Hello neko list,

First, I am very impressed with neko. You've done a fantastic job with it.

Second, I would like to use neko as a basis for creating my own language,
which will be a dynamically-typed language with functional features. In
fact, I'm so lazy that I even want to use nekoml as the language in which
I write my translator (to neko code), simply because that means I can copy
the code you've written and change the syntax as my heart desires (it's
pretty much just cosmetic changes + pattern matching [which is also
defined for nekoml so that's nice!]). It goes without saying that should
this ever get the point of release, credit will be given where it is due.

To this end, I would like some clarification as to the build process for
neko. As I understand it:
1. c files get compiled in order to produce the bytecode interpreter -- so
from this point on, in theory we can run ".n" files
2. standard libraries -- based on C-primitives -- get built.
3. neko, and nekoml are built from nekoml source
4. in the tools directory, we create a "nekoc" and "nekomlc" executable
*from the bootstrap bytecode for the relevant compiler*. NOTE: NOT from
the code built in step 3.

So as I see it, step 3 is not doing anything useful. More to the point, my
modus operandi will be to copy the "neko" directory into a new "my-lang"
directory, and alter the makefile so it does pretty much exactly what it
does to neko and nekoml, but to my-lang. Now, once that's done, I have a
bunch of bytecode files which represent my language, scattered in my-lang
directory (for example, Main.n, Lexer.n, etc). But how do I turn all of
those bytecode files into the single "nekoc.n" (or more accurately, a
"my-lang.n") that you find in the "boot" directory? Without this step, I
can't go to step 4 and therefore I can't create my compiler/translator.

Many thanks in advance,
MP

-- 
Neko : One VM to run them all
(http://nekovm.org)

Reply via email to