Time precision loss in first LagFilterFitPoints data chunks

[lenazh]

In some magnetic field data files the first LagFilterFitPoints chunks have wrong timing. The dt between the chunk arrival will appear to grow linearly with the chunk offset, then after LagFilterFitPoints'th point it resets to the correct value. The deviation is within 500 microseconds, but is still very noticeable and needs to be fixed. This occurs in roughly half of the samples. I don't know why this happens, but I suspect racing between the LegacyStorage and the thread filtering the data. Which is weird because they are supposedly in the same thread.

[lenazh]

This is caused by a race condition in the data file writer class; did Vincent fix it?

[lenazh]

Paper has been published, data file writer should be thrown away if you're using this code for anything