New error found with build 956

[Meteoalmoster]

Yesterday I installed build 956 and seemed to work fine, but I just went in front of my weather station dedicated PC (second thing after waking up, as usual ) and noticed that Cumulus stopped working some 6 hours ago with the following error message:



Not only this, but the error also seemed to somehow block the communications between the Davis console and Weatherlink, which I'm using in parallel with Cumulus thanks to VirtualVP 1.2.5. As soon as I restarted Cumulus, WL downloaded a bunch of data from the datalogger at once and continued to work normally.

As additional information, this kind of problem had never occurred with any of the previous versions of Cumulus I've used, including betas, and I've been working with this configuration for more than one year. In fact, as I mentioned yesterday, the first significant error I've ever faced with Cumulus came with build 955 and everything up to build 953 worked flawlessly.

Hope this helps identifying the problem, and in the mean time back to 1.9.0 . Thanks!

[steve]

Hope this helps identifying the problem

It's very little use at all, I'm afraid. For some reason the exception handler hasn't handled it, so I have no idea where this error occurred. I/O error 8 is "insufficient memory" and that address appears to be in some JPEG handling code, but I may have got that wrong, particularly as Cumulus doesn't use JPEGs.

[steve]

Actually, could you look in the Cumulus folder to see if a cumulus.elf file was created? The exception handler may have created one even though its dialogue box didn't appear.

[Meteoalmoster]

Hope this helps identifying the problem

It's very little use at all, I'm afraid. For some reason the exception handler hasn't handled it, so I have no idea where this error occurred. I/O error 8 is "insufficient memory" and that address appears to be in some JPEG handling code, but I may have got that wrong, particularly as Cumulus doesn't use JPEGs.

Steve, in fact there were two alert windows open and not only one, however the second was minimized at the bottom corner of the screen and I only saw it later, popping up for a moment when I was closing Cumulus down, therefore I could not capture it. However, I was able to read something about an "error getting database record" or something like this before it vanished.

Hope this does give you some clues... What about the fact of WL being also affected? It looks like Cumulus blocked the communications with the console, because WL was not reporting any error in its log and simply remained idle without even donloading data for several hours. Right when I closed Cumulus it started downloading again... 70 memory pages at once from the datalogger.

Well, I'm sure you'll figure it out and thank you once again for your response. Please let me know if I can be of any help.

Armand

[Meteoalmoster]

Actually, could you look in the Cumulus folder to see if a cumulus.elf file was created? The exception handler may have created one even though its dialogue box didn't appear.

Yes sir, there is such file. I'm trying to attach it (first time, hopefully I do it right ).

Armand

[steve]

Thank you; the error you reported isn't in there, unfortunately, but it seems it had a problem saving cumulus.ini file when it was closing:

2.1 Date : Sat, 23 Oct 2010 06:47:08 +0000
2.2 Address : 004D9678
2.3 Module Name : cumulus.exe
2.4 Module Version: 1.9.1.956
2.5 Type : EIniFileException
2.6 Message : Unable to write to C:\Cumulus\cumulus.ini.

I don't know whether this was related to the other problem, but it seems likely.