Backtracking parser

[lordmauve]

Adventurelib's command parser is inefficient when there are 2+ wildcards in a command. It tries all possible variations of assignment to each wildcard until one matches.

A more efficient parser would greedily match words until a constraint is violated and backtrack to see if any word in the input could fix it. This means it can fail fast if no such word exists.

[adrianogil]

Maybe parse module could help: https://pypi.org/project/parse/

[lordmauve]

I don't think so. Both regular expressions and the parse package are intended to match character-by-character. We don't need to do that because commands are defined as being word-split. So it's just a case of applying the same backtracking that regex/parse would do, but on a word-by-word basis.

Additionally, adventurelib is a one-file module that has no dependencies outside the standard library, which is a pretty nice property for beginners/educational contexts. I'd like to preserve that property if possible.