How do you overlap text in C terminal

Question

One of the training exercises in my C programming book goes like following: (translated from swedish)

If you use a writing/typing terminal you kan sometimes, just like a normal typewriter, underline a word by stepping back on the line and writing _ under the word. Write a program that writes the string in the variable s underlined.

my attempt looks like this:

#include <stdio.h>

int main(void) {
  char *s;
  printf("input: ");
  scanf("%ms", &s);
  for (int i=0; s[i] != '\0'; i++) {
    printf("%c\b_", s[i]);
  }
  printf("\n");
  return 0;
}

This does not work i expect terminal output to be:

$> gcc -Wall -std=c99 file.c  
$> ./file.c  
input: test  
test //underlined
$>

but instead i get:

$> gcc -Wall -std=c99 file.c  
$> ./file.c  
input: test  
____  
$> 

does anyone know how to do that or is the book wrong?

Answer 1

I haven't found a way to allow the literal rendering of one character over another, sadly (I didn't think this was possible with modern terminals, but I'm happy to be proven wrong as Wumpus Q. Wumbley has made the point that piping through less allows for this. I've not had time to look at how/why it works within less though).

I'd like to propose a workaround - there is a way to enable underscoring a little more directly using the ANSI code \033[4m.

Modifying your code:

#include <stdio.h>

int main()
{
  char *s;

  printf("input: ");
  scanf("%ms", &s);
  printf("\033[4m"); // Enable underlining
  printf("%s", s); // Print your string
  printf("\033[0m"); // Reset text attributes (disable underlining)
  printf("\n");
  return 0;
}

You can find the other ANSI graphic settings in the SGR table in the CSI Codes section on wikipedia, here.

The \033 specifies it's a terminal escape code, the combination of \e with [ specifies it's a terminal control signal, then the <number>m specifies we're doing an "SGR", "Select Graphic Rendition" command, and the number 4 as a parameter corresponds to an effect "Underline: Single". The 0 corresponds to a reset.

I welcome any corrections though, as I don't know much about the ANSI code format.

---

I've just skimmed through the less source, and I suspect it's converting sequences like _\ba into UTF underlined characters - but I haven't quite got to the bottom of it yet.