Hello! I need you help again.

I try to access a struct inside a callback.

My example code:

#[macro_use]
extern crate kiss_ui;

use kiss_ui::prelude::*;

use kiss_ui::button::Button;
use kiss_ui::container::Horizontal;
use kiss_ui::dialog::{self, AlertPopupBuilder};

pub struct Counter {
 pub value: u8,
}

impl<'a> Counter {
    fn new() -> Counter {
        Counter {value: 0}
    }

    fn inc(&mut self) {
        self.value += 1;
    }
}

#[macro_use]
extern crate lazy_static;
// cargo [dependencies]
// lazy_static = "0.1.*"

lazy_static! {
    static ref counter: Counter = Counter::new();
}
// works but is immutable 

// static mut counter: Counter = Counter::new();
// error: function calls in statics are limited to struct and enum constructors
// hmm isn't it a struct constructor?

fn main() {
//  let mut counter = Counter::new();

    kiss_ui::show_gui(||
        Dialog::new(
            Horizontal::new(
                children![               
                    Button::new()
                        .set_label("Message")
                        .set_onclick(show_counter_dialog),
                    Button::new()
                        .set_label("Close")
                        .set_onclick(close_dialog),
                ]
            )
            .set_elem_spacing_pixels(10)                   
        )
        .set_title("Button test!")
    )
}

// fn show_counter_dialog(_: Button, counter: Counter) { 
// =>  (expected a tuple with 1 elements, found one with 2 elements)
fn show_counter_dialog(_: Button) {
//  let mut counter = Counter::new();
// works, but i don't want to reset it 
    counter.inc();
    dialog::message_popup("Good job!", format!("You clicked the button {} times!", counter.value));
}
fn close_dialog(_: Button) -> CallbackStatus {
    println!("Closing dialog!");
    CallbackStatus::Close
}

I'm working on it since 3 days. :/ In PHP i just say something like "fn callback() use &counter {}".

Any Ideas how i get it done?

Thank you in advance!

TL;DR: Can and should kiss-ui and iup-rust coexist i.e. they have different goals?

I have been working on overhauling iup-rust this week to be idiomatic Rust instead of just a low-level wrappers, in the middle of this task I ended up finding this library (very nice btw) which builds upon IUP, instead of dropping all my work I choose to continue and see what happens.

But it still bites my mind that there's two kinda of similar libraries (at least at this moment that they are young), so I really needed to discuss this out over here to clear up my mind. I'm too new to the open source community and I'm not sure how everything ties together.

What's the direction this library is taking? Binding IUP or building a completly independent widget toolkit that just has IUP as backend? If not the later what should be done about two libraries doing the exactly same thing?

Best regards, nice library you are building here in a very quick pace :)

If i try to use Button::new().set_enabled(false) or .set_name() i get the this error:

<kiss_ui macros>:1:52: 1:62 error: the trait core::convert::From<()> is not implemented for the type > kiss_ui::BaseWidget [E0277] <kiss_ui macros>:1 ( $ ( $ child : expr ) , + , ) => ( [ $ ( $ child . into ( ) ) , + ] ) ; ( ) ^~~~~~~~~~ <kiss_ui macros>:1:1: 2:14 note: in expansion of children!

I don't know how to try it without using children! But i think is because Button don't implement set_enabled().

• KISS-UI is forbidden to run in more than one thread (went from thread-local flag to static atomic boolean), since I believe most GUI toolkits don't support more than one running instance at a time.
• Consolidated thread-local data to one object (may put in Box so we don't consume too much TLS)

And the big fix:

• String slices from widget metadata are now tracked so that they cannot be updated while there are references to them. Mostly non-breaking except for client code that depended on string accessors returning &str.

Closes #29.

The iup package on my system links to gtk-2, but kiss-ui unconditionally tries to link to gtk-3, making the build fail.

One possible solution to this letting the user decide whether he wants to link to gtk-3 through a cargo feature. But this might not be a very ideal solution.

Here is what I cooked up:

diff --git a/Cargo.toml b/Cargo.toml
index 6f005ea..c9c6b0d 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -5,6 +5,10 @@ authors = ["Austin Bonander <[email protected]>"]

 [dependencies.iup-sys]
 path = "iup-sys/"
+features = [""]

 [dependencies]
 libc = "*"
+
+[features]
+no-gtk3 = ["iup-sys/no-gtk3"]
diff --git a/iup-sys/Cargo.toml b/iup-sys/Cargo.toml
index 805e976..74a45e1 100644
--- a/iup-sys/Cargo.toml
+++ b/iup-sys/Cargo.toml
@@ -8,3 +8,6 @@ build = "build.rs"

 [dependencies]
 libc = "*"
+
+[features]
+no-gtk3 = []
diff --git a/iup-sys/build.rs b/iup-sys/build.rs
index 464af6c..ceee847 100644
--- a/iup-sys/build.rs
+++ b/iup-sys/build.rs
@@ -1,7 +1,7 @@
 fn main() {
     add_link("iup");

-    if cfg!(not(windows)) {
+    if cfg!(not(windows)) && std::env::var_os("CARGO_FEATURE_NO_GTK3").is_none(){
         add_link("gtk-3"); 
     }
 }

On exit of show_alert_message text gets dropped and its refcount underflows.

This message is printed to the console:

thread '<main>' panicked at 'arithmetic operation overflowed', C:\devel\Rust\kiss-ui\src\widget.rs:213
fatal runtime error: Could not unwind stack, error = 5

I always get this error, when trying to run the examples I followed the installation of iup for

3.13.0-53-generic #89-Ubuntu x86_64 libiup.so: cannot open shared object file: No such file or directory

I know that when installing the iup library it transfer this file to that directory

'libiup.so’ -> ‘/usr/lib64/libiup.so’

and make sure it is there

/usr/lib64$ ls -la | grep libiup
-rwxr-xr-x  1 root root  826955 May 27 16:49 libiup.so

Luckily, I was able to solve it via

sudo ln -s /usr/lib64/libiup.so  /usr/lib/libiup.so

System: Arch Linux lpt1 4.0.7-2-ARCH #1 SMP PREEMPT Tue Jun 30 07:50:21 UTC 2015 x86_64 GNU/Linux

Also tried this from a Debian Jessie container with the same results...

$ curl http://www.rust-ci.org/artifacts/put?t=$RUSTCI_TOKEN -v

• Trying 151.236.11.197...
• Connected to www.rust-ci.org (151.236.11.197) port 80 (#0)

  GET /artifacts/put?t=LBCp1Sz9ofbWbyN0GtpbICw0CZW4Hw+iOrw1zYJInHtS4/dJjrHjKJqy8aRB2rBPwc7Mt8sSgRVirj90XA0UwDD8nennQB9m8diXYOm03Itj+/2iKJCMchM8r/cNmaYeHelYcEHzLHykNaWSZ08gj0Bz3hDYVMI4ey4h1CpIdwM= HTTP/1.1 Host: www.rust-ci.org User-Agent: curl/7.43.0 Accept: /

  < HTTP/1.1 401 UNAUTHORIZED < Server: nginx/1.5.6 < Content-Type: text/html; charset=utf-8 < cache-control: max-age=0 < Content-Length: 12 < Accept-Ranges: bytes < Date: Wed, 08 Jul 2015 05:47:17 GMT < X-Varnish: 1493651606 < Via: 1.1 varnish < Connection: keep-alive < age: 0 < X-Cache: MISS <

• Connection #0 to host www.rust-ci.org left intact Unauthorized

Is there a way to use other commonly used widgets , such as ListViews and ComboBoxes? I can't seem to find it so I am assuming that it is not finished at the moment.

I just realized there's unsoundness in some of the APIs because of widgets implementing Copy and Clone:

• Widget::set_name() and Widget::get_name(): users can set a new name string and make the current one point to garbage in safe code. Even though set_name() takes self, the Copy impl makes ownership pointless and Clone allows creating a new owned alias.
• Label::set_text()/get_text(), TextBox::set_text()/get_text(): same thing.

Are people using these impls yet? It would break the progress_test example but as #27 shows, it's not all that intuitive a pattern anyways. Retrieving children by-name should probably be the preferred method, though... that can also produce unsoundness.

Maybe this API requires more thought. Or just return owned strings from these getters? That would be a lot of copying. Perhaps an unsafe method that returns the string slice, and then a safe method that returns a clone.

Users should be able to start a long-running task on a new thread, and be notified when it completes in a callback on the UI thread.

Maybe use threadpool crate and a Cargo feature, async, so users don't pay for what they don't need.

Hey, I installed IUP statically and I'm getting this error when trying to build my project. https://hastebin.com/ohikojorum.php

I'm probably just dumb and missed a step, but I can't seem to figure out why it can't find my GTK libs?

The docs for kiss_ui::show_gui state

"The closure argument should initialize and call .show()."

and the docs for kiss_ui::dialog::Dialog state

"This does not make the dialog appear on screen. .show() must be called after the dialog has been configured."

There doesn't appear to be any .show() method though.

I downloaded iup-3.14_Win64_dllw4_lib.zip, copied its contents in \Rust\bin\ and \Rust\lib\rustlib\x86_64-pc-windows-msvc\lib\. then I ceated a hello-world project with

[dependencies.kiss-ui]
git = "https://github.com/cybergeek94/kiss-ui"

and

#[macro_use]
extern crate kiss_ui;

use kiss_ui::container::Horizontal;
use kiss_ui::dialog::Dialog;
use kiss_ui::text::Label;


fn main() {
    kiss_ui::show_gui(|| {
        Dialog::new(
            Horizontal::new(
                children![
                    Label::new("Hello, world!"),
                ]
            )
        )
        .set_title("Hello, world!")
        .set_size_pixels(640, 480)
    });
}

Next I got this build error:

error: linking with `link.exe` failed: exit code: 1181
...
note: Non-UTF-8 output: LINK : fatal error LNK1181: \xed\xe5 \xf3\xe4\xe0\xe5\xf2\xf1\xff \xee\xf2\xea\xf0\xfb\xf2\xfc \xe2\xf5\xee\xe4\xed\xee\xe9 \x
f4\xe0\xe9\xeb \"iup.lib\"\r\n
error: aborting due to previous error

The instructions currently tell people to download the dllw4 version of iup which does not come with .lib libraries and is unsuitable for use with -msvc. For those targets the instructions should tell the user to download the dll12 version instead which definitely has .lib files. The instructions should also inform the user to place the .lib files where the linker can find them and not the .dll files. Also perhaps the instructions should tell the user how to use a build script override to tell it where to look for the .lib and .dll files instead of stuffing those files all over the place in weird directories.

I was just asked about UI libraries for rust. My suggestion of KISS-UI got met with the complaint that it's not on crates.

it would probably be a good idea to also add keywords like "ui" and "gui"

I haven't messed with C libraries much so I'm not sure if this is a problem with kiss-ui, IUP, my system, or me, but I'm getting a build error that I'm not sure how to resolve.

$ cargo build
   Compiling grim-edit v0.0.1 (file:///Users/erincalling/code/grim-edit)
error: linking with `cc` failed: exit code: 1
note: "cc" "-m64" "-L" "/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib" "-o" "/Users/erincalling/code/grim-edit/target/debug/grim_edit" "/Users/erincalling/code/grim-edit/target/debug/grim_edit.o" "-Wl,-force_load,/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib/libmorestack.a" "-Wl,-dead_strip" "-nodefaultlibs" "/Users/erincalling/code/grim-edit/target/debug/deps/libkiss_ui-3b13f82488719214.rlib" "/Users/erincalling/code/grim-edit/target/debug/deps/libiup_sys-4a23062e27aae405.rlib" "/Users/erincalling/code/grim-edit/target/debug/deps/liblibc-ef5cbad4ef5c7a1e.rlib" "/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib/libstd-4e7c5e5c.rlib" "/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib/libcollections-4e7c5e5c.rlib" "/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib/libunicode-4e7c5e5c.rlib" "/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib/librand-4e7c5e5c.rlib" "/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib/liballoc-4e7c5e5c.rlib" "/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib/liblibc-4e7c5e5c.rlib" "/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib/libcore-4e7c5e5c.rlib" "-L" "/Users/erincalling/code/grim-edit/target/debug" "-L" "/Users/erincalling/code/grim-edit/target/debug/deps" "-L" "/usr/local/Cellar/rust/1.0.0/lib/rustlib/x86_64-apple-darwin/lib" "-L" "/Users/erincalling/code/grim-edit/.rust/lib/x86_64-apple-darwin" "-L" "/Users/erincalling/code/grim-edit/lib/x86_64-apple-darwin" "-liup" "-lc" "-lm" "-lSystem" "-lpthread" "-lc" "-lm" "-lcompiler-rt"
note: Undefined symbols for architecture x86_64:
  "_gtk_range_set_value", referenced from:
      _gtkValSetValueAttrib in libiup.a(iupgtk_val.o)
      _gtkValKeyPressEvent in libiup.a(iupgtk_val.o)
      _gtkValChangeValue in libiup.a(iupgtk_val.o)
  "_gtk_adjustment_new", referenced from:
      _gtkValMapMethod in libiup.a(iupgtk_val.o)
  "_gtk_hscale_new", referenced from:
      _gtkValMapMethod in libiup.a(iupgtk_val.o)

{ 600 other symbols snipped }

ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)

error: aborting due to previous error
Could not compile `grim-edit`.

To learn more, run the command again with --verbose.

full, un-snipped error

This is on OSX 10.10.3 after installing IUP using the "precompiled binary" instructions in the README. I'm using cargo 0.4.0-nightly (b48cd46 2015-07-01) (built 2015-07-01) and rustc 1.0.0 (built 2015-05-16)

I'd appreciate any help you can offer!