one gtk_test::click leads to two clicked events

[liquidnight2]

I took the basic example https://github.com/gtk-rs/gtk-test/blob/master/tests/basic.rs and expanded it with an atomic counter and a loop with 10 executions. As we see later in the output, we get two closure evaluations, invoked by connect_clicked().

src/main.rs:

extern crate gtk;
extern crate gtk_test;

use std::sync::atomic::AtomicUsize;
use std::sync::atomic::Ordering;
use std::sync::Arc;

use gtk::Button;
use gtk::ButtonExt;
use gtk::ContainerExt;
use gtk::GtkWindowExt;
use gtk::Label;
use gtk::Orientation;
use gtk::Window;
use gtk::WindowType;
use gtk::WidgetExt;

fn main() {
    for _a in 0..12 {
        println!("===== TEST RUN  {} ", _a);
        button_with_counter();
    }
}

// without gedit on screen: works.
// with gedit on screen: doesnt work.
fn button_with_counter() {
    let click_counter = Arc::new(AtomicUsize::new(0));
    let ccc = click_counter.clone();

    gtk::init().unwrap();
    let window = Window::new(WindowType::Toplevel);
    let gbox = gtk::Box::new(Orientation::Vertical, 0);
    let label: gtk::Label = Label::new("button+counter");
    gbox.add(&label);

    let button1 = Button::new();
    button1.set_label(&"button-1");
    button1.connect_clicked(move |_| {
        println!(" button1 click received    ");
        (*ccc).fetch_add(1, Ordering::SeqCst);
    });

    gbox.add(&button1);
    window.add(&gbox);
    window.show_all();
    window.activate_focus();

    gtk_test::click(&button1);
    gtk_test::wait(300);

    let caw = (*click_counter).load(Ordering::SeqCst);
    println!(" counter after wait: {} ", caw);
    assert_eq!(1, caw);

    window.close();
    window.destroy();
    gtk::Inhibit(false);
}

Cargo.toml :

[package]
name = "gtktest-doppelclick"
version = "0.0.1"

[[bin]]
name = "doppelclick"
path = "src/main.rs"

[dependencies]
gtk-test = "0.2.0"

[dependencies.gtk]
 version = "0.5.0"
 features = ["v3_22"]

As result I get on my LinuxMint-19 VM:

===== TEST RUN 0
button1 click received
button1 click received
counter after wait: 2
thread 'main' panicked at 'assertion failed: (left == right)
left: 1,
right: 2', src/main.rs:54:5

On my gentoo host I get this result:

===== TEST RUN 0
button1 click received
counter after wait: 1
===== TEST RUN 1
button1 click received
counter after wait: 1
===== TEST RUN 2
button1 click received
button1 click received
counter after wait: 2
thread 'main' panicked at 'assertion failed: (left == right)
left: 1,
right: 2', src/main.rs:54:5
note: Run with RUST_BACKTRACE=1 for a backtrace.

Additional Hint:
On my gentoo host, i see a behaviour in 90% of the trials: {

• when another gtk program is on the screen, the test fails
• when no other gtk program is on the screen, the test works }
  For the remaining 10% of the trials I saw it either working continually, or not working.

Host System information:
Portage 2.3.51 (python 2.7.15-final-0, default/linux/amd64/17.0/desktop, gcc-6.4.0, glibc-2.27-r6, 4.17.8-gentoo x86_64)
System uname: Linux-4.17.8-gentoo-x86_64-AMD_A10-7870K
# rustup show
Default host: x86_64-unknown-linux-gnu
installed toolchains
stable-x86_64-unknown-linux-gnu
nightly-x86_64-unknown-linux-gnu
installed targets for active toolchain
i686-pc-windows-gnu
x86_64-pc-windows-gnu
x86_64-unknown-linux-gnu
active toolchain
stable-x86_64-unknown-linux-gnu (default)
rustc 1.32.0 (9fda7c223 2019-01-16)

VM System information:
Linux Mint 19 Cinnamon, Kernel 4.15
Installed package via console: libgtk-3-dev
# rustup show
Default host: x86_64-unknown-linux-gnu
rustc 1.32.0 (9fda7c223 2019-01-16)

[liquidnight2]

Further information: that bug appears on my LinuxMint in conjunction with focus change of the active window. What I just see:

• On a large console window or the console placed in center, the test doesnt work. Since LinuxMint places new applications somewhere centered on the screen.
• On empty screen with small console window, it works.
• If you open a small gedit window, it does not work.
• if you had focused this test, and the console only, the focus switch is only between those two, and the test works again.

[GuillaumeGomez]

You're not supposed to do anything while running those tests otherwise interactions might end in the wrong window. Otherwise, I was completely unable to get your error... :-/

[liquidnight2]

Yes I am aware that the test ist focus-sensitive and works when not disturbed. I confirm so far: in the console i type "cargo run" then i do nothing until the program has finished.

How did you try to replay the issue ? On your local machine ? What distro/gtk/windowmanager are you using ? Or a VM with a common linux ? Any hint how to get a better ability to make that replayable ? maybe Docker ?

[GuillaumeGomez]

No VM and Linux mint 19. That's pretty much it...