Thanks. Why is this better?

Thanks. Why is this better?

Hey :)

• The shbang is useless and the top helps with docs
• It keeps some bash features for completion, like it gives some autocompletion in some way
• Ex notation is recognised by a lot more editors
• Sometimes completion might fail and you might get an error

Ok. What do we gain from this:

-o bashdefault -o default

and more importantly, could this be NOT desired in some cases?
I do not see it in any other completion script (while I do see the other proposed changes).

Ok. What do we gain from this:

-o bashdefault -o default

and more importantly, could this be NOT desired in some cases? I do not see it in any other completion script (while I do see the other proposed changes).

The always complete thing

The always complete thing

I don't understand what that means. What behavior does it add exactly.

Also this line:

 _init_completion -s || return

breaks the tests. Specifically this test:

https://github.com/DannyBen/completely/blob/8819f5610618090df754175668577b1971d1b37c/spec/completely/integration_spec.rb#L12-L16

The result is an empty array. Also tried defining the locals before the _init_completion, but same.

I think I found the problem. Need to source source /usr/share/bash-completion/bash_completion in the test script.

I will probably create a new PR with the fixed tests to supersede this one, but unless there is a good reason to keep thises -o bashdefault -o default, I might remove them.

I am closing this in favor of #13 - let's continue discussion there.