I'm working on task #16096 currently, so this task is open for someone to take up !

Currently, Arithmetic needs its input files to be listed on the command-line within the Reverse polish notation. This is easy to use when the number of files is limited.

However, when there are hundreds/thousands of files to stack (for example with the 'sigclip-mean'), it can become inconvenient. Especially when the full path of each file needs to be mentioned. In some cases, this can even exceed the maximum allowed space for a command (which you can check with 'getconf ARG_MAX', on GNU Bash, its 2097152 bytes).

Of course, users can easily define shell hacks (like using shell variables within a loop), but it is more convenient if a list of files (with the number of files) can be loaded directly.

On the other hand, all the operators that take many inputs, first need to "pop" the number of inputs, then start loading them.

So to simplify the process, we can have a new operator in Arithmetic that will load a plain-text file (with file-names inside of it: one file per line), count how many non-commented lines there are, and then load all the images (in the given order) onto Arithmetic's Reverse-polish stack. And on top of the stack, it will place the total number of files.

For example if someone wants to stack all the FITS images in 'images.txt' by using the sigma-clipped mean of each pixel, they can simply use the command below (assuming the operator is called 'loadlist'):

astarithmetic images.txt loadlist 5 0.2 sigclip-mean

This task was suggested by Jesús Varela.