Adding a Directory Iterator

[DavidJAKoogler]

HAC is a much better scripting tool than Bash shell (and I like it better than Python scripts). Ada scripts are much more readable and intuitive than the strange hieroglyphics of Bash and sometimes strange side-effects of Python. I have written a dozen little HAC scripts so far and am pleased with the results. However, HAC's directory services are some what week and the lack of an iteration mechanism over a directory is a drawback for writing significant scripts.

I believe the underlying Ada.Directory package does have an iterator or a Read_Directory routine. There are also several useful path name routines in the package that should be exposed to HAC via the HAT package.

Motivation: try writing a script to apply a transformation to a set of files found in a directory or subdirectory(ies). Best I can do with HAC is to play around with the system operator and call upon the Unix ls command and parse the results. It is slow and painful.

[zertovitch]

Thanks for the feedback. I had in a corner of my head this feature gap (because I precisely miss often an iterator). So I will put it more on the top of the priority list...
Another cool aspect of using HAC as a scripting tool is the portability: people can use the same HAC script/program for Windows, Linux, and other platforms.

[DavidJAKoogler]

I am glad my suggestion aligns with your earlier observations. Directory iteration has always been odd under Unix/Linux operating systems, and yet you need this facility when doing things like searching for specific files. With HAC's better syntax, it should be much easier to write some of the complex shell scripts in a much more understandable form. While you are looking at Ada.Directory, you may want to add some of the path manipulation and parsing functions (basename, parent, extension, and so forth). Again, these are common operations you use in shell scripts. I know you can write most of these functions in HAC itself, but given that HAS is already using Ada.Directory and these functions are known to work in the package, it makes more sense to me to just expose those functions through HAC rather than re implement them in HAC itself. If you do much work with Python, you might find the Python PathLib functionality to be better than Ada.Directory and more appropriate for HAC as a scripting tool. I have noticed HAC's portability and that is a big win. Bash does not have a native implementation under Windows. Using MSYS or Mingwin is cheating :-) Dave Koogler