modify so gauge values always written at start time regardless of min_time_increment

[rjleveque]

Previously if min_time_increment > 0 was set for a gauge it wouldn't capture the gauge value until this long after the starting time value t1 specified for the gauge.

[mandli]

This should also be added to AMRClaw I suppose.

I was also trying to think of a way to do this without having to loop through the gauges twice but different types of data seems to make that difficult without looping twice.

[rjleveque]

By the way, I also tried to make it print the gauges at the final time in the calculation, since currently the gauge output is done before the call to stepgrid on each grid. But it's done that way so that ghost cell values have been filled in already, which in general are needed if interpolation is used to get the gauge value. (In practice we try to insure that gauges are at the centers of grid cells on the finest level but that's not always practical.). Anyway, I don't see any easy to fix this.