teem	/	hest	Comments, other information

Computer science graduate students should not be spending their time writing libraries for command-line parsing. I spent the time to write this because I couldn't find a simple C library with the features I was after: flagged and unflagged, variable and fixed arguments of many different types, with response files, and automatic generation of usage and option information. The primary motivation for this was to provide a consistent command-line interface for all the unrrdu tools. I can only hope that other people can find some use for this.

The current hest implementation has some characteristics which should probably be considered short-comings, but in the interests of time and simplicity I've decided to document them, rather than fix them. In order of decreasing importance:

• String parameters shouldn't contain space, because the simple methods of parsing the parameters into options can get confused.
• Only one level of response file expansion is supported: response files themselves can't name more response files in which to find further arguments.
• There is no simple way for the hestParse() caller to know what string or strings were parsed in order to specify the value for an option.
• Only one one unflagged variable parameter option. One could theoretically deal with a sequence of unflagged single variable parameter options at the very end of the command line (with usage info like "[<a> [<b> [<c>]]]"), but this requires more complicated logic.

Like much of teem, hest has been purified to minimize the chance of memory leaks, which is important given how much dynamic memory allocation it does.