Mirror of forum.nim-lang.org

399 :: Nimrod - Standard 1.0 ?

• enurlyx (orginal) [2014-03-12T13:55:22+01:00] view original

  Hello,

  is there a plan to create a specification, that describes all parts of nimrod, so that other compatible compilers could be created? Or is the plan to have only one compiler?

  With best regards, enurlyx

---

• Araq (orginal) [2014-03-12T20:06:44+01:00] view original

  The manual slowly evolves into a full specification, but it is of low priority. Alternative implementations have a hard time with or without spec as

  1. Nimrod is a complex language
  2. the AST that the compiler exposes for macros is quite bound to the implementation.

  For compiler writers it is not too hard to figure out all the details anyway because it's all open source. (Yes, reading source code can be easier than reading a bunch of abstract jargon: "Equivalence (=) between types is the smallest congruence such that the following holds...")

---

• Demos (orginal) [2014-03-12T20:33:57+01:00] view original

  There seem to be very few languages with meaningful standard documents. Most languages seem to be defined by how their main implementation behaves. Off the top of my head Java, c#, Javascript, and python are all like this, although python is a little less so. Even in the C/C++ world a whole lot of programs are written to the "whatever is accepted by GCC" standard.

---

• superfunc (orginal) [2014-03-14T18:20:30+01:00] view original

  With regards to C++, I have to disagree. The standard is comprehensive, and gcc/clang/vc++ are all bound to implement it to the letter. Each may have their own extensions, but code written to them, is written to the standard.

---