Question

I want to build a function to easily convert a string containing hex code (eg. “0ae34e”) into a string containing the equivalent ascii values and vice versa.
Do I have to cut the Hex string in pairs of 2 values and gue them together again or is there a convenient way to do that?

thanks

Answer 1

Based on binascii_unhexlify() function from Python:

#include <cctype> // is*

int to_int(int c) {
  if (not isxdigit(c)) return -1; // error: non-hexadecimal digit found
  if (isdigit(c)) return c - '0';
  if (isupper(c)) c = tolower(c);
  return c - 'a' + 10;
}

template<class InputIterator, class OutputIterator> int
unhexlify(InputIterator first, InputIterator last, OutputIterator ascii) {
  while (first != last) {
    int top = to_int(*first++);
    int bot = to_int(*first++);
    if (top == -1 or bot == -1)
      return -1; // error
    *ascii++ = (top << 4) + bot;
  }
  return 0;
}

Example

#include <iostream>

int main() {
  char hex[] = "7B5a7D";
  size_t len = sizeof(hex) - 1; // strlen
  char ascii[len/2+1];
  ascii[len/2] = '\0';

  if (unhexlify(hex, hex+len, ascii) < 0) return 1; // error
  std::cout << hex << " -> " << ascii << std::endl;
}

Output

7B5a7D -> {Z}

An interesting quote from the comments in the source code:

While I was reading dozens of programs that encode or decode the
formats here (documentation? hihi:-) I have formulated Jansen’s
Observation:

Programs that encode binary data in ASCII are written in such a style
that they are as unreadable as possible. Devices used include
unnecessary global variables, burying important tables in unrelated
sourcefiles, putting functions in include files, using
seemingly-descriptive variable names for different purposes, calls to
empty subroutines and a host of others.

I have attempted to break with this tradition, but I guess that that
does make the performance sub-optimal. Oh well, too bad…

Jack Jansen, CWI, July 1995.