanks @connor )

  • Farmwares work normally ( no OpenCV issue )
  • Sequences which include multiple Farmwares ( e.g. take-photo ) run to completion !

Still needing fixes

  • espeak “kind `a works” ™ ( very chopped-up audio and only on HDMI )
  • timing race around pipes/CeleryScript compiler/AST still shows up occasionally
    and aborts a running Sequence.
    Wonder if a fix similar to one of your farmware.ex fixes is a patch for this ? ( swallow up that :DOWN message so it won’t propagate )

Testing Pre-Release v8.2.4-rc8

  • Sequences with multiple Farmware calls now seem to be completely reliable.
    ( Deft refactoring by @connor )
  • espeak issues persist.

Now testing Pre-Release v9.0.0-rc4

  • espeak issues still persist ( nothing on analogue out ; very chopped-up audio on HDMI )
  • The Network Quality pings from the WebApp now result in one Logger message per “ping” !
    Also, the response from the device is published into MQTT ( wasn’t happening in v8.2.4 )

Picture this

Good catch- I have created a fix which will be deployed later this evening. Github Link.

Hi @RickCarlino . . notice your team is back at work big-time ( no Xmas break yet ? 0-: )

Just tried FBOS Pre-Release v9.0.0-rc9 and it won’t fly for me . . see image following.
I saw issues with FirmwareReset.NULL.init function in the log output (serial) and then Linux re-boot >> loop <<.

Thanks.

Hi @jsimmonds! I will take a look now, thanks for letting us know about this.

Glad to help !
@RickCarlino, on this topic, I was thinking recently that a little bit of automation to “smoke-test” these Pre-Release versions might be a good thing ?

Seems all fixed in Pre-Release v9.0.1-rc1.

Just noticed today that :mdns_lite seems to have stopped working ( responding ) . . maybe after a recent deps versions update ?

It’s a good idea @jsimmonds. That last RC was a bit rushed due to some internal things going on here at HQ.

I will look into the :mdns_lite issues. If you have any additional information, it might be a good issue for us to track on Github.