I tried to convert a string from Lua(5.1.5) to an integer, and check if the number is a valid integer(0~99999). However, I found lua_tonumber() has different behavior from lua_tointeger() when dealing with big integer.
int main()
{
int in;
double db;
lua_State* Lua = luaL_newstate();
luaL_openlibs(Lua);
lua_pushstring(Lua, "213232127162767162736718238168263816873");
db = lua_tonumber(Lua, -1);
in = lua_tointeger(Lua, -1);
printf("DOUBLE:%f\n", db); // DOUBLE:213232127162767176000119210017101447168.000000
printf("INT:%d\n", in); // INT:0
};
If I use lua_tointeger(), it returns 0 and will pass my checking.
I check both API description, but I still don't know why they have different behaviors. Are these behaviors machine-independent? Is using lua_tonumber() a better way?
Can I use the following code to check the result? (cross-platform)
if (!lua_isnumber(Lua, -1)) { //error }
result = lua_tonumber(Lua, -1);
if (result < 0 || result > 99999) { // error2 }
// pass
From the manual reference of lua_tointeger
:
Converts the Lua value at the given acceptable index to the signed integral type
lua_Integer
. The Lua value must be a number or a string convertible to a number (see §2.2.1); otherwise,lua_tointeger
returns0
The same is with lua_tonumber
. So when you get the return value of 0
means you are passing a string that is non-convertible. The number 213232127162767162736718238168263816873
is too big for an lua_Integer
(the same as ptrdiff_t
by default) type.
Are these behaviors machine-independent? Is using lua_tonumber() a better way?
In a way, it is machine-dependent, because the type lua_Integer
is machine-dependent. Again, the number 213232127162767162736718238168263816873
is too big for any normal integer type. Whether using lua_tonumber
or using lua_tointeger
is up to your need. If converting an integer that is less than 32-bit, for example, 0~99999 in your question, lua_tointeger
is your choice. It is very fast in a IEEE-754 compatible machine. The disadvantage of using lua_tonumber
is that it may mis-convert non-integer values.
Reference: A fast method to round a double to a 32-bit int explained
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With