Bind SPC to minibuffer-complete-word

[kaushalmodi]

• This binding will make the executions of commands like
  counsel-describe-variable, counsel-describe-function,
  counsel-M-x, etc magical!

• You will be able to use SPC to autocomplete the
  variable/function/command intelligently (just as it does in the
  emacs inbuilt execute-extended-command.

  Example: list-c SPC will autocomplete that to list-colors-display when
  executing counsel-M-x

[abo-abo]

Sorry, but I don't like this change. SPC should insert a space. This is part of the experience that I want Ivy to have: it should behave like a normal buffer if possible. So SPC should insert a space, like in a normal buffer; and C-w should call kill-region, like a normal buffer etc.

We can bind minibuffer-complete-word to C-z w if you wish. I think C-z could become a prefix for a lot of commands.

As for list-c SPC, you can get a similar behavior with ^list-c TAB. One problem with minibuffer-complete-word is that it automatically assumes that ^ is there. Feels not consistent.

Another idea is to make counsel-M-x auto-insert the initial ^. You could delete it if you want with DEL.

[abo-abo]

Sorry, but I don't like this change. SPC should insert a space. This is part of the experience that I want Ivy to have: it should behave like a normal buffer if possible. So SPC should insert a space, like in a normal buffer; and C-w should call kill-region, like a normal buffer etc.

We can bind minibuffer-complete-word to C-z w if you wish. I think C-z could become a prefix for a lot of commands.

As for list-c SPC, you can get a similar behavior with ^list-c TAB. One problem with minibuffer-complete-word is that it automatically assumes that ^ is there. Feels not consistent.

Another idea is to make counsel-M-x auto-insert the initial ^. You could delete it if you want with DEL.

[kaushalmodi]

I agree with your vision for the use of SPC key. I have bound that way in my init and it works fine because I use the fuzzy search regexp function.

It's alright that the counsel-M-x behavior stays the way it is (without autoinserting ^) because that will feel more than a patchwork and move away from the consistency with other functions like counsel-describe-function, etc

[kaushalmodi]

I agree with your vision for the use of SPC key. I have bound that way in my init and it works fine because I use the fuzzy search regexp function.

It's alright that the counsel-M-x behavior stays the way it is (without autoinserting ^) because that will feel more than a patchwork and move away from the consistency with other functions like counsel-describe-function, etc

[abo-abo]

It's alright that the counsel-M-x behavior stays the way it is (without autoinserting ^)

Actually, as I'm using it today, it works great. Auto-adding ^ brings back the feeling of old school non-ido completion. For me, 95% of the time, it doesn't matter that input starts with ^, but I get less candidates as a bonus, and usually there's no need to use arrows. And if I need more candidates, it's a simple C-a C-d to remove ^.

You should try it for a while. I was skeptical initially as well.

[abo-abo]

It's alright that the counsel-M-x behavior stays the way it is (without autoinserting ^)

Actually, as I'm using it today, it works great. Auto-adding ^ brings back the feeling of old school non-ido completion. For me, 95% of the time, it doesn't matter that input starts with ^, but I get less candidates as a bonus, and usually there's no need to use arrows. And if I need more candidates, it's a simple C-a C-d to remove ^.

You should try it for a while. I was skeptical initially as well.