[Screenshot] Inconsistent screenshot commands and behaviour.

[Atemo-C]

Note: Since this issue is not hardware or software-version/compatibility-specific, these details will be omitted as they are useless. However, it is to note that the latest version of Niri is being used for this, on NixOS 25.11.

---

Issue: Screenshot arguments

The arguments for niri msg action screenshot(*) are very inconsistent and thus confusing, sometimes being applicable, sometimes not, and with no explanation as to why.

The simplest example that explains it all:

---

Expected behaviour: Consistent handling of arguments

The -p and write-to-disk arguments should be handled for all types of screenshots (screenshot, screenshot-window, screenshot-screen).

Examples:

• niri msg action screenshot
  Take a screenshot of any area on the monitor.

• niri msg action screenshot -p=false
  Take a screenshot of any area on the monitor, but hide the cursor by default.

• niri msg action screenshot --write-to-disk=false
  Take a screenshot of any area on the monitor, but do not save to the disk when pressing Space or Enter (additionally, the message shown when taking such a screenshot should reflect that change by saying "to copy the screenshot" instead of "to save the screenshot", and that without having to do Ctrl+C).

• niri msg action screenshot-window -p=true --write-to-disk=false
  Take a screenshot of the currently focused window whilst keeping the cursor visible, and do not save to the disk.

---

Why?

• More consistent usage of arguments
• More user-friendly
• More intuitive and logical
• Makes scripting around Niri's built-in screenshot utility easier

[Atemo-C]

I have simply disabled saving screenshots entirely by now, which is not ideal, and it still shows "Press Space to save the screenshot.", which is another bug, but it is the only way I can work around this weird saving behaviour.

I now handle saving entirely within my own screenshot script, but this still does not fix the underlying issue, and I still have no way to toggle the cursor globally with a command-line argument.

Again, the issue persists, and is a bit annoying. I have considered switching to other screenshot utilities, but I like to keep things more tightly integrated with the environment, and since Niri provides its own utility, I accept the hacky workarounds to make it work like I want, though, I feel that I should not have to.

I hope this will get looked at and eventually fixed someday, even if it is not the highest of priority.

Cheers.

[llitz]

From what I can see:

screenshot - missing write-to-disk option (we are supposed to press ctrl+c instead)
screenshot-screen - has both options
screenshot-window - missing show-pointer option (is there even a way to force capture the pointer if you are taking a specific window screenshot?)

The only alternative for screenshot-window is to manually select the window are via screenshot, which is annoying at the best, but could be too late for some time sensitive cases. screenshot also doesn't allow one to simply "click on a window" to select it, which could be a nice alternative

[Atemo-C]

Yep. I guess this is not an area considered really important for Niri, especially since many people use other screenshot tools than the built-in one, but it would be nice to see this being fixed/changed for the better.

For now, my screenshot script works around Niri's limitations, but even then, it does not change things like the tooltip at the bottom when taking screenshots. Trivial to modify with a fork, but I prefer not forking software for those little things.

[nickjj]

Semi-related, it would also be nice if the commands themselves were consistently named.

Currently there's: screenshot, screenshot-window, screenshot-screen

It would be nice if it were: screenshot-region, screenshot-window, screenshot-screen

This way all 3 commands have a well named suffix and the suffix explicitly says what type of screenshot will be performed.

[Atemo-C]

I agree. My script takes that kind of logic, but instead of combining them like-this, they are standard arguments, like --this (and with many other fallbacks, why not…). I am unsure which approach would be better, but since Niri already uses this-format, it would make sense to roll with it.

Edit: Uploaded newer screenshot of my script.

[nickjj]

Yes, for scripting it would allow having "screenshot-${ACTION}" without needing to have extra logic to handle the empty case. Realistically coding around this is super easy but it would be nice to have 1 less condition to think about.

[nickjj]

I ended up starting #2649 to see if there's an opportunity for there to be screenshot tool enhancements so we don't need custom scripts.