https://linen.dev logo
Powered by
# haxe-ui
  • b
    followiing up on that ... would it make sense to have a 
    removeComponentById(string)
    as a shortcut ?
  • b
    yeah, probably
  • m
    Yes please that would be great.
  • c
    you might want to try this? https://github.com/MVCoconut/coconut.awt
  • b
    puh ... another framework
  • b
    if i need to use/dependon awt anyhow i would use it directly
  • r
    Than you for these! And yes, I was using removeComponent.
  • b
    ok, cool - removeComponent/At should work as intended now, so you can use either method
  • m
    I wish I was here for this discussion but just my 2 cents the scope would eventually be a subset of the java spec and simply transpile the tests and as much as the std as possible to achieve that.
  • m
    If compilers don't meet a well defined subset of spec compliance, they have to be considered toy projects for how important it is that compilers do what they are supposed to do similar to a car engine
  • b
    i think that would be useful, but i do also think that "real" world test would be nice also, xalan is a pretty huge lib - and arguably quite a useful one... kinda
  • m
    Sure but that's the goal, and that completely depends on supporting a subset of the spec which the language/std tests will do the best at demonstrating
  • b
    im certainly not disagreeing, im simply saying that the "proof of the pudding is in the eating"
  • m
    Fair enough :D how did the discussion end I'm only half way through
  • b
    that Zeta thought it would be "fairly simple" but that he has other things that, are probably, more useful to the community as a whole
  • m
    I see, yeah that's what I thought too, it's good motivation for starting haha
  • p
    'fairly simple' as in you only need a JVM bytecode transpiler. After that you just transpile openjdk, wire it up to openjdk's existing native code and done. But to make it truly portable you'd have to port openjdk to use Haxe std api's, which would be a lot more work.
  • b
    im confused now, wasnt the idea that you take a jar file and it would spit out .hx files?
  • p
    Yes, that's the idea. But to make the generated code portable to eg node.js or even the browser you'd need to adapt the Java std.
  • b
    right, yeah, gotcha
  • b
    i guess thats where the jdk tests come in
  • b
    hard to know how much to map, or if they can map, and if they cant, if its a major problem
  • b
    like java.swing can obviously be ignored, java.net... ... maybe?
  • b
    there is defo quite a bit though
  • b
    and i wonder if things like "beans" are gonna be able to be mapped
  • b
    lots of classes in there, and almost certainly are used everywhere
  • m
    Welcome to transpiler hell !
  • b
    ... ... no thanks... i reject the invitation! 😄
  • m
    You're only a few more steps away from it! First you enjoy writing a cross platform ui library with multiple backends, then you perhaps write a Haxe to c++ ardunio compiler and you become exposed, next thing you know you look at the sheer potential of getting 20k+ libraries into native Haxe and it's too tempting not to at least try a little bit over the weekend...
  • b
    haha
1...105610571058...1687Latest