Suspect bug <#RecentUV>
Posted: Tue 26 Nov 2013 9:38 am
I have found a problem with my version of Cumulus (1.9.4 build 1085)
I have created a big long webpage for the recent history as half hourly observations. (Even managed to get some small amount of JScript processing to convert the bearing to words direction eg NNW, thanks to the cumulus forum for the tips)
Part of this contains the webtag <#RecentUV> the issue is with a previous readings for that webtag, for example say <#RecentUV h=2 m=30> (2 hours and 30 min ago)
My process of events is as follows:
Cumulus processes the file and stores it locally (on my computer and not pushed to FTP every 10 minutes) Cumulus is set to update all web pages every 10min.
Cumulus toolbox then uploads it just after the half hour or hour time throughout the day. This way I get a half hourly obs page with the obs done at the right times. (not every 10 min)
What I have found is that when I manually choose "web update" in cumulus, the <#RecentUV h=2 m=30> gives the same data as <#RecentSolarRad h=2 m=30>
Now if I let Cumulus process automatically, it appears to work. That is, it shows the UV index value correctly.
Here is what I have found.
Checked the realtime data files and all looks Ok.
Cumulus graphs show fine. All other webtags related to UV index yield expected data.
If I run a test to a text file using Cumulus and just a sample spread of RecentUV times it gives the same result of showing the Solar Rad data values. When manually doing web update.
After letting it run for a while automatically (as described above), and getting good UV expected values, as soon as I choose Web update all the good UV index values are wiped and replaced with the solar Rad values.
Here is my test files output here
http://members.iinet.net.au/~nossis/wea ... output.txt
Here is the text file used to generate it the test
http://members.iinet.net.au/~nossis/wea ... s_test.txt
Here is my test obs page
http://members.iinet.net.au/~nossis/weather/obs.htm
To me the issue lies in how Cumulus is processing the web tag as everything else appears and behaves as excepted. And in saying that it is only when a "click on web update" is undertaken. If let it run normally as scheduled, then it shows good values. (well it did today until I ran a manual update and then in converted all my good values back to bad ones again)
So... looking forward to finding out if the issue is just with me or if it is in fact a replicable problem.
I have created a big long webpage for the recent history as half hourly observations. (Even managed to get some small amount of JScript processing to convert the bearing to words direction eg NNW, thanks to the cumulus forum for the tips)
Part of this contains the webtag <#RecentUV> the issue is with a previous readings for that webtag, for example say <#RecentUV h=2 m=30> (2 hours and 30 min ago)
My process of events is as follows:
Cumulus processes the file and stores it locally (on my computer and not pushed to FTP every 10 minutes) Cumulus is set to update all web pages every 10min.
Cumulus toolbox then uploads it just after the half hour or hour time throughout the day. This way I get a half hourly obs page with the obs done at the right times. (not every 10 min)
What I have found is that when I manually choose "web update" in cumulus, the <#RecentUV h=2 m=30> gives the same data as <#RecentSolarRad h=2 m=30>
Now if I let Cumulus process automatically, it appears to work. That is, it shows the UV index value correctly.
Here is what I have found.
Checked the realtime data files and all looks Ok.
Cumulus graphs show fine. All other webtags related to UV index yield expected data.
If I run a test to a text file using Cumulus and just a sample spread of RecentUV times it gives the same result of showing the Solar Rad data values. When manually doing web update.
After letting it run for a while automatically (as described above), and getting good UV expected values, as soon as I choose Web update all the good UV index values are wiped and replaced with the solar Rad values.
Here is my test files output here
http://members.iinet.net.au/~nossis/wea ... output.txt
Here is the text file used to generate it the test
http://members.iinet.net.au/~nossis/wea ... s_test.txt
Here is my test obs page
http://members.iinet.net.au/~nossis/weather/obs.htm
To me the issue lies in how Cumulus is processing the web tag as everything else appears and behaves as excepted. And in saying that it is only when a "click on web update" is undertaken. If let it run normally as scheduled, then it shows good values. (well it did today until I ran a manual update and then in converted all my good values back to bad ones again)
So... looking forward to finding out if the issue is just with me or if it is in fact a replicable problem.