Mirror of forum.nim-lang.org

786 :: nim2haxe?

• mapdog (orginal) [2015-01-23T20:44:40+01:00] view original

  How hard would it be to make a converter nim to haxe source. You could concentrate on a limited subset.

  Modules could be translated to a class. Proc becomes static function in that class. Import x expands to multiple static imports of all a class functions. {}; are added. Type Object makes a class. Stack allocation becomes heap.

---

• RayoGundead (orginal) [2015-01-24T02:36:24+01:00] view original

  It would be godsend if someone is making Haxe target for Nim.

  Since I'm working on an XSL library(xslt/xpath/xsl-fo) implementation in Nim, having the Haxe target will enable interface for other platforms. Think PHP, JS, etc. perfect for ending the Saxonica java implementation monopoly.

---

• RayoGundead (orginal) [2015-01-24T02:59:06+01:00] view original

  It is possible to make a Haxe target. The problem is lack of time to work on it, unless I get paid. Also, I don't know how hard it will be to implement.

---

• RayoGundead (orginal) [2016-03-09T04:52:50+01:00] view original

  Before even starting with the Haxe target, watch this first: Quaxe, infinity and beyond" by Daniel Glazman starting here.

---

• OderWat (orginal) [2016-03-09T12:44:13+01:00] view original

  Well. Nim already has JavaScript and (very) experimental PHP targets. Creating a Haxe target would probably work pretty similar. I am just not that confident that you really want to use something like Haxe if you already can target JS and PHP more directly with Nim.

---

• RayoGundead (orginal) [2016-03-17T05:01:00+01:00] view original

  I didn't know that Nim has PHP target. If it is stable enough then I can finally break free from Haxe.

  Most of my work requires PHP because wordpress and drupal, and I've been using Haxe to get away from PHP.

---

• OderWat (orginal) [2016-03-17T15:01:21+01:00] view original

  I don't think one can consider it as stable yet. Araq and me use it to explore how it could be used in our companies projects and have put some thoughts into this already. But PHP is not planned to be a "perfect" target ever. Because you would need to emulate memory access in PHP which does not really make sense. What we strife for is nice interoperability with >250K existing PHP code. Such that one can write "plugins" in Nim but reuses a lot of the existing PHP code. For this we made it possible to create and extend classes and handle the dynamic datatypes by defining types for them when used in the Nim context. This is all fairly experimental and currently I still refrain from using such code in production besides of "one shot" programs. You may want to try it out, but we removed the PHP compatibility layer from the main nim-compiler, because we are not ready to support the "public". But we could publish those parts again so you can check what it can do for you.

---