Hacker News Reader

Comment by TylerE | Hacker News Reader

I can’t see how that would handle any of the hard bits (like dependencies between arguments, mutual exclusion)...

To the extent the command-line program returns proper exit codes, or gives other hints via its output/side-effects as to whether it did anything useful, I could see it working up-to-a-point.

Discovering exclusive options, or options that must be specified together, sound like exactly the sorts of things it could deduce, when every invocation with/without certain pairs of options fails.

Perhaps wrapper-utilities like 'Gooey' or some sort of automated-probing could encourage a greater formalization for specifying de-facto command-line APIs. Then, command-line programs might want to offer not just the human-interpretable help texts and docs, but more rigorous usage-specs that could drive automatic wrappers.

Probing at options sounds dangerous. There is no guarantee at all that attempted invocations won't change state.

I wouldn't wanna let it probe 'rm' :D

Perhaps the probing could happen in a VM / container image which would reset after each invocation.

You'd of course do it inside a controlled environment, via virtualization or similar technologies which let you monitor, veto, or roll-back anything the analyzed program does. (And if it's malicious/sneaky enough to escape that, then it wasn't safe to use in a manual fashion, either.)

Reply on news.ycombinator.com