Creating time of the yesterday.ini

[nitrx]

Steve I wondered why the yesterday values are created after midnight ( I think about 01:00 hours) on my webpage the wrong values are shown for yesterday these are the values between midnight and 01:00 today

A zip of the diag is attached , latest .log of the diags map I've debugging not turned on.

[steve]

The clock on your computer went backwards at 01:00

21-2-2010 00:56:39.321 : Writing todayfile, LastUpdateTime = 21-2-2010 0:50:00 raindaystart = 74,0999984741211
21-2-2010 00:59:49.481 : Writing todayfile, LastUpdateTime = 21-2-2010 0:50:00 raindaystart = 74,0999984741211
21-2-2010 01:00:00.373 : Writing todayfile, LastUpdateTime = 21-2-2010 1:00:00 raindaystart = 74,0999984741211
21-2-2010 00:59:34.838 : === Day reset for 21-2-2010 0:59:34
21-2-2010 00:59:34.842 : Raintotal = 74,0999984741211 Raindaystart = 74,0999984741211
21-2-2010 00:59:34.842 : Rainyesterday (calibrated) set to 0
21-2-2010 00:59:34.842 : Raindaystart set to 74,0999984741211
21-2-2010 00:59:34.863 : Writing todayfile, LastUpdateTime = 21-2-2010 1:00:00 raindaystart = 74,0999984741211
21-2-2010 00:59:34.877 : === Day reset complete
21-2-2010 00:59:36.755 : Writing todayfile, LastUpdateTime = 21-2-2010 1:00:00 raindaystart = 74,0999984741211
21-2-2010 00:59:36.764 : Writing todayfile, LastUpdateTime = 21-2-2010 1:00:00 raindaystart = 74,0999984741211
21-2-2010 00:59:36.772 : Writing todayfile, LastUpdateTime = 21-2-2010 1:00:00 raindaystart = 74,0999984741211

I guess this is the time you have it set to correct the clock from the internet or wherever? This causes Cumulus a bit of a problem, as it thinks it has to do the day reset again, as the hour has become zero. This is a bit tricky to fix; the rollover code is quite complicated already. Would it be possible for you to change it so that your clock gets corrected at 02:00 or later?

[nitrx]

My clock is synchronised by the timeserver from time.windows.com there seems not to be an option to set the interval for setting the update of the time in Win 7 for now I put the syncronitisation off and will do this manually thanks for investigation.

[steve]

there seems not to be an option to set the interval for setting the update of the time in Win 7

I hadn't noticed that; that's a nuisance.

[nitrx]

I'll write Microsoft for an update

[steve]

It looks like it does it once a week. If you force an update at a convenient time, it seems it will do the next update exactly one week later. I've just done mine and it says it will next update at 13:04 on 28/02/10.

[nitrx]

That's the solution , I hope , I did updat it now and it says the next synchronisation wil be 28th of febr 14:17 so I keep an eye on it. Do you expect problems at the end of march concerning daylight savings ?

[steve]

Do you expect problems at the end of march concerning daylight savings ?

It usually works OK. What happens exactly depends on the type of station and whether you have Cumulus running at the time the clocks change. The simple case is where Cumulus is actually running; you just get an apparent gap in your data between 01:00 and 02:00. If Cumulus wasn't running, what you get depends on whether the station timestamps logger entries (Fine Offset doesn't), when you stop and start Cumulus, when exactly the PC clock and the station clock changed etc etc. It's not ideal but DST is such a stupid concept anyway. Cumulus 2 uses UTC.

[nitrx]

I get virtual data lost well we see what happens UTC should be better for these things and some countries have complete other daylight savings or don't have it. I check out how official meteostations deal with it

[beteljuice]

I check out how official meteostations deal with it

Modern history dictates that ALL meto info is Zulu (UT) reporting. Originally navy, now METAR / TAF (airport type data).