-/* $Id: term.c,v 1.123 2009/11/06 10:31:32 kristaps Exp $ */
+/* $Id: term.c,v 1.131 2010/04/08 07:05:38 kristaps Exp $ */
/*
* Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
*
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+
+#include <sys/types.h>
+
#include <assert.h>
#include <ctype.h>
#include <stdio.h>
#include "mdoc.h"
#include "main.h"
-/* FIXME: accomodate non-breaking, non-collapsing white-space. */
-/* FIXME: accomodate non-breaking, collapsing white-space. */
-
static struct termp *term_alloc(enum termenc);
static void term_free(struct termp *);
-
-static void do_escaped(struct termp *, const char **);
-static void do_special(struct termp *,
- const char *, size_t);
-static void do_reserved(struct termp *,
- const char *, size_t);
-static void buffer(struct termp *, char);
-static void encode(struct termp *, char);
+static void spec(struct termp *, const char *, size_t);
+static void res(struct termp *, const char *, size_t);
+static void buffera(struct termp *, const char *, size_t);
+static void bufferc(struct termp *, char);
+static void adjbuf(struct termp *p, size_t);
+static void encode(struct termp *, const char *, size_t);
void *
perror(NULL);
exit(EXIT_FAILURE);
}
- p->maxrmargin = 78;
p->enc = enc;
return(p);
}
* Flush a line of text. A "line" is loosely defined as being something
* that should be followed by a newline, regardless of whether it's
* broken apart by newlines getting there. A line can also be a
- * fragment of a columnar list.
+ * fragment of a columnar list (`Bl -tag' or `Bl -column'), which does
+ * not have a trailing newline.
*
- * Specifically, a line is whatever's in p->buf of length p->col, which
- * is zeroed after this function returns.
- *
- * The usage of termp:flags is as follows:
+ * The following flags may be specified:
*
* - TERMP_NOLPAD: when beginning to write the line, don't left-pad the
* offset value. This is useful when doing columnar lists where the
size_t vbl; /* number of blanks to prepend to output */
size_t vsz; /* visual characters to write to output */
size_t bp; /* visual right border position */
+ size_t hyph; /* visible position of hyphen */
int j; /* temporary loop index */
size_t maxvis, mmax;
- static int overstep = 0;
/*
* First, establish the maximum columns of "visible" content.
assert(p->offset < p->rmargin);
- maxvis = (int)(p->rmargin - p->offset) - overstep < 0 ?
+ maxvis = (int)(p->rmargin - p->offset) - p->overstep < 0 ?
/* LINTED */
- 0 : p->rmargin - p->offset - overstep;
- mmax = (int)(p->maxrmargin - p->offset) - overstep < 0 ?
+ 0 : p->rmargin - p->offset - p->overstep;
+ mmax = (int)(p->maxrmargin - p->offset) - p->overstep < 0 ?
/* LINTED */
- 0 : p->maxrmargin - p->offset - overstep;
+ 0 : p->maxrmargin - p->offset - p->overstep;
bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
* (starting with the CSI) aren't counted. A space
* generates a non-printing word, which is valid (the
* space is printed according to regular spacing rules).
+ * Collect the number of printable characters until the
+ * first hyphen, if found. Hyphens aren't included if
+ * they're the first character (so `Fl' doesn't break)
+ * or second consecutive character (`Fl -').
*/
/* LINTED */
- for (j = i, vsz = 0; j < (int)p->col; j++) {
+ for (j = i, vsz = 0, hyph = 0; j < (int)p->col; j++) {
if (j && ' ' == p->buf[j])
break;
- else if (8 == p->buf[j])
+ if (8 == p->buf[j])
vsz--;
else
vsz++;
+ if (j > i && '-' == p->buf[j] && 0 == hyph)
+ if ('-' != p->buf[j - 1])
+ hyph = vsz;
}
/*
* beginning of a line, one between words -- but do not
* actually write them yet.
*/
+
vbl = (size_t)(0 == vis ? 0 : 1);
/*
* Find out whether we would exceed the right margin.
- * If so, break to the next line. (TODO: hyphenate)
- * Otherwise, write the chosen number of blanks now.
+ * If so, break to the next line, possibly after
+ * emittign character up to a hyphen. Otherwise, write
+ * the chosen number of blanks.
*/
+
if (vis && vis + vbl + vsz > bp) {
+ /*
+ * Has a hyphen been found before the breakpoint
+ * that we can use?
+ */
+ if (hyph && vis + vbl + hyph <= bp) {
+ /* First prepend blanks. */
+ for (j = 0; j < (int)vbl; j++)
+ putchar(' ');
+
+ /* Emit up to the character. */
+ do {
+ if (31 == p->buf[i])
+ putchar(' ');
+ else
+ putchar(p->buf[i]);
+ if (8 != p->buf[i])
+ vsz--;
+ } while ('-' != p->buf[i++]);
+
+ /* Emit trailing decoration. */
+ if (8 == p->buf[i]) {
+ putchar(p->buf[i]);
+ putchar(p->buf[i + 1]);
+ }
+ }
+
putchar('\n');
if (TERMP_NOBREAK & p->flags) {
for (j = 0; j < (int)p->rmargin; j++)
putchar(' ');
vis = 0;
}
- /* Remove the overstep width. */
+
+ /* Remove the p->overstep width. */
+
bp += (int)/* LINTED */
- overstep;
- overstep = 0;
+ p->overstep;
+ p->overstep = 0;
} else {
for (j = 0; j < (int)vbl; j++)
putchar(' ');
vis += vbl;
}
- /*
- * Finally, write out the word.
- */
- for ( ; i < (int)p->col; i++) {
+ /* Write out the [remaining] word. */
+ for ( ; i < (int)p->col; i++)
if (' ' == p->buf[i])
break;
-
- /* The unit sep. is a non-breaking space. */
- if (31 == p->buf[i])
+ else if (31 == p->buf[i])
putchar(' ');
else
putchar(p->buf[i]);
- }
+
vis += vsz;
}
p->col = 0;
- overstep = 0;
+ p->overstep = 0;
if ( ! (TERMP_NOBREAK & p->flags)) {
putchar('\n');
if (TERMP_HANG & p->flags) {
/* We need one blank after the tag. */
- overstep = /* LINTED */
+ p->overstep = /* LINTED */
vis - maxvis + 1;
/*
* move it one step LEFT and flag the rest of the line
* to be longer.
*/
- if (overstep >= -1) {
- assert((int)maxvis + overstep >= 0);
+ if (p->overstep >= -1) {
+ assert((int)maxvis + p->overstep >= 0);
/* LINTED */
- maxvis += overstep;
+ maxvis += p->overstep;
} else
- overstep = 0;
+ p->overstep = 0;
} else if (TERMP_DANGLE & p->flags)
return;
static void
-do_special(struct termp *p, const char *word, size_t len)
+spec(struct termp *p, const char *word, size_t len)
{
const char *rhs;
size_t sz;
- int i;
rhs = chars_a2ascii(p->symtab, word, len, &sz);
-
- if (NULL == rhs) {
-#if 0
- fputs("Unknown special character: ", stderr);
- for (i = 0; i < (int)len; i++)
- fputc(word[i], stderr);
- fputc('\n', stderr);
-#endif
- return;
- }
- for (i = 0; i < (int)sz; i++)
- encode(p, rhs[i]);
+ if (rhs)
+ encode(p, rhs, sz);
}
static void
-do_reserved(struct termp *p, const char *word, size_t len)
+res(struct termp *p, const char *word, size_t len)
{
const char *rhs;
size_t sz;
- int i;
rhs = chars_a2res(p->symtab, word, len, &sz);
-
- if (NULL == rhs) {
-#if 0
- fputs("Unknown reserved word: ", stderr);
- for (i = 0; i < (int)len; i++)
- fputc(word[i], stderr);
- fputc('\n', stderr);
-#endif
- return;
- }
- for (i = 0; i < (int)sz; i++)
- encode(p, rhs[i]);
+ if (rhs)
+ encode(p, rhs, sz);
}
-/*
- * Handle an escape sequence: determine its length and pass it to the
- * escape-symbol look table. Note that we assume mdoc(3) has validated
- * the escape sequence (we assert upon badly-formed escape sequences).
- */
-static void
-do_escaped(struct termp *p, const char **word)
+void
+term_fontlast(struct termp *p)
{
- int j, type, sv, t, lim;
- const char *wp;
+ enum termfont f;
- wp = *word;
- type = 1;
-
- if ('\0' == *(++wp)) {
- *word = wp;
- return;
- }
+ f = p->fontl;
+ p->fontl = p->fontq[p->fonti];
+ p->fontq[p->fonti] = f;
+}
- if ('(' == *wp) {
- wp++;
- if ('\0' == *wp || '\0' == *(wp + 1)) {
- *word = '\0' == *wp ? wp : wp + 1;
- return;
- }
- do_special(p, wp, 2);
- *word = ++wp;
- return;
-
- } else if ('*' == *wp) {
- if ('\0' == *(++wp)) {
- *word = wp;
- return;
- }
+void
+term_fontrepl(struct termp *p, enum termfont f)
+{
- switch (*wp) {
- case ('('):
- wp++;
- if ('\0' == *wp || '\0' == *(wp + 1)) {
- *word = '\0' == *wp ? wp : wp + 1;
- return;
- }
+ p->fontl = p->fontq[p->fonti];
+ p->fontq[p->fonti] = f;
+}
- do_reserved(p, wp, 2);
- *word = ++wp;
- return;
- case ('['):
- type = 0;
- break;
- default:
- do_reserved(p, wp, 1);
- *word = wp;
- return;
- }
- } else if ('s' == *wp) {
- /* This closely follows mandoc_special(). */
- if ('\0' == *(++wp)) {
- *word = wp;
- return;
- }
+void
+term_fontpush(struct termp *p, enum termfont f)
+{
- t = 0;
- lim = 1;
-
- if (*wp == '\'') {
- lim = 0;
- t = 1;
- ++wp;
- } else if (*wp == '[') {
- lim = 0;
- t = 2;
- ++wp;
- } else if (*wp == '(') {
- lim = 2;
- t = 3;
- ++wp;
- }
+ assert(p->fonti + 1 < 10);
+ p->fontl = p->fontq[p->fonti];
+ p->fontq[++p->fonti] = f;
+}
- if (*wp == '+' || *wp == '-')
- ++wp;
- if (*wp == '\'') {
- if (t) {
- *word = wp;
- return;
- }
- lim = 0;
- t = 1;
- ++wp;
- } else if (*wp == '[') {
- if (t) {
- *word = wp;
- return;
- }
- lim = 0;
- t = 2;
- ++wp;
- } else if (*wp == '(') {
- if (t) {
- *word = wp;
- return;
- }
- lim = 2;
- t = 3;
- ++wp;
- }
+const void *
+term_fontq(struct termp *p)
+{
- if ( ! isdigit((u_char)*wp)) {
- *word = --wp;
- return;
- }
+ return(&p->fontq[p->fonti]);
+}
- for (j = 0; isdigit((u_char)*wp); j++) {
- if (lim && j >= lim)
- break;
- ++wp;
- }
- if (t && t < 3) {
- if (1 == t && *wp != '\'') {
- *word = --wp;
- return;
- }
- if (2 == t && *wp != ']') {
- *word = --wp;
- return;
- }
- ++wp;
- }
- *word = --wp;
- return;
+enum termfont
+term_fonttop(struct termp *p)
+{
- } else if ('f' == *wp) {
- if ('\0' == *(++wp)) {
- *word = wp;
- return;
- }
+ return(p->fontq[p->fonti]);
+}
- switch (*wp) {
- case ('3'):
- /* FALLTHROUGH */
- case ('B'):
- p->metamask = p->metafont;
- p->metafont |= METAF_BOLD;
- break;
- case ('2'):
- /* FALLTHROUGH */
- case ('I'):
- p->metamask = p->metafont;
- p->metafont |= METAF_UNDER;
- break;
- case ('P'):
- sv = p->metamask;
- p->metamask = p->metafont;
- p->metafont = sv;
- break;
- case ('1'):
- /* FALLTHROUGH */
- case ('R'):
- p->metamask = p->metafont;
- p->metafont &= ~METAF_UNDER;
- p->metafont &= ~METAF_BOLD;
- break;
- default:
- break;
- }
- *word = wp;
- return;
+void
+term_fontpopq(struct termp *p, const void *key)
+{
- } else if ('[' != *wp) {
- do_special(p, wp, 1);
- *word = wp;
- return;
- }
+ while (p->fonti >= 0 && key != &p->fontq[p->fonti])
+ p->fonti--;
+ assert(p->fonti >= 0);
+}
- wp++;
- for (j = 0; *wp && ']' != *wp; wp++, j++)
- /* Loop... */ ;
- if ('\0' == *wp) {
- *word = wp;
- return;
- }
+void
+term_fontpop(struct termp *p)
+{
- if (type)
- do_special(p, wp - j, (size_t)j);
- else
- do_reserved(p, wp - j, (size_t)j);
- *word = wp;
+ assert(p->fonti);
+ p->fonti--;
}
void
term_word(struct termp *p, const char *word)
{
- const char *sv;
+ const char *sv, *seq;
+ int sz;
+ size_t ssz;
+ enum roffdeco deco;
sv = word;
case(')'):
/* FALLTHROUGH */
case(']'):
- /* FALLTHROUGH */
- case('}'):
if ( ! (TERMP_IGNDELIM & p->flags))
p->flags |= TERMP_NOSPACE;
break;
}
if ( ! (TERMP_NOSPACE & p->flags))
- buffer(p, ' ');
+ bufferc(p, ' ');
if ( ! (p->flags & TERMP_NONOSPACE))
p->flags &= ~TERMP_NOSPACE;
- for ( ; *word; word++)
- if ('\\' != *word)
- encode(p, *word);
- else
- do_escaped(p, &word);
+ /* FIXME: use strcspn. */
+ while (*word) {
+ if ('\\' != *word) {
+ encode(p, word, 1);
+ word++;
+ continue;
+ }
+
+ seq = ++word;
+ sz = a2roffdeco(&deco, &seq, &ssz);
+
+ switch (deco) {
+ case (DECO_RESERVED):
+ res(p, seq, ssz);
+ break;
+ case (DECO_SPECIAL):
+ spec(p, seq, ssz);
+ break;
+ case (DECO_BOLD):
+ term_fontrepl(p, TERMFONT_BOLD);
+ break;
+ case (DECO_ITALIC):
+ term_fontrepl(p, TERMFONT_UNDER);
+ break;
+ case (DECO_ROMAN):
+ term_fontrepl(p, TERMFONT_NONE);
+ break;
+ case (DECO_PREVIOUS):
+ term_fontlast(p);
+ break;
+ default:
+ break;
+ }
+
+ word += sz;
+ if (DECO_NOSPACE == deco && '\0' == *word)
+ p->flags |= TERMP_NOSPACE;
+ }
+
+ /*
+ * Note that we don't process the pipe: the parser sees it as
+ * punctuation, but we don't in terms of typography.
+ */
if (sv[0] && 0 == sv[1])
switch (sv[0]) {
case('('):
/* FALLTHROUGH */
case('['):
- /* FALLTHROUGH */
- case('{'):
p->flags |= TERMP_NOSPACE;
break;
default:
}
-/*
- * Insert a single character into the line-buffer. If the buffer's
- * space is exceeded, then allocate more space by doubling the buffer
- * size.
- */
static void
-buffer(struct termp *p, char c)
+adjbuf(struct termp *p, size_t sz)
{
- size_t s;
-
- if (p->col + 1 >= p->maxcols) {
- if (0 == p->maxcols)
- p->maxcols = 256;
- s = p->maxcols * 2;
- p->buf = realloc(p->buf, s);
- if (NULL == p->buf) {
- perror(NULL);
- exit(EXIT_FAILURE);
- }
- p->maxcols = s;
+
+ if (0 == p->maxcols)
+ p->maxcols = 1024;
+ while (sz >= p->maxcols)
+ p->maxcols <<= 2;
+
+ p->buf = realloc(p->buf, p->maxcols);
+ if (NULL == p->buf) {
+ perror(NULL);
+ exit(EXIT_FAILURE);
}
- p->buf[(int)(p->col)++] = c;
}
static void
-encode(struct termp *p, char c)
+buffera(struct termp *p, const char *word, size_t sz)
{
-
- if (isgraph((u_char)c)) {
- if (p->under || METAF_UNDER & p->metafont) {
- buffer(p, '_');
- buffer(p, 8);
- }
- if (p->bold || METAF_BOLD & p->metafont) {
- buffer(p, c);
- buffer(p, 8);
+
+ if (p->col + sz >= p->maxcols)
+ adjbuf(p, p->col + sz);
+
+ memcpy(&p->buf[(int)p->col], word, sz);
+ p->col += sz;
+}
+
+
+static void
+bufferc(struct termp *p, char c)
+{
+
+ if (p->col + 1 >= p->maxcols)
+ adjbuf(p, p->col + 1);
+
+ p->buf[(int)p->col++] = c;
+}
+
+
+static void
+encode(struct termp *p, const char *word, size_t sz)
+{
+ enum termfont f;
+ int i;
+
+ /*
+ * Encode and buffer a string of characters. If the current
+ * font mode is unset, buffer directly, else encode then buffer
+ * character by character.
+ */
+
+ if (TERMFONT_NONE == (f = term_fonttop(p))) {
+ buffera(p, word, sz);
+ return;
+ }
+
+ for (i = 0; i < (int)sz; i++) {
+ if ( ! isgraph((u_char)word[i])) {
+ bufferc(p, word[i]);
+ continue;
}
+
+ if (TERMFONT_UNDER == f)
+ bufferc(p, '_');
+ else
+ bufferc(p, word[i]);
+
+ bufferc(p, 8);
+ bufferc(p, word[i]);
}
- buffer(p, c);
}
git.ckatri.com / mandoc.git / blobdiff