int
main (int argc, char * argv[])
{
if (argc == 1)
{
/* Display all the break opportunities in the input string. */
char *input = read_file (stdin);
int length = strlen (input);
char *breaks = malloc (length);
int i;
u8_possible_linebreaks ((uint8_t *) input, length, "UTF-8", breaks);
for (i = 0; i < length; i++)
{
switch (breaks[i])
{
case UC_BREAK_POSSIBLE:
/* U+2027 in UTF-8 encoding */
putc (0xe2, stdout); putc (0x80, stdout); putc (0xa7, stdout);
break;
case UC_BREAK_MANDATORY:
/* U+21B2 (or U+21B5) in UTF-8 encoding */
putc (0xe2, stdout); putc (0x86, stdout); putc (0xb2, stdout);
break;
case UC_BREAK_PROHIBITED:
break;
default:
abort ();
}
putc (input[i], stdout);
}
free (breaks);
return 0;
}
else
return 1;
}
void
ulc_possible_linebreaks (const char *s, size_t n, const char *encoding,
char *p)
{
if (n > 0)
{
if (is_utf8_encoding (encoding))
u8_possible_linebreaks ((const uint8_t *) s, n, encoding, p);
else
{
/* Convert the string to UTF-8 and build a translation table
from offsets into s to offsets into the translated string. */
size_t *offsets = (size_t *) malloc (n * sizeof (size_t));
if (offsets != NULL)
{
uint8_t *t;
size_t m;
t = u8_conv_from_encoding (encoding, iconveh_question_mark,
s, n, offsets, NULL, &m);
if (t != NULL)
{
char *q = (char *) (m > 0 ? malloc (m) : NULL);
if (m == 0 || q != NULL)
{
size_t i;
/* Determine the possible line breaks of the UTF-8
string. */
u8_possible_linebreaks (t, m, encoding, q);
/* Translate the result back to the original string. */
memset (p, UC_BREAK_PROHIBITED, n);
for (i = 0; i < n; i++)
if (offsets[i] != (size_t)(-1))
p[i] = q[offsets[i]];
free (q);
free (t);
free (offsets);
return;
}
free (t);
}
free (offsets);
}
/* Impossible to convert. */
#if C_CTYPE_ASCII
if (is_all_ascii (s, n))
{
/* ASCII is a subset of UTF-8. */
u8_possible_linebreaks ((const uint8_t *) s, n, encoding, p);
return;
}
#endif
/* We have a non-ASCII string and cannot convert it.
Don't produce line breaks except those already present in the
input string. All we assume here is that the encoding is
minimally ASCII compatible. */
{
const char *s_end = s + n;
while (s < s_end)
{
*p = (*s == '\n' ? UC_BREAK_MANDATORY : UC_BREAK_PROHIBITED);
s++;
p++;
}
}
}
}
}
int
u8_width_linebreaks (const uint8_t *s, size_t n,
int width, int start_column, int at_end_columns,
const char *o, const char *encoding,
char *p)
{
const uint8_t *s_end;
char *last_p;
int last_column;
int piece_width;
u8_possible_linebreaks (s, n, encoding, p);
s_end = s + n;
last_p = NULL;
last_column = start_column;
piece_width = 0;
while (s < s_end)
{
ucs4_t uc;
int count = u8_mbtouc_unsafe (&uc, s, s_end - s);
/* Respect the override. */
if (o != NULL && *o != UC_BREAK_UNDEFINED)
*p = *o;
if (*p == UC_BREAK_POSSIBLE || *p == UC_BREAK_MANDATORY)
{
/* An atomic piece of text ends here. */
if (last_p != NULL && last_column + piece_width > width)
{
/* Insert a line break. */
*last_p = UC_BREAK_POSSIBLE;
last_column = 0;
}
}
if (*p == UC_BREAK_MANDATORY)
{
/* uc is a line break character. */
/* Start a new piece at column 0. */
last_p = NULL;
last_column = 0;
piece_width = 0;
}
else
{
/* uc is not a line break character. */
int w;
if (*p == UC_BREAK_POSSIBLE)
{
/* Start a new piece. */
last_p = p;
last_column += piece_width;
piece_width = 0;
/* No line break for the moment, may be turned into
UC_BREAK_POSSIBLE later, via last_p. */
}
*p = UC_BREAK_PROHIBITED;
w = uc_width (uc, encoding);
if (w >= 0) /* ignore control characters in the string */
piece_width += w;
}
s += count;
p += count;
if (o != NULL)
o += count;
}
/* The last atomic piece of text ends here. */
if (last_p != NULL && last_column + piece_width + at_end_columns > width)
{
/* Insert a line break. */
*last_p = UC_BREAK_POSSIBLE;
last_column = 0;
}
return last_column + piece_width;
}