markv58

Bugs, I hate bugs! Found one. Didn't test for all possible combinations of RtoP, CtoF and FtoC. Testing for a bit then triggering the update, nothing big but important to keep the unnecessary polyglot activity curtailed.

Virtual 1.2.0 (stable) - Temperature Nodes - Code cleaned, timestamp added to log when a variable is changed, there is no driver to add to the node display at present so that did not happen. The values will only update now if a variable has changed since the last check and the counter will increase until that change happens. Data storage only happens when there is a change.

Now that it looks like the Virtual Temp nodes are finally doing what they're suppose to do, I am working on the final tweaks. 1. Stop the node from updating the temp panel and storing data if there is no change in the value from the previous pull. This will take a load off of the polyglot and reduce writes to the sd card or drive. This is in progress and is working well, the Since Last Update will continue to count up until there is a change to the Current value. 2. Add a time stamp to the lower right area beside the Convert F to C or C to F where it is blank and indicate of the last time the value updated. 3. Stripping out un-needed logger items and dressing up the code. This will hopefully be the last update on those nodes, it's pretty simple stuff so there should be no bugs, stay tuned.

@Teken Once again I think you can pull negative numbers, trial and error with dozens of parsing regex and methods I believe I've landed on one that will reliably and consistently do the job. Watch for the update, I think it's 1.0.20, and I think you owe me a beer.

I am seeing some very strange errors in logs. It seems that parsing negative numbers may or may not work and kicks up an error, when that happens it skips the update and moves on. I guess I'll work on the regex and try to find something that works consistently. Makes no sense at all what is happening.

I still think the best way in most situations is to have a program do the work, that way the only time the temp value is set is when there is an actual change.

I just added another delay to the long poll of .1 seconds v1.0.19. The delay you can set is a parseDelay in custom config params. It is set to .1 as default but you can increase it if you want. It adds a delay between the pull and the parse of the data to offset any latency there may be in the network. Virtual uses the rest interface to get the value from ISY and while that is very quick it could be delayed.

Virtual v1.0.18: Negative values are parsed correctly without error. Increased Variable IDs to 1000 for @Teken however the drop down is gone so scrolling to or direct input (recommended) then hitting Enter or the button to the left is required to register the change. Variable 71 has a value of -155, Converted from Raw and Converted from F to C. Update Profile on this one.

Thanks for the reminder, I keep forgetting.

@Teken It looks like the parser does not like negative numbers, I finally got an error. Must dig in and wrestle that bug out! After some design work, I had decided to stand the energy node solo, it's the right call.

@Teken It looks like Raw to Prec is doing what it's supposed to do. Setting the variable causes the raw data to be 4 digits and the raw to prec divides by 10. Looking at your log, it appears that setting prec to 1 doesn't do anything other that add the extra digit which is always 0. Set your variable prec to 0, reset your statistics and you should be good to go.

@gviliunas Great! I still don’t know what was going on. This morning looking thru your log it seemed that you were on a different version, some of the info was not there that should have been. Maybe there was a glitch in the update. I tested pulling from a non existing variable and the data came back as all 0 with no errors so that may be the lowest temp problem there. When you reset the statistics, the lowest is set to 129 and highest to -60 or -30 depending on the c or f node so that any temp has an extreme to compare to at first run after the reset.

@tmorse305 I'm running all nodes with no errors, mind boggled. @Teken @gviliunas v1.0.17 adds some logger info to help sort out what is going on. Update, run for a few minutes and give me some log info. Also can you tell me what Polyglot version you are on, Thanks. If I understand correctly, the values are updating properly but there is still an error reported, is this correct?

@Teken @gviliunas I tried everything I could think of to crash one of my polyglot and no luck, even non-existent variable IDs passed. So all I can figure is there some latency in the system. My ISY is plugged into the router so that may be why. I added and error handler that should capture errors and not let the nodeserver crash. I can't test it cause I can't get an error so you guys are the testers. I added another Custom Configuration Parameter: key - parseDelay where you can add a pause, in seconds, to let the data have time to populate. I would start at .1 and work up. I set mine at 1 just to see what happens. Let me know how it works for you and log bits of errors will be helpful. Thanks for helping out! I really would like to put the Temperature Nodes to bed so I can start on the Energy Node. v1.0.16 has been triggered, no update profile, just update and restart.

@Teken @gviliunas How many pulls are you guys doing. I am not getting any errors so it must be an overloading. I will insert a tiny wait like Teken mentioned and see if that clears the bug.

@gviliunas Good log info! The error got kicked down the road, now I can resolve this. The data did pull correctly so it must be a timing thing. Back at my desk in 3 hours. Should not take long, way too bouncy to try a fix on the laptop.

@Teken @gviliunas I removed the offending line and triggered an update. I have no way to test so update and let me know. No need for a profile update just a nodeserver restart. That may move the error further down the code or it may have been the LOGGER printing before the return values were set.

I don't know why that is happening but I do know how to stop it. Got held up on an IT install, on the road home. finally. I'll try to do the fix en route and update.

@Teken I can see that the problem is one of the nodes is pulling from a variable does not exist. That will get fixed. I can't dive in any deeper for a couple of hours but it’s nothing serious.

Prec 0 or 1, it can handle both. Those button have given me problems on all nodeservers. Setting the value from the drop down should be enough to set the value and it should change in the panel above. Hitting the button should only resend the set value command.Your confirmation is the proper setting in the reporting panel, that's why I added them. You may or may not see the Busy flag pop on. I have no limits applied and have gotten no answer as to if there is a limit of nodes or node types allowed and I have asked. I suppose you could have so many nodes pushing and pulling that it could bog down the polyglot with all of that activity but it seems unlikely on Polisy as that thing has some power. I test everything on 2 RPi polyglot and 1 Polisy and everything is working properly. Raw to Prec seems to be needed on any variable that is set the Prec 1. Other things: Make sure the variable exists that you are pulling from, I will add a safety check to ensure that is doesn't try to update from a non-existent variable. You can only pull from 1 variable in each node, I will add a safety check to disallow 2 pull actions. If you send me a log dump I will look through it and see if things are working properly, no sensitive data is in the logs. Redact anything you think necessary and DM or email markv58git@gmail.com. Teken, thanks for all your input and driving the Temperature nodes to this point! The Energy node may possibly give you just as much grief but hopefully not. I plan to get the whole thing ready and tested before release so maybe the bugs will be less and there will be only a couple of profile updates.

Be absolutely certain that the variable you are pulling from exists! If it does not exist there will most likely be an error that will crash the nodeserver. I will add a routine to check that the variable is valid when I get back to my desk tomorrow evening and run some tests to see what error I need to look for.

@Teken Try this, go to the Virtual folder and delete all .db files in the folder then restart the nodeserver. This shouldn't affect any of the settings in the nodes. Restart the nodeserver.

@Teken Maybe something glitched during the update. Do a git pull in the virtual folder and restart.

I have joy joy at my place.

And set your Raw to Prec