Yes, the 2 devices went offline because I moved the fridge and the power supplies were destroyed so today the new power supply came in, they been like 2 days offline.
I did not have complaints from my wife when the power supplies were destroyed because was her fault I had to move the fridge, I had a sensor on top of it and she pushed it and fall behind the fridge....
I've been using your driver for a few weeks. The previous version did slow down my hub. .86 is running pretty damn good. I've not "felt" any slow downs, I'm not running any tests or anything but I'm no longer noticing my actions lagging.
But, there's always a "but".. I am having trouble with hue,sat,level settings. I just can't seem to get a good red. If I hit the preset for "fade to red" it looks fantastic. If I set the values manually 100/100/100 the lights just turn off. If I do 99/99/100 I get pink. I'm using the controller RGBW 5-pin device.
Thank you for the excellent work, after spending $100 on Hue and Lightify bulbs it has been immensly gratifying to take $30 and light 15 feet of cabinetry. Keep up the great work, sorry about the "but".
I don't think the crashes were WebSockets client specific. I've had two WS clients running since my last message in this thread and everything seems to be okay with them.
However, I believe I might have an idea of one of the reasons I have periodic crashes. I related a story here:
I didn't explicitly say why my hub crashed last night but didn't crash every other night. If my last feature request I mention that some log messages take a long time. I was debugging an interface to another vendor's hub using the WS client. Some of the responses from the WS client were pages of JSON and I was logging them in trace.
I have a feeling that either large log message or large amounts of small log messages may be the culprit in at least some of my crashes. Sort of a negative feedback loop too when it starts to slow down. For the most part I am probably more careful than 99% of HE users with my log messages. Sometimes when I'm developing a new driver I have to turn on trace level debugging though and that might be when disaster strikes. I'm going to watch it more closely now.
So, long story short... you can probably feel okay about the WS client.
We’re still looking into it. I’m using an alternative version with Telnet currently (releasing Friday if it keeps working) that doesn’t slow down. I’m using a new protocol of Socket for Hubitat here, so there may be some issues with that.
I’ll also take the logging into account. I’m out traveling this week, so I know I’d better make sure I don’t push any bad code and force the hub down while I’m gone
@adamkempenich Adam, I have an RGB light that I'm not able to restore a previous state after capturing. I thought I remembered that you had said this was not possible, so I wasn't surprised it didn't work. But searching this thread doesn't yield any results when I look for that. Is it possible? Thanks!
Baby steps here... I know things have been pretty quiet with this driver.
While updating this has taken a while, it looks like we've finally caught the bug that was slowing down hubs with this driver. I've been updating the drivers, cleaning them up to match other platform drivers, and testing them.
There is a new option to choose between telnet or socket. You should use TELNET until Hub update 2.1 is officially released.
Unfortunately, Telnet won't let you refresh data from your device. Socket does, but that's where the bug was hiding.
I'll keep updating this list as time goes on, but the following devices are now on 0.87:
All devices have been updated to 0.87 --> Edit your device settings to use telnet <-- --> Optionally, reboot your hub to ensure all old connections were closed.
WW+CW Bulb drivers will need to be copied/pasted in from GitHub (the importURL was wrong)
I see that 2.1 has dropped publicly. Let me run some tests on my end and ensure everything works as expected, and I'll make sure socket is safe to use for y'all
I'll be traveling for work the next couple of weeks, but each device seems stable and functional. I'll keep working on it, even from the road, but wanted to give a fair heads up that I won't be able to really do any tests for a while.