/**
* Upcase the first letter of the word.
* XXX FIXME This works 'most of the time', but is not technically correct.
* This is because towlower() and towupper() are locale dependent, and also
* because the byte-counts might not match up, e.g. German ß and SS.
* The correct long-term fix is to use ICU or glib g_utf8_strup(), etc.
*/
void upcase_utf8_str(char *to, const char * from, size_t usize)
{
wchar_t c;
int i, nbl, nbh;
char low[MB_LEN_MAX];
mbstate_t mbs;
memset(&mbs, 0, sizeof(mbs));
nbh = mbrtowc (&c, from, MB_CUR_MAX, &mbs);
if (nbh < 0)
{
prt_error("Error: Invalid multi-byte string!");
return;
}
c = towupper(c);
nbl = wctomb_check(low, c);
/* Check for error on an in-place copy */
if ((nbh < nbl) && (to == from))
{
/* I'm to lazy to fix this */
prt_error("Error: can't upcase multi-byte string!");
return;
}
/* Upcase */
for (i=0; i<nbl; i++) { to[i] = low[i]; }
if ((nbh == nbl) && (to == from)) return;
from += nbh;
to += nbl;
safe_strcpy(to, from, usize-nbl);
}
/**
* Reads in one word from the file, allocates space for it,
* and returns it.
*/
static const char * get_a_word(Dictionary dict, FILE * fp)
{
char word[MAX_WORD+4]; /* allow for 4-byte wide chars */
const char * s;
wint_t c;
mbstate_t mbss;
int j;
do {
c = fgetwc(fp);
} while ((c != WEOF) && iswspace(c));
if (c == WEOF) return NULL;
memset(&mbss, 0, sizeof(mbss));
for (j=0; (j <= MAX_WORD-1) && (!iswspace(c)) && (c != WEOF);)
{
j += wctomb_check(&word[j], c, &mbss);
c = fgetwc(fp);
}
if (j >= MAX_WORD) {
word[MAX_WORD] = 0x0;
prt_error("Fatal Error: The dictionary contains a word that "
"is too long. The word was: %s", word);
exit(1);
}
word[j] = '\0';
s = string_set_add(word, dict->string_set);
return s;
}
/**
* Downcase the first letter of the word.
* XXX FIXME This works 'most of the time', but is not technically correct.
* This is because towlower() and towupper() are locale dependent, and also
* because the byte-counts might not match up, e.g. German ß and SS.
* The correct long-term fix is to use ICU or glib g_utf8_strup(), etc.
*/
void downcase_utf8_str(char *to, const char * from, size_t usize, locale_t locale)
{
wchar_t c;
int i, nbl, nbh;
char low[MB_LEN_MAX];
mbstate_t mbs;
/* Make sure it doesn't contain garbage in case of an error */
if (to != from) strcpy(to, from);
memset(&mbs, 0, sizeof(mbs));
nbh = mbrtowc (&c, from, MB_CUR_MAX, &mbs);
if (nbh < 0)
{
prt_error("Error: Invalid UTF-8 string!\n");
return;
}
c = towlower_l(c, locale);
nbl = wctomb_check(low, c);
/* Check for error on an in-place copy */
if ((nbh < nbl) && (to == from))
{
/* I'm to lazy to fix this */
prt_error("Error: can't downcase UTF-8 string!\n");
return;
}
/* Downcase */
for (i=0; i<nbl; i++) { to[i] = low[i]; }
if ((nbh == nbl) && (to == from)) return;
from += nbh;
to += nbl;
safe_strcpy(to, from, usize-nbl);
}
/**
* This reads the next token from the input into token.
* Return 1 if a character was read, else return 0 (and print a warning).
*/
static int link_advance(Dictionary dict)
{
wchar_t c;
int nr, i;
int quote_mode;
dict->is_special = FALSE;
if (dict->already_got_it != '\0')
{
dict->is_special = is_special(dict->already_got_it, &dict->mbss);
if (dict->already_got_it == WEOF) {
dict->token[0] = '\0';
} else {
dict->token[0] = dict->already_got_it; /* specials are one byte */
dict->token[1] = '\0';
}
dict->already_got_it = '\0';
return 1;
}
do { c = get_character(dict, FALSE); } while (iswspace(c));
quote_mode = FALSE;
i = 0;
for (;;)
{
if (i > MAX_TOKEN_LENGTH-3) { /* 3 for multi-byte tokens */
dict_error(dict, "Token too long");
return 0;
}
if (quote_mode) {
if (c == '\"') {
quote_mode = FALSE;
dict->token[i] = '\0';
return 1;
}
if (iswspace(c)) {
dict_error(dict, "White space inside of token");
return 0;
}
/* Although we read wide chars, we store UTF8 internally, always. */
nr = wcrtomb(&dict->token[i], c, &dict->mbss);
if (nr < 0) {
#ifndef _WIN32
dict_error2(dict, "Unable to read UTF8 string in current locale",
nl_langinfo(CODESET));
fprintf (stderr, "\tTry setting the locale with \"export LANG=en_US.UTF-8\"\n");
#else
dict_error(dict, "Unable to read UTF8 string in current locale");
#endif
return 0;
}
i += nr;
} else {
if (is_special(c, &dict->mbss))
{
if (i == 0)
{
dict->token[0] = c; /* special toks are one char always */
dict->token[1] = '\0';
dict->is_special = TRUE;
return 1;
}
dict->token[i] = '\0';
dict->already_got_it = c;
return 1;
}
if (c == 0x0) {
if (i == 0) {
dict->token[0] = '\0';
return 1;
}
dict->token[i] = '\0';
dict->already_got_it = c;
return 1;
}
if (iswspace(c)) {
dict->token[i] = '\0';
return 1;
}
if (c == '\"') {
quote_mode = TRUE;
} else {
/* store UTF8 internally, always. */
nr = wctomb_check(&dict->token[i], c, &dict->mbss);
if (nr < 0) {
#ifndef _WIN32
dict_error2(dict, "Unable to read UTF8 string in current locale",
nl_langinfo(CODESET));
fprintf (stderr, "\tTry setting the locale with \"export LANG=en_US.UTF-8\"\n");
#else
dict_error(dict, "Unable to read UTF8 string in current locale");
#endif
return 0;
}
i += nr;
}
}
c = get_character(dict, quote_mode);
}
return 1;
}
