Exa: Offer official aliases

[clarfonthey]

As discussed in #195, a lot of people use aliases for exa similar to those for ls. It'd be nice if we could offer official aliases for these (similar to ls, ll, la, and tree) so that users have access to these aliases without having to set up anything different in their shells.

Personally, I vote for xs, xl, xa, and maybe xt; most people seem to prefer these.

What makes the most sense is to simply parse the currently-skipped first argument in args_os(), using that to determine which alias is in use. That way, people can simply symlink exa to aliases with these names to make them work, and it allows distributing a single binary while still reaping these benefits.

This would also be a huge benefit over ls, as most people create aliases for ll and la when ls itself does not officially support these.

[ogham]

(I mixed up the labels because I misunderstood the request)

There’s some discussion over in #332 about ‘ls mode’, where we’d do one of the things you suggest: see if the binary is being invoked as ls, rather than exa, and offer a different command-line interface if it does.

This is a similar feature, except it only has to support exa’s current command-line interface, rather than inventing a new one!

However I’m not sure about this feature’s scope. I haven’t run a poll of how often these aliases are used (maybe I should); if it supports xl, no doubt there’d be a feature request to support exal or something similar immediately after. And if you want to have shell aliases, wouldn’t it be the same amount of work creating the symlink as it would creating the alias?

[hyiltiz]

x is harder to type than two characters ls.