Add option to rm -rf irrespective of trash

[62832]

Heya. I was slightly annoyed by the fact that using NNN_TRASH completely disabled the ability to rm -rf within nnn, so I wanted to add a separate option for ^X that would just use rm -rf regardless of the trash backend being used.

Hope this is alright.

[KlzXS]

I think this is a fine addition. It always seemed redundant to me that there were two keybinds for the exact same thing bound to the exact same key.

Though it is a bit dangerous for the people who've grown accustomed to using ^X for deleting files with trash enabled. The last change (there have been additions since, but those wouldn't affect existing workflows) in keybinds was in April 2022 and that was much less impactful than this would be.

[N-R-K]

You can accomplish the same thing with a simple plugin. Bind it to x and then call it with alt+x. Don't think we need to have this builtin, and I especially don't think we should bind it to ^x and change existing bind.

[KlzXS]

On the other hand, I think this is a useful bultin. It's a bit odd that you can't specify to permanently delete a file if you configure trash. Maybe even going as far as to say that most file managers have such a feature. So, delegating it to a one-liner plugin seems a bit off.

I do agree about the keybind tough. I don't expect it to affect anybody realistically, but who knows. This would have been perfect back before keybinds were stabilized, but apparently no one thought of that at the time.

[KlzXS]

...unless we do something like with batch rename where we call a plugin if it's found and otherwise do the builtin thing.

It's a bit convoluted, but would at the same time keep the existing behavior for most users while allowing something extra for those willing to opt in. Also, there was some talk recently about supporting more trash managers. I didn't follow that through, but I imagine giving them a script they can edit would allow them better compatibility than just building all those options directly in.

What do you think?

[N-R-K]

do something like with batch rename where we call a plugin if it's found and otherwise do the builtin thing.

That might be a good idea. Would allow people to use other trash utilities or add special logic to skip trash and rm directly and what not.

[KlzXS]

@62832 the plugin in question is .nmv and the corresponding symbol in code is UTIL_NMV. In case you want to take a look at that.

[jarun]

@KlzXS there is one problem. For users who use the type-to-nav mode, ^X is the natural way to trash/delete. In my case, I still use rm -rf but someone used to trashing files would inadvertently remove the files.

So I am inclined to agree with @N-R-K on this aspect.

However, I think we can have the best of both the worlds. We can add a program option -X (capital-X) which would force rm -rf with ^X when a different trash is configured.

What do you think?

@62832 hold on till we conclude this discussion.

[jarun]

@62832 I am fine with a new keybind X for rm -rf always.

Also, please squash all the commits into one for easy maintenance.

[62832]

Should be done now, having addressed @N-R-K's concerns and switched to using SHIFT+X as the bind for rm -rf. Ready when you are.

[N-R-K]

Small documentation nit. Code seems ok to me.

[KlzXS]

If we are happy with the new names of the actions and the description in the help this should be good to go.

Also a reminder that we should update the wiki after this is merged.

[jarun]

We are good! I'll take care of the Wiki update. Thanks guys!

[62832]

Is that failing ubuntu-patches test going to be a problem?

[jarun]

./patches/check-patches.sh fails in CI. @62832 please have a look.

./patches/check-patches.sh
7
O_COLEMAK=1 O_GITSTATUS=0 O_NAMEFIRST=0 O_RESTOREPREVIEW=0 [FAILED]
8
O_COLEMAK=0 O_GITSTATUS=1 O_NAMEFIRST=0 O_RESTOREPREVIEW=0 [SUCCESS]
9
O_COLEMAK=1 O_GITSTATUS=1 O_NAMEFIRST=0 O_RESTOREPREVIEW=0 [FAILED]
10
O_COLEMAK=0 O_GITSTATUS=0 O_NAMEFIRST=1 O_RESTOREPREVIEW=0 [SUCCESS]
11
O_COLEMAK=1 O_GITSTATUS=0 O_NAMEFIRST=1 O_RESTOREPREVIEW=0 [FAILED]
12
O_COLEMAK=0 O_GITSTATUS=1 O_NAMEFIRST=1 O_RESTOREPREVIEW=0 [SUCCESS]
13
O_COLEMAK=1 O_GITSTATUS=1 O_NAMEFIRST=1 O_RESTOREPREVIEW=0 [FAILED]
14
O_COLEMAK=0 O_GITSTATUS=0 O_NAMEFIRST=0 O_RESTOREPREVIEW=1 [SUCCESS]
15
O_COLEMAK=1 O_GITSTATUS=0 O_NAMEFIRST=0 O_RESTOREPREVIEW=1 [FAILED]
16
O_COLEMAK=0 O_GITSTATUS=1 O_NAMEFIRST=0 O_RESTOREPREVIEW=1 [SUCCESS]
17
O_COLEMAK=1 O_GITSTATUS=1 O_NAMEFIRST=0 O_RESTOREPREVIEW=1 [FAILED]
18
O_COLEMAK=0 O_GITSTATUS=0 O_NAMEFIRST=1 O_RESTOREPREVIEW=1 [SUCCESS]
19
O_COLEMAK=1 O_GITSTATUS=0 O_NAMEFIRST=1 O_RESTOREPREVIEW=1 [FAILED]
20
O_COLEMAK=0 O_GITSTATUS=1 O_NAMEFIRST=1 O_RESTOREPREVIEW=1 [SUCCESS]
21
O_COLEMAK=1 O_GITSTATUS=1 O_NAMEFIRST=1 O_RESTOREPREVIEW=1 [FAILED]
22
make: *** [Makefile:326: checkpatches] Error 1
23
Error: Process completed with exit code 2.

[jarun]

@62832 please refer to: https://github.com/jarun/nnn/tree/master/patches#resolving-patch-conflicts

[62832]

@jarun Should be sorted now, conflicts have been resolved with the Colemak patch.

[jarun]

Re-ran CI.

[jarun]

Thank you!