edit: Command line questions

links

history

item nametags

- Is there a document somewhere that describes command line principles?
- How should command line info be stored in the executable?
  - Perhaps an array like {short_switch, long_switch, data_type, short_description, long_description, default_value, sample_value} ?
- What should the standard switches be?  (see the [[http://www.gnu.org/software/coreutils/manual/html_mono/coreutils.html#Common-options:GNU CoreUtils Common Options]] for examples)
  - One common option could be --explain, which will take a given command string and list each argument along with its [calculated] value and the short_description from above.
  - Another could be --noaction or --donothing, where validation routines are run on the arguments but no actual changes are made.
- Might intellisense for a shell be useful, if parameter info is stored as described above?
- What is the standard for passing a list of filenames to a CLI app which are more than a simple "*.txt"?  stdin?
- standard ways of getting data in and out of a CLI program?
  - first we have the pipeline, but that typically allows only one input stream and two output streams (stdout and stderr)
  - backticks for executing commands and dumping their contents
    - insane escaping when one nests these

==Notes
- [[http://argtable.sourceforge.net/:argtable]] is a C utility to parse GNU-style command line arguments.

=Components
- aliasing of commands
- wildard matching
- start other programs easily
- command history retrieval and modification
- handling of multiple line input
- redirection to file (>, >>, stdout & stderr)

=Structure
There are switch arguments, like `-file a.txt`, as well as positional arguments, which either immediately follow a program name or otherwise aren't valid switch arguments (any the switch already would already have an argument).

some permissive license goes here
contact

tsunami