New bug introduced in 0.30.28 for FAST LEDs

[Brian Cox]

Hi guys,

I already brought this up here ( https://groups.google.com/forum/#!topic/mpf-users/Evrl00hDwx0 ), but wanted to re-post since this is almost certainly a bug, similar to the switch debounce bug posted earlier, and wanted to tag it as such.

I'm using a FAST controller with RGB LEDs, and this was working on version 0.30.23, and then broke when upgraded to 0.30.28.

I did a test where I changed my config to use INT instead of HEX, like so:

fast:

  ports: /dev/tty.usbserial-1454B, /dev/tty.usbserial-1454C

  config_number_format: int

...and then purged my other configs of any device that used *HEX* as a number format (coils, switches and LEDs).

I then set up a single small show using this reduced set of LEDs, and that show worked.

Aside from figuring out this specific bug, can I reiterate my request for instructions on how to roll back MPF to working versions ???????

Rolling back would be extremely helpful to track down these new bugs as they are introduced, especially in light of no documentation,

Thanks,

BC

[Brian Madden]

Dude you gotta give us more than 6 hours to reply on over a holiday weekend before reiterating your questions with six question marks. :)

I saw Jan answered your other thread about how to roll back to an older specific version of MPF. Those kind of questions are things you can google since they relate to installing Python packages and pip in general, so you don't have to wait for us to get back to you with answers like that.

Can you try rolling back to 0.30.27 instead of .23? In looking at the changes from .23 to .28, that would be my best guess as to where the source of the bug could be if it turns out to be a bug. Same as Jan said in the other thread, also run it with debug: true in one of the LEDs so we can see all the debug details in the log.

As for no documentation, I'm working on it as fast as I can, but I don't think I'll ever get documentation that says "This is a bug." :) Be sure to use the new docs site which is docs.missionpinball.org (not .com).

I'll  be spending time doing labor day cookout type things today, so I probably won't be able to respond until tonight or tomorrow, but if you could, post the -v log with a debug: true, and also try in .27 and see if that fixes it.

Thanks!