GitLab

anyextract did not check plausibility of selected time window. Let the
program abort with an error message, if time window is empty.

binary executables of tests remain in this subdirectory and should be
ignored

specify the tidal constituents used for tide removal in a source code
comment

- let deconv provide tolerance with respect to sampling interval
- fix comment in teseco

randomseries allows to produce several incoherent test series without
setting a seed value individually. This is much more straight-foward.

revision of tide removal code:
- both versions (C and Fortran): do case of nstep=1 explicitly
  seife by default averages over serval samples and uses the tidal
  hamonic as well as its derivatives for corrections; this appears to
  have saved computation time in the very early times but is of no
  relevance nowadays; by not averaging the code (and the computation)
  becomes much more straight-forward and less error-prone
- C-version: add M3 frequency
  The number of frequencies was fixed to 6 in the original version; to
  allow for the addition of frequencies, the dimensions of arrays now
  are det by C preprocessor macros; the function seife_gauss is modified
  accordingly to allow for a larger system of equations

Originally any2matlab converted the file format string (format ID and
subformat specifications) to lower case. This appears unnecessary and in
fact results in failing code, if for example the modifier
'estimateNframes' has to be passed with 'mseed'. There currently is no
obvious reason, why case conversion should be done in any2matlab. If it
should turn out that existing matlab code relies on any2matlab doing
this conversion, the matlab code has to be fixed, doing the conversion
in matlab before calling any2matlab.