                 Changes in version 0.2.3 (2026-07-27)                  

  - Fix: cwt() incorrectly flattened matrix/data.table input into a
    single row because is.numeric() also returns TRUE for numeric
    matrices, causing the result to be returned as a plain vector
    instead of a data.table.
  - Fix: cwt() with summed_wavelet = TRUE collapsed the result to a
    plain vector whenever the input had a single row or a single scale,
    even when the input was a genuine matrix/data.table, due to apply()
    silently dropping dimensions.
  - Fix: cwt() accepted scales = 0 and variance = 0, which silently
    produced Inf/NaN values instead of raising the documented error.
  - Fix: resampling_FWHM() silently returned 0 for new bands whose FWHM
    window fell outside the range of wavelengths, indistinguishable from
    a real zero reading. These bands are now returned as NA.

                         Changes in version 0.2                         

  - Fix: Magnitude of response by variance in cwt.
  - Add a function for fast resampling using FWHM.

                 Changes in version 0.1.1 (2024-05-28)                  

  - First submission to CRAN
