TAB completion not showing complete path about z HOT 8 CLOSED

Hi,

This is definitely not intended, but I can reproduce this as well.

However, it doesn't look like something I can easily fix. The code has already been written to generate the completions once, after the function definition: https://github.com/jethrokuan/z/blob/master/conf.d/z.fish#L40

Completions do indeed work after manually running __z_complete. This seems to be a situation where the lines here:

https://github.com/jethrokuan/z/blob/master/conf.d/z.fish#L23-L27

takes a while to define the function, and completes only after __z_complete is run, causing the completions to be produced before the function is defined and are hence invalid.

from z.

I'm not really familiar with Fish code, so I'm not sure how to help here. You can count me in for tests, though. :)

from z.

what I was trying to get is that there's nothing much I can do here, the behaviour we're seeing is (I'm guessing) a result of some race condition that arises from how fish evaluates the code.

from z.

Probably worth mentioning @faho then, and see if he's got some clue about this.

from z.

Simple: There is already a thing called "j", and fish ships completions for it.

Since z does not ship a completion file, it'll still autoload that one, and that includes:

complete -c j -e

which will erase all completions for the "j" command.

from z.

Thanks @faho, I confirm that using the default z instead of j works as expected.

from z.

Thanks @faho, closing as it's resolved

from z.

Well, actually I'd like to have j working, but it's ok. :)

from z.