named queue stops working?

[mikeporter]

I guess the thing I don't understand is why you are using queues in the first place. This is a pretty straight-forward low speed data logging application. Why bother with queues when you can just write the data out to some sort of temporary file or database? When you have a measurement process that's going to take a long time you don't want to hold all the data in memory because it's too easily lost.

Mike...

[Ha-Jo]

Hello Mike,
thank you for answering to my message. I didn't keep much data in the queue, but during water measurements it is essential to compensate some time, the water needs to flow through pipes from outside the ship to the setup. Along this way the temperature of the probe is changing. To calculate some gas saturations in an equilibration prozess accurate, the scientist, who developt the set-up, needs to know the original temperature of the probe. As a compromise he wants to be able to change the delay in units of the measurement cycle, which is one minute. Actually two readings are hold for two cycles each, to compensate about two minutes the probe needs on its way through a pipe.
I thought, it is a convenient way to use a queue as some sort of shiftregister with an adjustable length. I didn't expect any problems because the maximum needed length may be about 10 cycles and there are no special requirements for timing.
A third queue was used to collect strings from LabVIEW error messages around the VIs of the program to write them in a log file centrally. This queue is examined very often to dequeue any element in it immediately.
The program works fine on the first journey to the destination port but at the run back there were only calibration- an air-data stored in the files. My first idea was a bug in my program, but I couldn't find it. So I decided to put little strings in nearly all branches of the program flow to give me some hints about what really happens. Again I used a named queue to collect them before writing on hard disk. Now the stored water measurements are disappearing about 45 minutes before the mode changes from water to air measurement the first time and my new "InfoLog.txt"-file stops at the same time. Thats why I'm thinking, the problem is related to my queues, but honestly, I don't really understand it.
Tuesday night I will replace my "DataDelay"-queues by an array solution. Actual I'm not at work this week, but it's my last chance to provide a functioning program till next week. Then the scientist will be on board during the journey to make some fine-tuning with his instruments.

Ha-Jo