Status: https://docs.rs/crate/libnotcurses-sys/

The build log shows that it can't compile libnotcurses-sys, since it expects it to be already installed, and it isn't.

If the notcurses package were already in Ubuntu 20.04 (the version used by the Build environment for Rust crates, then we could open an issue asking to include it.

More info:

• Add a dependency to the build environment
• Metadata for custom builds

Until then maybe we could also modify build.rs to download & compile notcurses manually, like in this example: build.rs & Cargo.toml

Also:

• Move poc/example to poc/src/bin so that each file compiles by default.
• Fix the poc/menu description reporting: there was a mistake in string convention

Some quickstart could be added to the README like:

git clone https://github.com/dankamongmen/libnotcurses-sys; cd libnotcurses-sys
cd example/poc
cargo build

On fetching input, Escape key event doesn't seem to be detected the first time. It takes two key-presses to detect Escape key for the first time after which it is detected at each key press.

Reproducible example -

use libnotcurses_sys::{widgets::*, *};
use env_logger;
use log::info;
use nix::poll::{ poll, PollFd, PollFlags };

fn main() -> NcResult<()> {
    env_logger::init();
    reader()
}

fn reader() -> NcResult<()> {
    let mut nc = unsafe { Nc::new()? };
    let mut stp = unsafe { nc.stdplane() };
    
    let input_fd = PollFd::new(
        unsafe { c_api::notcurses_inputready_fd(nc as *mut Nc) },
        PollFlags::POLLIN);

    let mut input_details = NcInput::new_empty();
    loop {
        if let Ok(_) = poll(&mut [input_fd], -1) {
            let recorded_input = nc.get_nblock(Some(&mut input_details))?;
            info!("Recieved {:?} {:?}", recorded_input, input_details);
        }
    }

    unsafe { nc.stop()? };
    Ok(())
}

Viewing the logs RUST_LOG=trace cargo run would reveal that each other keyboard input (characters, Tab etc) is detected on one key-press, only Escape key takes two.

The behavior is the same if I don't use polling and simply loop delayed get(). My Escape key appears to be registering perfectly for other applications so I doubt it is the issue.

Terminals - Alacritty, XTerm.

Please confirm this if this behavior is reproducible and is a bug.

I've been meaning to draw empty cells across a line, yet am unable to debug why gradient function is throwing error. Here's a simple reproducible example.

use libnotcurses_sys::*;

fn main() -> NcResult<()> {
    let nc = unsafe { Nc::with_flags(NcOptions::SUPPRESS_BANNERS)? };
    let splane = unsafe { nc.stdplane() };

    let header_bg_channel = NcChannel::from_rgb8(
            57,
            100,
            200,
        );
    let header_fg_channel = NcChannel::from_rgb8(
            200,
            200, 
            200
        );

    let header_combined_channel = NcChannels::combine(header_fg_channel, header_bg_channel);
    splane.gradient(
        Some(0),
        Some(0),
        Some(1),
        None,
        " ",
        0,
        header_combined_channel,
        header_combined_channel,
        header_combined_channel,
        header_combined_channel,
    )?;
    nc.render()?;
    Ok(())
}

In some rare cases, it does render with garbage values filled in. Am I doing something wrong or making some invalid assumptions?

OS: Arch Terminals: Xterm, Alacritty, Kitty

The situation

At the moment this library is in a complicated situation in regard to its version number.

It is currently not following semver, but the version is tied to the one the notcurses C library, as an indication of the API that is supposed to be compatible with. Any breaking changes that this library makes outside of the ones from the notcurses C lib API are not reflected anywhere in the version number. This is not ideal.

For example, some methods need to be declared unsafe in #12, and this doesn't have anything with the notcurses C lib.

I found some discusions regarding the problem of using semver for wrappers/binding libs:

• https://stackoverflow.com/questions/16418183/what-are-the-best-practices-to-do-semver-semantic-versioning-of-a-wrapper-libr
• https://github.com/semver/semver/issues/352

Several thoughts:

• It makes most sense to have a semver version number for the development of this library, independently of any other library.
• The release of notcurses 3.0 is near. We could upgrade to 3.0 in sync with it and at the same time we adopt semver. The problem is if we keep the rapid development and iteration cycle of the bindings it would mean increasing the major version number very often.
• Ideally we should be in a major 0 version right now (Major version zero (0.y.z) is for initial development. Anything MAY change at any time. The public API SHOULD NOT be considered stable.) (semver.md)
• I'd really like to be able to evolve API of the Rust bindings until I feel it's stable enough, (I don't like to rush to 1.0)
• Maybe we could treat the current version 2.X as development version. Or maybe even upgrade to 3.X and treat that one as the develoment version, signaling the change by bumping. And keep upgrading the version number from then on, detached from the version number increases of the notcurses C lib, as if it where a 0.X version (and indicate that in the docs). I like this option the most.
• The compatibility with which notcurses C lib specific version should also be clearly indicated in the docs.

ncvisual_render() has been deprecated, and will be removed for ABI3. ncvisual_blit() has replaced it, with slightly different semantics. this was necessary to respect pile logic. you'll want to reflect this.

it might also be wise to check out the dankamongmen/3.0.0 branch where all ABI3 work is happening.

w00t w00t!

With a new application (cargo init hello_world), and notcurses = 3.0.1 as a dependency in cargo.toml, and notcurses 3.0.8 installed via homebrew, running cargo build produced the following error:

Compiling libnotcurses-sys v3.6.0
error[E0308]: mismatched types
    --> $HOME/.cargo/registry/src/github.com-1ecc6299db9ec823/libnotcurses-sys-3.6.0/src/plane/methods.rs:1610:59
     |
1610 |             unsafe { c_api::ncpile_render_to_buffer(self, &mut buf, &mut len.try_into().unwrap()) },
     |                                                           ^^^^^^^^ expected `i8`, found `u8`
     |
     = note:    expected raw pointer `*mut *mut i8`
             found mutable reference `&mut *mut u8`

doc tests worked in the past, but for a while they've been failing.

I've createda minimal example to investigate the issue:

In this example, it hangs in the running 1 test line, until a key is pressed, then it fails continues:

$ cargo test

running 0 tests
test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s
   Doc-tests issue-test
running 1 test
test src/lib.rs - (line 13) ... FAILED
failures:
---- src/lib.rs - (line 13) stdout ----
Test executable failed (terminated by signal).
failures:
    src/lib.rs - (line 10)
test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 2.71s
error: test failed, to rerun pass '--doc'

Running it with strace (via cargo-with, installable by running cargo install cargo-with ) like this:

$ cargo with "strace -s1000" -- test

produces the following output strace.txt

Before publishing the first release we should make sure the features are on par with a given version on the original library:

pending changes against v2.4.2:

• [x] ncsubproc
• [x] tabbed widget
• [x] ncdirect_stream
• [x] ncdirect_styles
• [x] ncdirect_supported_styles
• [x] ncpile_render_to_buffer
• [x] ncpile_render_to_file
• [x] nccells_ascii_box
• [x] nccells_heavy_box
• [x] nccells_light_box
• [x] notcurses_hostname
• [x] notcurses_accountname
• [x] notcurses_enter_alternate_screen
• [x] notcurses_leave_alternate_screen
• [x] ncvisual_subtitle_plane
• [x] ncvisual_from_palidx
• [x] ncplane_boundlist
• [x] ncplane_cursor_move_rel
• [x] ncplane_erase_region
• [x] ncplane_scrollup
• [x] ncplane_scrollup_child
• [x] ncstrwidth_valid
• [x] notcurses_getvec
• [x] NCOPTION_DRAIN_INPUT & NCDIRECT_OPTION_DRAIN_INPUT

Hi, before I ask, let me tell that I'm myself a beginner in Rust.

Doubt

NcPlane::new_child has the following signature -

pub fn new_child<'parent, 'plane, 'opts>(parent: &'parent mut NcPlane, options: &'opts NcPlaneOptions) ->
NcResult<&'plane mut NcPlane>

The lifetime 'plane is problematic here, in my opinion, as the compiler has no way to "tie" this to another existing lifetime (like 'opts or 'parent). This leads to assign the reference to the newly created child plane, an unknown lifetime. And because of this is it tricky to use this plane as a field of struct since constraining the reference with the lifetime constraints of the struct is not viable.

If I recall correctly, all children planes would be destroyed on destruction of parent and hence it makes sense to constraint the child reference to live as long as parent ('parent lifetime), instead of an arbitrary lifetime ('plane).

So, I want to ask how should I go about it? Because I'm using a child plane as a field in a struct, something along the lines of -

struct Widget<'a> {
  plane: &'a mut NcPlane
  ... other fields
}

And assigning Widget { plane : NcPlane::new_child(parent_plane, ...) } just leads to compilation complications as we nest structs and all. Am I doing something wrong?

After asking a dumb question in dankamongmen/notcurses#1883 about planes, I'm opening this issue in order to create an practical example that can fix the holes in my understanding.

after the split from the original repo we no longer have automatic testing, we should fix that.

pending configurations:

• [ ] .drone.yml
• [ ] .github/workflows/
  • [ ] macos_test.yml
  • [x] ubuntu_test.yml
  • [ ] windows_test.yml

feature branches

• platform-windows → windows_test.yml
• platform-macos → macos_test.yml

I've submitted rust-libnotcurses-sys for building on Fedora Rawhide, and am running into a problem with regards to a size_t argument. This type has implementation-defined size. Here's a failed build; they can all be found linked from here.

  --> src/bindings.rs:24:5
   |
24 |     __va_list_tag,
   |     ^^^^^^^^^^^^^ no `__va_list_tag` in `bindings::ffi`
error[E0432]: unresolved import `crate::ffi::__va_list_tag`
 --> src/plane/reimplemented.rs:5:5
  |
5 | use crate::ffi::__va_list_tag;
  |     ^^^^^^^^^^^^^^^^^^^^^^^^^ no `__va_list_tag` in `bindings::ffi`
error[E0308]: mismatched types
   --> src/notcurses/methods.rs:495:75
    |
495 |         error![unsafe { crate::notcurses_render_to_buffer(self, &mut buf, &mut len) }]
    |                                                                           ^^^^^^^^ expected `u32`, found `u64`
    |
    = note:    expected raw pointer `*mut u32`
            found mutable reference `&mut u64`
error[E0308]: mismatched types
   --> src/plane/reimplemented.rs:167:55
    |
167 |     unsafe { crate::ncplane_putnstr_yx(plane, -1, -1, size, cstring![gclustarr]) }
    |                                                       ^^^^ expected `u32`, found `u64`
    |
help: you can convert a `u64` to a `u32` and panic if the converted value doesn't fit
    |
167 |     unsafe { crate::ncplane_putnstr_yx(plane, -1, -1, size.try_into().unwrap(), cstring![gclustarr]) }
    |                                                       ^^^^^^^^^^^^^^^^^^^^^^^^
error[E0308]: mismatched types
  --> src/signal.rs:15:23
   |
15 |         Self { __val: [0; 16] }
   |                       ^^^^^^^ expected an array with a fixed size of 32 elements, found one with 16 elements