Use clipr::dr_clipr()

[gvdr]

Clipboard is not generally available on Linux. Following the minimum example provided (copying to clipboard y <- 1:4) I got:

reprex()
#> No input provided and clipboard is not available.
#> Rendering reprex...
#> Unable to put result on the clipboard. How to get it:
#>   * Capture what reprex() returns.
#>   * Use `outfile = "foo"` to request output in specific file.
#>   * See the temp file:
#>     - /tmp/RtmpSKU1b1/file46ab29804967_reprex.md

Clipr provides also a solution:

clipr_available()
#> [1] FALSE
read_clip()
#> Error: Clipboard on X11 requires 'xclip' (recommended) or 'xsel'.

And in fact, having installed xclip, everything seems to be working fine. 👍

However, the solution is maybe not so straightforward to find? I'd suggest to put a line somewhere suggesting Linux users to install xclip if they want to run reprex.

[jennybc]

@mdlincoln What would be the most official way for me to delegate messaging about this to clipr? I already use clipr_available() internally, but it returns logical. What I really need is a way put the right message in front of the user's eyeballs when clipr_available() returns FALSE.

[mdlincoln]

Hm, good question. Possible solution:

Add a message = TRUE/FALSE argument to clipr_available() that would send a message() with that installation suggestion text when returning FALSE?
a. Or a warning()?

[jennybc]

Here in reprex, I would handle with a closure so that I only message upon the first attempt to use the clipboard. Maybe you need a clipr "doctor" function that either emits an "all clear" or OS-appropriate advice on how to make the clipboard available? For now I can presumably catch the error from read_clip() to approximate this.

[mdlincoln]

That does sound more elegant. I'll have to read up a bit on how to build that in to clipr, and will keep you posted.

[jennybc]

I think this is more elaborate than you need, but something along the lines of here::dr_here() would be really useful.

https://github.com/krlmlr/here/blob/93593ee1e04d456d33199795c1d6491bd9b84164/R/here.R#L25-L54

You have all the ingredients (checking availability and suggesting remedies), just not packaged in a doctor function.

devtools::dr_github() is also inspiring (and also more involved than what clipr needs).

[gvdr]

Just as a side: my idea was more humbly to just include a line warning of this potential problem in the README.md. Obviously, the solution that you are building is much stronger.

[mdlincoln]

Indeed @gvdr, but your issue definitely highlights something that clipr should handle more transparently, especially since many people will be using clipr when it's called by other packages like reprex.

[jennybc]

I'm adding one bullet to my message and will upgrade the approach when/if clipr::dr_clipr() exists.

> Sys.setenv("CLIPBOARD_AVAILABLE" = FALSE)
> reprex(1:4)
Rendering reprex...
Unable to put result on the clipboard. How to get it:
  * Unix-like systems may require explicit installation of xclip or xsel. 
  * Capture what reprex() returns.
  * Use `outfile = "foo"` to request output in specific file.
  * Use `outfile = NA` to request output in working directory.
  * For now, see the temp file:
    - /private/var/folders/vr/gzrbtprx6ybg85y5pvwm1ct40000gn/T/RtmpOnN9lT/file735b4d40d67_reprex.md

[mdlincoln]

I actually have some time to look at this over the next day - I'll keep you updated @jennybc!

[mdlincoln]

OK, the latest build of clipr now has a dr_clipr() method. When you can, mind giving it a spin and let me know if you've any suggestions before I mint a CRAN release?

[jennybc]

I'm going to reopen this so I don't forget to test drive dr_clipr()!

[mdlincoln]

@jennybc I totally forgot about minting a new CRAN release =_= but I will by the end of this week! Let me know if you spot any issues with dr_clipr() before then :)

[jennybc]

OK. I don't think I've given it a whirl yet but will try do to do! reprex is also basically ready for a new release. I am waiting on a few dependencies to appear or update on CRAN.