static void init_special_chars()
{
const char *p;
for (p = "':^`~"; *p; p++)
for (const char *q = "aeiouy"; *q; q++) {
// Use a variable to work around bug in gcc 2.0
char c = cmupper(*q);
init_two_char_letter(*p, *q, *p, c);
}
for (p = "/l/o~n,coeaeij"; *p; p += 2) {
// Use variables to work around bug in gcc 2.0
char c0 = cmupper(p[0]);
char c1 = cmupper(p[1]);
init_two_char_letter(p[0], p[1], c0, c1);
}
init_two_char_letter('v', 's', 'v', 'S', "s");
init_two_char_letter('v', 'z', 'v', 'Z', "z");
init_two_char_letter('o', 'a', 'o', 'A', "a");
init_two_char_letter('T', 'p', 'T', 'P', THORN_SORT_KEY);
init_two_char_letter('-', 'd', '-', 'D');
store_token("\\(ss", TOKEN_LOWER, 0, "SS");
store_token("\\[ss]", TOKEN_LOWER, 0, "SS");
store_token("\\(Sd", TOKEN_LOWER, "d", "\\(-D");
store_token("\\[Sd]", TOKEN_LOWER, "d", "\\[-D]");
store_token("\\(hy", TOKEN_HYPHEN);
store_token("\\[hy]", TOKEN_HYPHEN);
store_token("\\(en", TOKEN_RANGE_SEP);
store_token("\\[en]", TOKEN_RANGE_SEP);
}
static void init_ascii()
{
const char *p;
for (p = "abcdefghijklmnopqrstuvwxyz"; *p; p++) {
char buf[2];
buf[0] = *p;
buf[1] = '\0';
store_token(strsave(buf), TOKEN_LOWER);
buf[0] = cmupper(buf[0]);
store_token(strsave(buf), TOKEN_UPPER);
}
for (p = "0123456789"; *p; p++) {
char buf[2];
buf[0] = *p;
buf[1] = '\0';
const char *s = strsave(buf);
store_token(s, TOKEN_OTHER, s);
}
for (p = ".,:;?!"; *p; p++) {
char buf[2];
buf[0] = *p;
buf[1] = '\0';
store_token(strsave(buf), TOKEN_PUNCT);
}
store_token("-", TOKEN_HYPHEN);
}
void token_info::upper_case(const char *start, const char *end,
string &result) const
{
if (type != TOKEN_LOWER) {
while (start < end)
result += *start++;
}
else if (other_case)
result += other_case;
else {
while (start < end)
result += cmupper(*start++);
}
}
map_init::map_init()
{
int i;
for (i = 0; i < 256; i++)
map[i] = csalnum(i) ? cmlower(i) : '\0';
for (i = 0; i < 256; i++) {
if (cslower(i)) {
inv_map[i][0] = i;
inv_map[i][1] = cmupper(i);
inv_map[i][2] = '\0';
}
else if (csdigit(i)) {
inv_map[i][0] = i;
inv_map[i][1] = 0;
}
else
inv_map[i][0] = '\0';
}
}