Re: Slight discrepency in the documentation for

[Mark Edgar <medgar@...>]

Eric Raible wrote:
> The first return value of pairs(t) is rather a private function that
> invokes that same code (luaB_next) that 'next' invokes.  But it does
> _not_ return 'next' per se.
>
> So either the documentation should be fixed, or the following patch
> could be applied which makes the reference manual true.  This patch
> both the advantage and the disadvantage that if the value of 'next'
> is changed, then pairs(t) will see that change.

Your patch will break if "next" is not available in the global table or 
is reassigned to something else for some reason.

The following should work (untested!):
--- lbaselib.c.orig     2006-10-09 12:29:55.000000000 -0700
+++ lbaselib.c  2007-02-07 16:13:09.000000000 -0700
@@ -621,11 +621,13 @@
   luaL_register(L, "_G", base_funcs);
   lua_pushliteral(L, LUA_VERSION);
   lua_setglobal(L, "_VERSION");  /* set global _VERSION */
   /* `ipairs' and `pairs' need auxliliary functions as upvalues */
   auxopen(L, "ipairs", luaB_ipairs, ipairsaux);
-  auxopen(L, "pairs", luaB_pairs, luaB_next);
+  lua_getfield(L, -1, "next");
+  lua_pushcclosure(L, luaB_pairs, 1);
+  lua_setfield(L, -2, "pairs");
   /* `newproxy' needs a weaktable as upvalue */
   lua_createtable(L, 0, 1);  /* new table `w' */
   lua_pushvalue(L, -1);  /* `w' will be its own metatable */
   lua_setmetatable(L, -2);
   lua_pushliteral(L, "kv");

With this change, the pairs upvalue is same function as _G.next instead 
of a new one which refers to the same C function.

					-Mark