[teqc] Decimating high rate data on linux 2005Dec16 release

[Lou Estey]

> Yes, a known problem with that release due to a change I put
> in on 15 Dec.  That version will decimate up to the first
> millisecond reset in the rx time tag and then stop.
> If you are starting with raw data, the workaround is to
> translate with `+smtt` and then the decimation will work
> fine.  If you only have RINEX, then you have to backtrack
> to an earlier version of teqc.

My mistake: the code change for this bit was on 28 Nov, and it
was tested on just about everything ... except RINEX with
millisecond resets in the time tags.

At the UNAVCO Facility archive, we are migrating away from
RINEX with millisecond resets in the time tags and smooth
phase/pseudorange observables, opting instead for `teqc +smtt`
generated RINEX with smooth time tags and millisecond (equivalent)
jumps in the phase/pseudorange observables.  The latter is
just easier to work with for many reasons.

--lou