Chapter 14: Processing and I/O

Level 2 — Cognition

Text processing, formatted I/O, character encodings -- how C handles data flowing in and out, and how Auto simplifies it.

Input and output are where programs meet the outside world. C's I/O library, centered on stdio.h, provides a flexible but error-prone set of tools. This chapter covers text processing, formatted input, character encodings, and binary streams -- and how Auto provides safer alternatives.

---

14.1 Text Processing

C provides character classification functions in <ctype.h>:

// C Deep Dive: character classification
#include <ctype.h>
int ch = 'A';
isalpha(ch);    // true: alphabetic
isdigit(ch);    // false: not a digit
isupper(ch);    // true: uppercase
islower(ch);    // false: not lowercase
isspace(ch);    // false: not whitespace
toupper(ch);    // 'A' -> 'A'
tolower(ch);    // 'A' -> 'a'

These functions have a critical caveat: passing a value not representable as unsigned char or EOF is undefined behavior. In practice, this means passing a char from a platform where char is signed can cause UB for characters with values above 127.

String operations in <string.h>:

// C Deep Dive: string operations
strlen("hello");           // 5
strcpy(dst, src);          // copy -- dst must be large enough!
strcat(dst, src);          // concatenate -- dst must be large enough!
strcmp("abc", "def");      // negative: abc < def
strncmp("abc", "abd", 2);  // 0: first 2 chars equal
strchr("hello", 'l');      // pointer to first 'l'
strstr("hello", "ell");    // pointer to "ell"

The strcpy and strcat functions are responsible for many buffer overflow vulnerabilities. C11 added strncpy_s and strcat_s as safer alternatives, but they remain underused.

Auto's approach: safe string operations. Auto strings are length-tracked and immutable. Operations that could overflow simply do not arise:

// Auto: safe text processing
fn count_words(text str) int {
    var count int = 0
    var in_word bool = false
    for i in 0..len(text) {
        let ch str = str(text[i])
        if ch == " " {
            in_word = false
        } else {
            if !in_word {
                count = count + 1
            }
            in_word = true
        }
    }
    count
}