Zero for today's low humidity/wind chill/apparent temp

[steve]

There seems to a problem where sometimes some of today's "lows" get set to zero a few minutes after midnight, and hence, depending on your real conditions, some or all of these don't get replaced with correct values later in the day. I thought that I could reproduce the problem, as I saw this in my today.ini file, but I've realised that I can't, and what I saw was because I had been using someone else's today.ini file, investigating problems with the beta.

I'm pretty sure this is nothing to with the 'this month/year highs and lows' problem, so I've started a new thread. I think the problem is related to a particular pattern of usage (i.e. the times when you typically start and stop Cumulus), and I think I already know how to make the problem go away, but I'd like to understand exactly what causes it. So, if you have this problem right now, please zip up your diags folder and your today.ini, and attach here.

[Ned]

Here's mine

[steve]

I can't see which of the values is wrong?

[Ned]

Sorry, misread your post, still thinking about the years's high/low problem

[steve]

Sorry, misread your post, still thinking about the years's high/low problem

You still have a problem there? I thought that was all resolved now?

[Synewave]

Steve,

I seem to have a 0 for yesterday:

Humidity: 0% at 00:03 from my http://www.synewave.co.uk/gcsweather/todayyesterday.php page.

Do you want me to send anything?

[steve]

Do you want me to send anything?

If your diags files go back far enough to cover the period over midnight 1st/2nd Nov, then they will be useful.

[Synewave]

Unfortunately not. Interesting though as the webtag I use for that value is <#humYL>, i.e. not one of the new ones.

[Synewave]

Actually, I found a backup!

01/11/2010 23:50:00.669 : Writing todayfile, LastUpdateTime = 01/11/2010 23:50:00 raindaystart = 1094.09997558594
02/11/2010 00:00:00.190 : === Day reset for 02/11/2010
02/11/2010 00:00:00.212 : Error writing dayfile.txt: I/O error 32
02/11/2010 00:00:00.212 : Copy the following data into dayfile.txt
02/11/2010 00:00:00.212 :
02/11/2010 00:00:00.224 : Error closing dayfile.txt: I/O error 103
02/11/2010 01:32:55.427 : Writing todayfile, LastUpdateTime = 01/11/2010 23:50:00 raindaystart = 1094.09997558594

Here's the actual file:

[steve]

Ah, excellent. Looks like it's simply a knock-on from one of the fixed bugs. Keep an eye out for it (the zero value) happening again, though.

[Greebo]

I just restarted Cumulus this morning (8:15 am) and noticed that my wind chill all time record is now 0.0 again (and appears blank in the all time record editor as posted in my other thread)
My todays minimum apparent temp is now 11.5 too and it should be 15.2...

here's the last few lines from the alltimelog.txt:
2010-11-04 06:28 15.100 "Low wind chill" 2010-11-04 06:14 15.200
2010-11-04 06:56 0.510 "High rain rate" 2010-11-03 10:57 0.500
2010-11-04 06:57 0.510 "High hourly rain" 2010-11-03 10:37 0.500
2010-11-04 07:08 15.000 "Low wind chill" 2010-11-04 06:28 15.100
2010-11-04 07:19 15.264 "Low apparent temp" 2010-11-03 19:35 15.300
2010-11-04 07:19 14.900 "Low wind chill" 2010-11-04 07:08 15.000
2010-11-04 07:19 15.241 "Low apparent temp" 2010-11-04 07:19 15.264
2010-11-04 07:19 15.234 "Low apparent temp" 2010-11-04 07:19 15.241
2010-11-04 08:15 0.000 "Low wind chill" 2010-11-04 07:19 14.900

relevant diag files attached.

[Synewave]

Ah, excellent. Looks like it's simply a knock-on from one of the fixed bugs. Keep an eye out for it (the zero value) happening again, though.

Thanks Steve, this morning it's corrected itself. I'll keep an eye on it.

[steve]

I just restarted Cumulus this morning (8:15 am) and noticed that my wind chill all time record is now 0.0 again (and appears blank in the all time record editor as posted in my other thread)

There was a problem with the messaging between the data reading thread and the main thread at about the time of your zero windchill; it's likely this caused the zero value:

4/11/2010 08:15:06.675 : Error in thread messaging: Access violation at address 00A96DB8 in module 'cumulus.exe'. Read of address 00000008

I have no idea about the apparent temperature, there's just not enough for me to go on.

I've fixed the all-time editor problem.

[Greebo]

Thanks, I'll see how the windchill goes...