Toggle options like --no-multi-line or --no-hscroll

[amariusz]

I wonder if it's possible to change (or even better to toggle) --no-multi-line or --no-hscroll without restarting fzf. I'm using --read0 to obtain multi-line list and sometimes, depending on query results, it would be convenient to toggle multiline on and off (along with no-hscroll).

I've seen some examples using transform but there are no actions related to these command line options.

Is there any other way?

[junegunn]

No, there currently isn't a way. But adding it would be trivial. What is the exact use case?

[amariusz]

Thank you for replying! That's good to know that internal logic is not preventing such behaviour after fzf is executed :)

My usecase is similar to your example:

declare -f |
  perl -0 -pe 's/^}\n/}\n\0/gm' |
  bat --plain --language bash --color always |
  fzf --read0 --ansi --layout reverse --multi --highlight-line

The length of my input entries (basically text documents) varies greatly - from 5 to 1000 lines. Most of them are short and browsing with fzf list is enough. But depending on the query I may end up with those bulky ones. They don't fit on the screen, occupies whole list and most of the time the match is not visible anyway. In such cases I prefer to 'fold' the document on the list to singe line with --no-multi-line, prevent it from scrolling with --no-hscroll (because first line has some crucial metadata like creation date), and use preview to browse the document.

Currently I need to restart fzf with different options for such cases but using shortcut(s) to toggle them when needed would more comfortable.

[amariusz]

Something related I was thinking about - would it be also trivial to toggle --read0 option this way?

The use case would be 2-stage filtering, similar to your example here: https://junegunn.github.io/fzf/tips/ripgrep-integration/
First stage is filtering whole 'documents' with --read0 - for example only those that have (or don't have) specific #hashtag. Then switching modes with something like ctrl-t which also turns off --read0 and filtering lines within those documents with separate query.

This is already possible with another very useful approach - select all matching entries ('documents'), execute vim with {+f} and use BLines to perform second stage filtering.

[junegunn]

No. It's very different from toggle-multi-line which only affects the user interface. --read0 is used to control how fzf parses the input stream. Once read, changing the option afterwards won't affect the items already read. To change the list, you have to use reload.

[amariusz]

You're right, this is important distiction.
I was actually thinking about using reload. Sorry for linking wrong example - I've realized there are few examples of rg integration in your documentation. This is the one I had in mind: https://github.com/junegunn/fzf/blob/master/ADVANCED.md#switching-between-ripgrep-mode-and-fzf-mode

I guess this could probably work if instead of rg, another fzf process with --read0 --print0 and --filter would be used to do the first stage filtering (whole documents) and then after pressing ctrl-f (and enabling search) line filtering would be in effect.

[junegunn]

See ca5e633 and e60a9a6.

[amariusz]

Thank you very much! Changes seem OK to me :)