Change in `M`

Change in M: with individual addressing, too large an input now gives no output, instead of causing an error

Extension to `Y$`. Correction to `Zc`

• Y$ (vpa) with negative second input gives output without rounding the last decimals. First input must be a real or imaginary scalar

• Zc with single input now allows sparse input in MATLAB (previously it only worked in Octave)

Correction to previous release

Ye in/out spec and help seem not to have been updated properly in release 20.8.1. This release corrects that.
(One of these days I should learn to use Git / Github properly).

Generalized `Ye` funcion

Ye with logical input now makes the matrix square if needed. If there is a second input it specifies exponent for power graph

Function `e` extended. Octave bug that affected function `z` corrected

• Empty second input in e specifies square matrix
• Octave bug that affected z with range input corrected. Thanks to @Steabybox for finding it out.

Correction to parser

• r (rand) and Xr (randn) can now have 0 outputs. This is useful to specify the seed of the RNG. Suggested by @flawr

• Correction to parser. It incorrectly identified ] or } within strings as closing symbols for array literals. Thanks to @Sanchises for noticing!

Various improvements

• Refactored setxor_comp (which patches setxor for Octave compatibility)

• Small correction in setdiff_comp (which patches setdiff for Octave compatibility)

• Octave compatibility: mode now accepts char input

• Replace 'MATLAB' by 'Octave' if appropriate in run-time error messages

• \ now accepts complex first input, and operates independently on its real and imaginary parts

Correction to 20.5.0

Octave compatibility for c (round always down)

Octave compatibility: patched a bug in ismember function with complex inputs

Generalized Y^ (matrix power) as suggested by @Sanchises: compute sum of matrix powers if there are more than one exponent in second input

Ye: matrix exponential / "infinite" graph power, logical version (see help / documentation). Thanks @Sanchises for the idea and discussion

An Octave bug (https://savannah.gnu.org/bugs/?52542) that affected X~ has been patched. Thanks to @Sanchises for noticing it

Matrix exponential and others

Octave compatibility for c (round always down)

Octave compatibility: patched a bug in ismember function with complex inputs

Generalized Y^ (matrix power) as suggested by @Sanchises: compute sum of matrix powers if there are more than one exponent in second input

Ye: matrix exponential / "infinite" graph power, logical version (see help / documentation). Thanks @Sanchises for the idea and discussion

An Octave bug (https://savannah.gnu.org/bugs/?52542) that affected X~ has been patched. Thanks to @Sanchises for noticing it

FFT-related functions added

• Bug corrected in &Y@. Thanks to @Sanchises!

• fft (FFT), fftn (n-dimensional FFT) added. Including Octave compatibility for fftn when the second input contains zeros

• fftshift added.