//input: *c=='[' **pc==':'
static u16 bracket_class(u8 *c,u8 **pc,u8 **sc,u8 not,u8 sc_folded)
{
u8 char_class[CHAR_CLASS_MAX+1];//don't forget the 0 terminating char
u16 r=bracket_char_class_get(c,pc,not,sc_folded,&char_class[0]);
if(r!=OK) return r;
if((STREQ(char_class,"alnum")&&is_alnum(**sc))
||(STREQ(char_class,"alpha")&&is_alpha(**sc))
||(STREQ(char_class,"blank")&&is_blank(**sc))
||(STREQ(char_class,"cntrl")&&is_cntrl(**sc))
||(STREQ(char_class,"digit")&&is_digit(**sc))
||(STREQ(char_class,"graph")&&is_graph(**sc))
||(STREQ(char_class,"lower")&&is_lower(**sc))
||(STREQ(char_class,"print")&&is_print(**sc))
||(STREQ(char_class,"punct")&&is_punct(**sc))
||(STREQ(char_class,"space")&&is_space(**sc))
||(STREQ(char_class,"upper")&&is_upper(**sc))
||(STREQ(char_class,"xdigit")&&is_xdigit(**sc)))
return bracket_matched(c,pc,not);
*c=*(*pc)++;
return OK;
}
void init_tty(void)
{
#ifdef USE_HPUX_TERMIO
struct ltchars chars;
#endif
#ifdef USE_SGTTY
struct ltchars chars;
#endif
char bs[2], dline[2], bword[2], refresh[2], lnext[2];
bs[0] = dline[0] = bword[0] = refresh[0] = lnext[0] = '\0';
no_tty = !isatty(STDIN_FILENO) || !isatty(STDOUT_FILENO);
cbreak_noecho_mode();
if (!no_tty) {
#ifdef USE_TERMIOS
*bs = old_tty.c_cc[VERASE];
*dline = old_tty.c_cc[VKILL];
# ifdef VWERASE /* Not POSIX, but many systems have it. */
*bword = old_tty.c_cc[VWERASE];
# endif
# ifdef VREPRINT /* Not POSIX, but many systems have it. */
*refresh = old_tty.c_cc[VREPRINT];
# endif
# ifdef VLNEXT /* Not POSIX, but many systems have it. */
*lnext = old_tty.c_cc[VLNEXT];
# endif
#endif
#ifdef USE_HPUX_TERMIO
*bs = old_tty.c_cc[VERASE];
*dline = old_tty.c_cc[VKILL];
if (ioctl(STDIN_FILENO, TIOCGLTC, &chars) < 0) perror("TIOCGLTC ioctl");
else {
*bword = chars.t_werasc;
*refresh = chars.t_rprntc;
/* *lnext = chars.t_lnextc; */ /* ?? Screw it, use default. */
}
#endif
#ifdef USE_SGTTY
*bs = old_tty.sg_erase;
*dline = old_tty.sg_kill;
if (ioctl(STDIN_FILENO, TIOCGLTC, &chars) < 0) perror("TIOCGLTC ioctl");
else {
*bword = chars.t_werasc;
*refresh = chars.t_rprntc;
*lnext = chars.t_lnextc;
}
#endif
}
bs[1] = dline[1] = bword[1] = refresh[1] = lnext[1] = '\0';
/* Note that some systems use \0 to disable, others use \377; we must
* check both. Also, some seem to leave garbage in some of the fields,
* so we'll ignore anything that isn't a control character.
*/
if (is_cntrl(*bs) && *bs && *bs != '\b' && *bs != '\177')
add_ibind(bs, "/DOKEY BSPC");
if (is_cntrl(*bword) && *bword) add_ibind(bword, "/DOKEY BWORD");
/* if (is_cntrl(*dline) && *dline) add_ibind(dline, "/DOKEY DLINE"); */
if (is_cntrl(*refresh) && *refresh) add_ibind(refresh, "/DOKEY REFRESH");
if (is_cntrl(*lnext) && *lnext) add_ibind(lnext, "/DOKEY LNEXT");
}
static void
output_tables (const char *filename, const char *version)
{
FILE *stream;
unsigned int ch;
stream = fopen (filename, "w");
if (stream == NULL)
{
fprintf (stderr, "cannot open '%s' for writing\n", filename);
exit (1);
}
fprintf (stream, "escape_char /\n");
fprintf (stream, "comment_char %%\n");
fprintf (stream, "\n");
fprintf (stream, "%% Generated automatically by gen-unicode-ctype for Unicode %s.\n",
version);
fprintf (stream, "\n");
fprintf (stream, "LC_IDENTIFICATION\n");
fprintf (stream, "title \"Unicode %s FDCC-set\"\n", version);
fprintf (stream, "source \"UnicodeData.txt, PropList.txt\"\n");
fprintf (stream, "address \"\"\n");
fprintf (stream, "contact \"\"\n");
fprintf (stream, "email \"[email protected]\"\n");
fprintf (stream, "tel \"\"\n");
fprintf (stream, "fax \"\"\n");
fprintf (stream, "language \"\"\n");
fprintf (stream, "territory \"Earth\"\n");
fprintf (stream, "revision \"%s\"\n", version);
{
time_t now;
char date[11];
now = time (NULL);
strftime (date, sizeof (date), "%Y-%m-%d", gmtime (&now));
fprintf (stream, "date \"%s\"\n", date);
}
fprintf (stream, "category \"unicode:2001\";LC_CTYPE\n");
fprintf (stream, "END LC_IDENTIFICATION\n");
fprintf (stream, "\n");
/* Verifications. */
for (ch = 0; ch < 0x110000; ch++)
{
/* toupper restriction: "Only characters specified for the keywords
lower and upper shall be specified. */
if (to_upper (ch) != ch && !(is_lower (ch) || is_upper (ch)))
fprintf (stderr,
"%s is not upper|lower but toupper(0x%04X) = 0x%04X\n",
ucs_symbol (ch), ch, to_upper (ch));
/* tolower restriction: "Only characters specified for the keywords
lower and upper shall be specified. */
if (to_lower (ch) != ch && !(is_lower (ch) || is_upper (ch)))
fprintf (stderr,
"%s is not upper|lower but tolower(0x%04X) = 0x%04X\n",
ucs_symbol (ch), ch, to_lower (ch));
/* alpha restriction: "Characters classified as either upper or lower
shall automatically belong to this class. */
if ((is_lower (ch) || is_upper (ch)) && !is_alpha (ch))
fprintf (stderr, "%s is upper|lower but not alpha\n", ucs_symbol (ch));
/* alpha restriction: "No character specified for the keywords cntrl,
digit, punct or space shall be specified." */
if (is_alpha (ch) && is_cntrl (ch))
fprintf (stderr, "%s is alpha and cntrl\n", ucs_symbol (ch));
if (is_alpha (ch) && is_digit (ch))
fprintf (stderr, "%s is alpha and digit\n", ucs_symbol (ch));
if (is_alpha (ch) && is_punct (ch))
fprintf (stderr, "%s is alpha and punct\n", ucs_symbol (ch));
if (is_alpha (ch) && is_space (ch))
fprintf (stderr, "%s is alpha and space\n", ucs_symbol (ch));
/* space restriction: "No character specified for the keywords upper,
lower, alpha, digit, graph or xdigit shall be specified."
upper, lower, alpha already checked above. */
if (is_space (ch) && is_digit (ch))
fprintf (stderr, "%s is space and digit\n", ucs_symbol (ch));
if (is_space (ch) && is_graph (ch))
fprintf (stderr, "%s is space and graph\n", ucs_symbol (ch));
if (is_space (ch) && is_xdigit (ch))
fprintf (stderr, "%s is space and xdigit\n", ucs_symbol (ch));
/* cntrl restriction: "No character specified for the keywords upper,
lower, alpha, digit, punct, graph, print or xdigit shall be
specified." upper, lower, alpha already checked above. */
if (is_cntrl (ch) && is_digit (ch))
fprintf (stderr, "%s is cntrl and digit\n", ucs_symbol (ch));
if (is_cntrl (ch) && is_punct (ch))
fprintf (stderr, "%s is cntrl and punct\n", ucs_symbol (ch));
if (is_cntrl (ch) && is_graph (ch))
fprintf (stderr, "%s is cntrl and graph\n", ucs_symbol (ch));
if (is_cntrl (ch) && is_print (ch))
fprintf (stderr, "%s is cntrl and print\n", ucs_symbol (ch));
if (is_cntrl (ch) && is_xdigit (ch))
fprintf (stderr, "%s is cntrl and xdigit\n", ucs_symbol (ch));
/* punct restriction: "No character specified for the keywords upper,
lower, alpha, digit, cntrl, xdigit or as the <space> character shall
be specified." upper, lower, alpha, cntrl already checked above. */
if (is_punct (ch) && is_digit (ch))
fprintf (stderr, "%s is punct and digit\n", ucs_symbol (ch));
if (is_punct (ch) && is_xdigit (ch))
fprintf (stderr, "%s is punct and xdigit\n", ucs_symbol (ch));
if (is_punct (ch) && (ch == 0x0020))
fprintf (stderr, "%s is punct\n", ucs_symbol (ch));
/* graph restriction: "No character specified for the keyword cntrl
shall be specified." Already checked above. */
/* print restriction: "No character specified for the keyword cntrl
shall be specified." Already checked above. */
/* graph - print relation: differ only in the <space> character.
How is this possible if there are more than one space character?!
I think susv2/xbd/locale.html should speak of "space characters",
not "space character". */
if (is_print (ch) && !(is_graph (ch) || /* ch == 0x0020 */ is_space (ch)))
fprintf (stderr,
"%s is print but not graph|<space>\n", ucs_symbol (ch));
if (!is_print (ch) && (is_graph (ch) || ch == 0x0020))
fprintf (stderr,
"%s is graph|<space> but not print\n", ucs_symbol (ch));
}
fprintf (stream, "LC_CTYPE\n");
output_charclass (stream, "upper", is_upper);
output_charclass (stream, "lower", is_lower);
output_charclass (stream, "alpha", is_alpha);
output_charclass (stream, "digit", is_digit);
output_charclass (stream, "outdigit", is_outdigit);
output_charclass (stream, "blank", is_blank);
output_charclass (stream, "space", is_space);
output_charclass (stream, "cntrl", is_cntrl);
output_charclass (stream, "punct", is_punct);
output_charclass (stream, "xdigit", is_xdigit);
output_charclass (stream, "graph", is_graph);
output_charclass (stream, "print", is_print);
output_charclass (stream, "class \"combining\";", is_combining);
output_charclass (stream, "class \"combining_level3\";", is_combining_level3);
output_charmap (stream, "toupper", to_upper);
output_charmap (stream, "tolower", to_lower);
output_charmap (stream, "map \"totitle\";", to_title);
output_widthmap (stream);
fprintf (stream, "END LC_CTYPE\n");
if (ferror (stream) || fclose (stream))
{
fprintf (stderr, "error writing to '%s'\n", filename);
exit (1);
}
}
inline bool is_cntrl<unicode::char_t> (unicode::char_t c)
{
return c.value <= 127 && is_cntrl(static_cast<char>(c.value));
}
