Question

Given a counted string (either an array of characters, or a wrapper like std::string), is there a “proper” way to escape and/or unescape it in C or C++, such that “special” characters (like the null character) become C-style-escaped and “normal” characters stay the way they are?

Or do I have to do it by hand?

Answer 1

This is a function to process a single character:

/*
** Does not generate hex character constants.
** Always generates triple-digit octal constants.
** Always generates escapes in preference to octal.
** Escape question mark to ensure no trigraphs are generated by repetitive use.
** Handling of 0x80..0xFF is locale-dependent (might be octal, might be literal).
*/

void chr_cstrlit(unsigned char u, char *buffer, size_t buflen)
{
    if (buflen < 2)
        *buffer = '\0';
    else if (isprint(u) && u != '\'' && u != '\"' && u != '\\' && u != '\?')
        sprintf(buffer, "%c", u);
    else if (buflen < 3)
        *buffer = '\0';
    else
    {
        switch (u)
        {
        case '\a':  strcpy(buffer, "\\a"); break;
        case '\b':  strcpy(buffer, "\\b"); break;
        case '\f':  strcpy(buffer, "\\f"); break;
        case '\n':  strcpy(buffer, "\\n"); break;
        case '\r':  strcpy(buffer, "\\r"); break;
        case '\t':  strcpy(buffer, "\\t"); break;
        case '\v':  strcpy(buffer, "\\v"); break;
        case '\\':  strcpy(buffer, "\\\\"); break;
        case '\'':  strcpy(buffer, "\\'"); break;
        case '\"':  strcpy(buffer, "\\\""); break;
        case '\?':  strcpy(buffer, "\\\?"); break;
        default:
            if (buflen < 5)
                *buffer = '\0';
            else
                sprintf(buffer, "\\%03o", u);
            break;
        }
    }
}

And this is the code to handle a null-terminated string (using the function above):

void str_cstrlit(const char *str, char *buffer, size_t buflen)
{
    unsigned char u;
    size_t len;

    while ((u = (unsigned char)*str++) != '\0')
    {
        chr_cstrlit(u, buffer, buflen);
        if ((len = strlen(buffer)) == 0)
            return;
        buffer += len;
        buflen -= len;
    }
    *buffer = '\0';
}