https://linen.dev logo
Join Discord
Powered by
# haxe-ui
  • a
    No idea
  • b
    yeah, its a UI lib for haxe: http://haxeui.org/
  • b
    😃
  • b
    @cool-psychiatrist-49311 - im sure tink will do this, and lots lots more, but i really dont want the dependency... one of the things is not have any dependancy to anthing 3rd party, except the core, the backend and any framework requirements (openfl, lime, kha, etc)
  • c
    sounds like ufront
  • c
    and ultimately becomes too bulky to be maintainable
  • b
    the bigger issue with ufront is probably that the maintainer went missing 😄
  • b
    which is a good way to kill any project
  • c
    That's the smaller issue I'd say. If the library itself is more maintainable it would be picked up by someone (me)
  • c
    If its components were small and isolated it would be more easily maintained. Because they can be tested separately.
  • b
    by that logic, the tink ecosystem would be fine if you and juraj stopped maintaining it. I kind of doubt that
  • c
    Well some of the libraries are complex by itself. e.g. tink_web. But it is irrelevant when I want to improve its dependencies. e.g. tink_http I once wanted to improve the http layer of ufront but turns out i failed. Because changing some low level parts requires me to update hundreds of other places just to compile and test my change.
  • c
    Yeah that was a fundamental change. But in tink style at least I can confine my scope at the first place.
  • b
    there's always two sides to the argument
  • b
    in a mono-repo, you can at least update everything at once during a refactor
  • b
    with many small repos, you need a thousand commits for that
  • b
    i think, also, in this case, it depends on the context... if haxeui was pretty much ALL macro based, then it would make sense... but in fact its not really, only a relatively small part is, and its only for "sugar" so having a dependency to tink seems overboard...
  • b
    speaking of sugar... adding some nice new stuff while im refactoring:
  • b
    Copy code
    @:xml("
<hbox>
    <button id='button1' text='Button 1' />
    <button id='button2' text='Button 2' />
    <button id='button3' text='Button 3' />
</hbox>
")
class Test1 extends Box {
    @:bind(button1.text) public var button1Text:String = "bobbbby!";
    @:bind(button2.text) public var button2Text:String;
    @:bind(button3.text) public var button3Text:String;

    public function new() {
        super();
        button3Text = "Changed!";
    }
    
    @:bind(button1, MouseEvent.CLICK)
    public function onButton1Click(e:UIEvent) {
        button2Text = "Changed!";
    }
}
  • b
    hm, maybe the new inline markup could be used for that
  • b
    not sure if you can use that in metadata
  • b
    yeah, going to have a play new inline markup also
  • b
    but i want things to be haxe3 compatible too
  • b
    do you know how to add a new class in a macro that is accessible outside of the macro?
  • b
    ive tried a few things and turning up blanks... im no doubt just doing things wrong
  • b
    ive tried this as an onAfterTyping callback:
  • b
    Copy code
    private static function fn(a:Array<ModuleType>) {
        trace(a.length);
        if (gen == true) {
            return;
        }
            var c = macro class TestClass {
                public function new() {
                }
                public var x:Int = 101;
            }
               
            ;
            Context.defineType(c);
            gen = true;
    }
  • b
    (just a messy test)
  • b
    it compiles, but i cant seem to find a TestClass
  • c
    If I don't use the data binding feature will the relevant code get compiled to the output?
12345...1687Latest