Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Always read user input from /dev/tty #396

Open
lotabout opened this issue Mar 10, 2020 · 13 comments
Open

Always read user input from /dev/tty #396

lotabout opened this issue Mar 10, 2020 · 13 comments
Labels
enhancement in_progress is being worked on

Comments

@lotabout
Copy link

Is your feature request related to a problem? Please describe.
I'm trying to port my CLI util skim to crossterm for windows support. And one of the use cases of skim is to accept piped input:

echo "input" | sk

In this case the stdin do not generate user inputs while I still hope the user inputs are handled correctly.

Describe the solution you'd like
When creating file descriptor, use /dev/tty all the time.

@TimonPost
Copy link
Member

TimonPost commented Mar 10, 2020

We already use /dev/tty, this one is not supported for windows of course. There we use winapi. Crossterm can be used through a pipe, please see this example for a demo.

@lotabout
Copy link
Author

lotabout commented Mar 10, 2020

@TimonPost Not sure if I understand correctly. I believe the example shows how to write to stderr instead of stdout. While I am trying to do is to read piped input(stdin).

For example, expecting the following example to work:

$ echo "abc"| cargo run --example event-read
    Finished dev [unoptimized + debuginfo] target(s) in 0.05s
     Running `target/debug/examples/event-read`
Blocking read()
 - Keyboard, mouse and terminal resize events enabled
 - Hit "c" to print current cursor position
 - Use Esc to quit

Error: IoError(Os { code: 25, kind: Other, message: "Inappropriate ioctl for device" })

@TimonPost
Copy link
Member

TimonPost commented Mar 24, 2020

I've tried to always use /dev/tty here:

/// Creates a file descriptor pointing to the standard input or `/dev/tty`.
pub fn tty_fd() -> Result<FileDesc> {
    let (fd, close_on_drop) = (
        fs::OpenOptions::new()
            .read(true)
            .write(true)
            .open("/dev/tty")?
            .into_raw_fd(),
        true,
    );

    Ok(FileDesc::new(fd, close_on_drop))
}

But doesn't seem to fix the issue. Any idea why it might not detect an IOCTL device while using a pipe ?

@TimonPost
Copy link
Member

In fact, this resulted in an error for ITerm terminals. #404. On Linux, it seems to behave just fine, but on macos no input can be read anymore.

@lotabout
Copy link
Author

@TimonPost Just to make sure, were tcgetattr and tcsetattr operated on /dev/tty as well?

(Disclaimer: I haven't checked the solution)

@TimonPost
Copy link
Member

Ah, no they did not, I will open a branch with those changes. If someone can validate the working on mac os we can merge it.

@extrawurst
Copy link
Contributor

@TimonPost I can test it on Mac, let me know

@TimonPost
Copy link
Member

Can someone try this branch on mac os: #407

@sophiajt
Copy link
Contributor

sophiajt commented Apr 6, 2020

@TimonPost - I can give it a try. Is there something in particular that needs testing?

@TimonPost
Copy link
Member

You could run the interactive example in the examples folder. And in there run the input demo. It should work. Though on the PR #407 and #406 there are people who did the testing already. They reported invalid input. I left a comment with some ideas on there.

@TimonPost TimonPost added enhancement in_progress is being worked on labels Apr 9, 2020
@sophiajt
Copy link
Contributor

@TimonPost I noticed with the latest master that #406 is fixed but I do still see the issue that @lotabout mentions in the comment above:

% echo "abc" | cargo run --example event-read
   Compiling crossterm v0.17.3 (/Users/test/Source/crossterm)
    Finished dev [unoptimized + debuginfo] target(s) in 3.42s
     Running `target/debug/examples/event-read`
Blocking read()
 - Keyboard, mouse and terminal resize events enabled
 - Hit "c" to print current cursor position
 - Use Esc to quit

Error: IoError(Os { code: 25, kind: Other, message: "Inappropriate ioctl for device" })

@TimonPost
Copy link
Member

Yea, there is someplace some argument is given, I supposed to libc, which is invalid. I haven't found that place yet. Regarding @Iotabout his comment, his ideas are implemented by this PR.

@austinjones
Copy link

I ran into this issue in the tab terminal multiplexer. I was trying to add support for stdin input in a raw mode app:

echo query | tab

I hit the same error as described above.

irh added a commit to koto-lang/koto that referenced this issue Oct 22, 2021
This will allow for the REPL to be used on Windows, but has the side-effect of
making the REPL untestable via stdin, due to
crossterm-rs/crossterm#396

I haven't run into a failing test here in recent times except for occasionally
leaving in some debug output, which has been caught a few times. I'll add a
specific test for ensuring trace isn't left in later in the branch.
irh added a commit to koto-lang/koto that referenced this issue Oct 22, 2021
This will allow for the REPL to be used on Windows, but has the side-effect of
making the REPL untestable via stdin, due to
crossterm-rs/crossterm#396
irh added a commit to koto-lang/koto that referenced this issue Oct 22, 2021
This will allow for the REPL to be used on Windows, but has the side-effect of
making the REPL untestable via stdin, due to
crossterm-rs/crossterm#396
irh added a commit to koto-lang/koto that referenced this issue Oct 22, 2021
This will allow for the REPL to be used on Windows, but has the side-effect of
making the REPL untestable via stdin, due to
crossterm-rs/crossterm#396
irh added a commit to koto-lang/koto that referenced this issue Oct 23, 2021
This will allow for the REPL to be used on Windows, but has the side-effect of
making the REPL untestable via stdin, due to
crossterm-rs/crossterm#396
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement in_progress is being worked on
Projects
None yet
Development

No branches or pull requests

5 participants