Long Long, decimals and input validation in C

Question

Currently I'm using TCC as it's the easiest thing to get setup on windows. Simply unzip and you're ready to go. However I'm open to other compilers, GCC, whatever microsoft has on offer etc.

My problem is that I need to validate the input to a size 16 array of integers. Here are my rules:

if number is under 15 (including negative values) then input is valid
if number is under -2147483648 then -2147483648

if number is over 2147483647 then 15
else if number is over 15 then mod 16

if the number is a decimal, remove decimal point and validate again

Considering I'm using C, the last point scares me, and I'll come back to it later. For now I'm just trying to handle the first 4 conditions.

The problem I'm running into is that trying to test for the outer bounds results in Integer overflows and screws up my checks. So I've made a temporary array of long longs to hold the input for validation purposes. The moment everything is successfully validated it should fit in an array of Integers, so I will (somehow) copy the long longs from the temp array to the actual one and start the program as normal.

I've messed around with long longs and trying to do what I want to do, but my code is getting messy fast and everything is so vague and machine dependant in C so when something goes wrong I don't know whether it's me and my crappy coding, or the fact that my machine is different to everyone else's that is causing the error. I am going to stick at it cause I know this sort of thing can be investigated and worked out, however I don't want to waste too much time on it so I'll ask SO and see if there's a shortcut.

The decimal validation part I've got various ideas on how to approach, but I'm not hopeful. What's your opinion?

Anyone who wants to know why I'm doing this: It doesn't matter, I can solve the higher level problem that requires this array quite easily and it will work for all valid inputs. However I'm just being pedantic right now, hence this question.

Answer 1

First, your conditions have some problems. If a number is under -2147483648, it's also under 15, so that check never matches (neither being a decimal for numbers under 15).

Second, you can check for overflows with strtol (check errno for ERANGE) and then compare with your limits (though there's no need if your long has 32-bits and is in two's complement).

As for decimals, if you always want to remove the decimal point (not what you're saying you want because you condition that on a series of other conditions failing), you can setup a preprocessing step that removes periods from the string. It can easily be done in-place with two pointers – a read pointer and a write pointer.