Re: How to load more than a file from C

[Sean Conner <sean@...>]

It was thus said that the Great Igor Trevisan once stated:
> Hi Thijs,
> 
> thanks for answering!
> 
> On 30 April 2014 09:48, Thijs Schreijer <thijs@thijsschreijer.nl> wrote:
> 
> >From your description it seems that you "run files" and not specifically
> > Lua code or functions. Especially your wording "meld them in a single
> > bytecode".
> >
> I want to execute a function witten in Lua calling it from my
> main application written in C.
> 
> [...]
> 
> 
> > And though I don't understand exactly what and how you're trying to do
> 
> 
> I recall the example I used answering to a previous email in the thread.
> I hope it can help.
> 
> Suppose I have defined func_y in file_y.lua
> and I want to use it within func_x that's in file_x.lua.
>  func_x is the main lua function that has to be called and
> executed from my C code.
> file_x.lua and file_y.lua are both in a USB stick used
> in my embedded system to make lua functionalities avaliable
> when needed.
> Now the question is: how can I load and execute successfully
> func_x?

  Somehow, your C application will need to know where file_x.lua resides. 
For sake of argument, let's just pass it in on the command line.  So we
have (I'm assuming Lua 5.2):

#include <stdlib.h>
#include <stdio.h>
#include <lua.h>
#include <lauxlib.h>
#include <lualib.h>

int main(int argc,char *argv[])
{
  lua_State *L;
  int        rc;
  
  /* check to see if Lua file specified */

  if (argc == 1)
  {
    fprintf(stderr,"usage: %s luafile\n",argv[0]);
    return EXIT_FAILURE;
  }

  /* create Lua state */

  L = luaL_newstate();
  if (L == NULL)
  {
    fprintf(stderr,"could not initialize Lua\n");
    return EXIT_FAILURE;
  }

  /* initialize standard Lua modules like io and math */

  luaL_openlibs(L);

  /* load our file */

  rc = luaL_loadfile(L,argv[1]);
  if (rc != LUA_OK)
  {
    const char *errmsg = lua_tostring(L,-1);
    fprintf(stderr,"%s: %s\n",argv[1],errmsg);
    lua_close(L);
    return EXIT_FAILURE;
  }

  /* load the file */

  rc = lua_pcall(L,0,0,0);
  if (rc != LUA_OK)
  {
    const char *errmsg = lua_tostring(L,-1);
    fprintf(stderr,"executing %s: %s\n",argv[1],errmsg);
    lua_close(L);
    return EXIT_FAILURE;
  }

  /* now call our function */

  lua_getglobal(L,"func_x");
  rc = lua_pcall(L,0,0,0);
  if (rc != LUA_OK)
  {
    const char *errmsg = lua_tostring(L,-1);
    fprintf(stderr,"func_x() = %s",errmsg);
    lua_close(L);
    return EXIT_FAILURE;
  }

  lua_close(L);
  return EXIT_SUCCESS;
}

That's fine if file_x.lua doesn't require other Lua files.  If it does, then
things get ... system specific.  file_x.lua needs to know where file_y.lua
resides, and there are multiple ways of having file_x.lua load the file.  It
can do:

	dofile "file_y.lua"

or

	do
	  local x = loadfile "file_y.lua"
	  x()
	end

or

	require "file_y.lua"

  The first two are pretty much equivalent and assume that no path is
required to specify the file.  The last depends upon how the environment
variable LUA_PATH is set, and the internal value package.path.  

  -spc