In practice it's _worse_ than that because you probably don't want a "long", you probably want a particular size like a 64 bit integer. So you have to add ifdefs to call either strtol or strtoll depending on the size of "long" and "long long".
And if you are using base 16 then strtol will allow an optional "0x" prefix. So if you didn't want that you have to check for it manually.
Strtol also accepts leading whitespace so if you didn't want that you have to test manually for it.
Don't pass a zero base thinking it means base ten. This works almost all the time but misinterprets a leading zero to mean octal.
>> you probably don't want a "long", you probably want a particular size like a 64 bit integer. So you have to add ifdefs to call either strtol or strtoll depending on the size of "long" and "long long".
Great, and what happens when you accidentally assign something greater than 2^31 to an int32_t when using strtol? You won't benefit from a range error if LONG_MAX is 2^63, and now you have to make sure to handle any implementation defined behavior.
That's true. If the range you're interested in is not the same as the range of long or long long, you'll have to check the range yourself before you go on and use the value. No ifdefs required. If you're not happy doing it yourself, I do recommend strtonum or any of the alternatives that allow you to explicitly specify the range you're interested in. I don't see the point in littering the standard library with functions having hard-coded range for every range you might be interested in.
And if you are using base 16 then strtol will allow an optional "0x" prefix. So if you didn't want that you have to check for it manually.
Strtol also accepts leading whitespace so if you didn't want that you have to test manually for it.
Don't pass a zero base thinking it means base ten. This works almost all the time but misinterprets a leading zero to mean octal.
Good luck!