static int
pg_wc_isupper(pg_wchar c)
{
switch (pg_regex_strategy)
{
case PG_REGEX_LOCALE_C:
return (c <= (pg_wchar) 127 &&
(pg_char_properties[c] & PG_ISUPPER));
case PG_REGEX_LOCALE_WIDE:
#ifdef USE_WIDE_UPPER_LOWER
if (sizeof(wchar_t) >= 4 || c <= (pg_wchar) 0xFFFF)
return iswupper((wint_t) c);
#endif
/* FALL THRU */
case PG_REGEX_LOCALE_1BYTE:
return (c <= (pg_wchar) UCHAR_MAX &&
isupper((unsigned char) c));
case PG_REGEX_LOCALE_WIDE_L:
#if defined(HAVE_LOCALE_T) && defined(USE_WIDE_UPPER_LOWER)
if (sizeof(wchar_t) >= 4 || c <= (pg_wchar) 0xFFFF)
return iswupper_l((wint_t) c, pg_regex_locale);
#endif
/* FALL THRU */
case PG_REGEX_LOCALE_1BYTE_L:
#ifdef HAVE_LOCALE_T
return (c <= (pg_wchar) UCHAR_MAX &&
isupper_l((unsigned char) c, pg_regex_locale));
#endif
break;
}
return 0; /* can't get here, but keep compiler quiet */
}
wint_t towlower_l(wint_t wc, locale_t locale) {
if (iswupper_l(wc, locale)) {
return wc - 'A' + 'a';
}
return wc;
}
void wctype_check_functions(wint_t i, wctype_t t, wctrans_t tr, locale_t l)
{
(void)iswalnum(i);
(void)iswalnum_l(i, l);
(void)iswalpha(i);
(void)iswalpha_l(i, l);
(void)iswblank(i);
(void)iswblank_l(i, l);
(void)iswcntrl(i);
(void)iswcntrl_l(i, l);
(void)iswctype(i, t);
(void)iswctype_l(i, t, l);
(void)iswdigit(i);
(void)iswdigit_l(i, l);
(void)iswgraph(i);
(void)iswgraph_l(i, l);
(void)iswlower(i);
(void)iswlower_l(i, l);
(void)iswprint(i);
(void)iswprint_l(i, l);
(void)iswpunct(i);
(void)iswpunct_l(i, l);
(void)iswspace(i);
(void)iswspace_l(i, l);
(void)iswupper(i);
(void)iswupper_l(i, l);
(void)iswxdigit(i);
(void)iswxdigit_l(i, l);
(void)towctrans(i, tr);
(void)towctrans_l(i, tr, l);
(void)towlower(i);
(void)towlower_l(i, l);
(void)towupper(i);
(void)towupper_l(i, l);
(void)wctrans((const char *)1234);
(void)wctrans_l((const char *)1234, l);
(void)wctype((const char *)1234);
(void)wctype_l((const char *)1234, l);
}
void shorten_name(const char *name,
char short_name[512], char shortest_name[512])
{
wchar_t w_name[512];
wchar_t w_short_name[512];
wchar_t w_shortest_name[512];
const wchar_t *cur_word, *wchar_ptr;
wchar_t *cur_short_word, *cur_shortest_word;
int unabbrev = 0;
int i, len, new_len, capital;
if (!name)
return;
mbsrtowcs_l(w_name, &name, ARRAY_SIZE(w_name), NULL, l);
/* TODO: also skip anything in parenthesis from the short names */
/* TODO: instead of calling wcscasecmp_l all the time it'd be more
* effective to lower case w_name once and use plain wcscmp or mem
* compare since the phrases we search for are all lower case already.
* Might also want to take "collation" into account (wcsxfrm_l the
* string once and use memcmp instead of wcscoll_l).
*/
cur_word = w_name;
cur_short_word = w_short_name;
cur_shortest_word = w_shortest_name;
while (1) {
while (*cur_word && !iswalnum_l(*cur_word, l))
*cur_short_word ++ = *cur_shortest_word ++ = *cur_word ++;
if (!*cur_word)
break;
/* TODO: use a hash of some kind instead of iterating over arrays */
/* Go through possible abbreviations from top to bottom */
for (i = 0; i < ARRAY_SIZE(abbrevs); i += 2)
if (!wcsncasecmp_l(abbrevs[i], cur_word, wcslen(abbrevs[i]), l)) {
len = wcslen(abbrevs[i]);
/* Check that we matched a full word */
if (iswalnum_l(cur_word[len], l))
continue;
capital = iswupper_l(*cur_word, l);
cur_word += len;
new_len = wcslen(abbrevs[i + 1]);
memcpy(cur_short_word, abbrevs[i + 1],
new_len * sizeof(wchar_t));
/*
* If original was capitalised then capitalise the abbreviation
* as well, if it was lower case.
*/
if (capital)
*cur_short_word = towupper_l(*cur_short_word, l);
/* Make sure shortest_word doesn't end up being empty */
if (!*cur_word && !unabbrev) {
memcpy(cur_shortest_word, cur_short_word,
new_len * sizeof(wchar_t));
cur_shortest_word += new_len;
}
cur_short_word += new_len;
/*
* Avoid excess whitespace in short and shortest
* when a word is replaced with "".
* TODO: this may require more complicated logic to get
* the corner cases right.
*/
if (new_len == 0) {
if (cur_short_word > w_short_name &&
iswspace_l(cur_short_word[-1], l))
cur_short_word --;
if (cur_shortest_word > w_shortest_name &&
iswspace_l(cur_shortest_word[-1], l))
cur_shortest_word --;
}
/*if (new_len != len)*/
break;
}
if (i < ARRAY_SIZE(abbrevs))
continue;
/* Go through possible given names from top to bottom */
for (i = 0; i < ARRAY_SIZE(given_names); i ++)
if (!wcsncasecmp_l(given_names[i], cur_word,
wcslen(given_names[i]), l)) {
len = wcslen(given_names[i]);
/* Check that we matched a full word */
if (iswalnum_l(cur_word[len], l))
continue;
/*
* If this is the final part of the name, and it matches a
* given name then that's most likely somebody's surname which
* happens to also be a possibble given name. In that case
* do not abbreviate or omit it.
*/
if (!cur_word[len])
continue;
cur_word += len;
*cur_short_word++ = given_names[i][0];
*cur_short_word++ = L'.';
/*
* Avoid excess whitespace in shortest when a word is
* replaced with "".
* TODO: this may require more complicated logic to get
* the corner cases right.
*/
if (cur_shortest_word > w_shortest_name &&
iswspace_l(cur_shortest_word[-1], l))
cur_shortest_word --;
break;
}
if (i < ARRAY_SIZE(given_names))
continue;
/* Nothing matched, copy the current word as-is */
while (iswalnum_l(*cur_word, l))
*cur_short_word ++ = *cur_shortest_word ++ = *cur_word ++;
unabbrev += 1;
}
*cur_short_word = 0;
*cur_shortest_word = 0;
wchar_ptr = w_short_name;
wcsrtombs_l(short_name, &wchar_ptr, 512, NULL, l);
wchar_ptr = w_shortest_name;
wcsrtombs_l(shortest_name, &wchar_ptr, 512, NULL, l);
}
