Added default print of `~' with empty `Pa' (not documented with OpenBSD, but still...

[mandoc.git] / term.c

diff --git a/term.c b/term.c
index 935b29e384217f629844a7e62795c508c24b51d4..f7b926976a8e6211f833d33d893dd52e4fe89f68 100644 (file)
--- a/term.c
+++ b/term.c
@@ -1,4 +1,4 @@
-/*     $Id: term.c,v 1.79 2009/06/17 18:42:42 kristaps Exp $ */
+/*     $Id: term.c,v 1.91 2009/07/21 13:34:13 kristaps Exp $ */

 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
  *
 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
  *


@@ -31,15 +31,13 @@ extern      int               mdoc_run(struct termp *,
 
 static struct termp     *term_alloc(enum termenc);
 static void              term_free(struct termp *);
 
 static struct termp     *term_alloc(enum termenc);
 static void              term_free(struct termp *);

-static void              term_pword(struct termp *, const char *, int);
-static void              term_pescape(struct termp *, 
-                               const char *, int *, int);
+static void              term_pescape(struct termp *, const char **);

 static void              term_nescape(struct termp *,
                                const char *, size_t);
 static void              term_chara(struct termp *, char);
 static void              term_encodea(struct termp *, char);
 static void              term_nescape(struct termp *,
                                const char *, size_t);
 static void              term_chara(struct termp *, char);
 static void              term_encodea(struct termp *, char);

-static int               term_isopendelim(const char *, int);
-static int               term_isclosedelim(const char *, int);
+static int               term_isopendelim(const char *);
+static int               term_isclosedelim(const char *);

 
 
 void *
 
 
 void *


@@ -105,17 +103,17 @@ term_alloc(enum termenc enc)
        if (NULL == (p = malloc(sizeof(struct termp))))
                err(1, "malloc");
        bzero(p, sizeof(struct termp));
        if (NULL == (p = malloc(sizeof(struct termp))))
                err(1, "malloc");
        bzero(p, sizeof(struct termp));

-       p->maxrmargin = 80;
+       p->maxrmargin = 78;

        p->enc = enc;
        return(p);
 }
 
 
 static int
        p->enc = enc;
        return(p);
 }
 
 
 static int

-term_isclosedelim(const char *p, int len)
+term_isclosedelim(const char *p)

 {
 
 {
 

-       if (1 != len)
+       if ( ! (*p && 0 == *(p + 1)))

                return(0);
 
        switch (*p) {
                return(0);
 
        switch (*p) {


@@ -146,10 +144,10 @@ term_isclosedelim(const char *p, int len)
 
 
 static int
 
 
 static int

-term_isopendelim(const char *p, int len)
+term_isopendelim(const char *p)

 {
 
 {
 

-       if (1 != len)
+       if ( ! (*p && 0 == *(p + 1)))

                return(0);
 
        switch (*p) {
                return(0);
 
        switch (*p) {


@@ -176,8 +174,7 @@ term_isopendelim(const char *p, int len)
  * Specifically, a line is whatever's in p->buf of length p->col, which
  * is zeroed after this function returns.
  *
  * Specifically, a line is whatever's in p->buf of length p->col, which
  * is zeroed after this function returns.
  *

- * The variables TERMP_NOLPAD, TERMP_LITERAL and TERMP_NOBREAK are of
- * critical importance here.  Their behaviour follows:
+ * The usage of termp:flags is as follows:

  *
  *  - TERMP_NOLPAD: when beginning to write the line, don't left-pad the
  *    offset value.  This is useful when doing columnar lists where the
  *
  *  - TERMP_NOLPAD: when beginning to write the line, don't left-pad the
  *    offset value.  This is useful when doing columnar lists where the


@@ -187,7 +184,15 @@ term_isopendelim(const char *p, int len)
  *    columns.  In short: don't print a newline and instead pad to the
  *    right margin.  Used in conjunction with TERMP_NOLPAD.
  *
  *    columns.  In short: don't print a newline and instead pad to the
  *    right margin.  Used in conjunction with TERMP_NOLPAD.
  *

- *  - TERMP_NONOBREAK: don't newline when TERMP_NOBREAK is specified.
+ *  - TERMP_TWOSPACE: when padding, make sure there are at least two
+ *    space characters of padding.  Otherwise, rather break the line.
+ *
+ *  - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and
+ *    the line is overrun, and don't pad-right if it's underrun.
+ *
+ *  - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when
+ *    overruning, instead save the position and continue at that point
+ *    when the next invocation.

  *
  *  In-line line breaking:
  *
  *
  *  In-line line breaking:
  *


@@ -200,12 +205,19 @@ term_isopendelim(const char *p, int len)
  *  Otherwise, the line will break at the right margin.  Extremely long
  *  lines will cause the system to emit a warning (TODO: hyphenate, if
  *  possible).
  *  Otherwise, the line will break at the right margin.  Extremely long
  *  lines will cause the system to emit a warning (TODO: hyphenate, if
  *  possible).

+ *
+ *  FIXME: newline breaks occur (in groff) also occur when a single
+ *  space follows a NOBREAK (try `Bl -tag')
+ *
+ *  FIXME: there's a newline error where a `Bl -diag' will have a
+ *  trailing newline if the line is exactly 73 chars long.

  */
 void
 term_flushln(struct termp *p)
 {
        int              i, j;
  */
 void
 term_flushln(struct termp *p)
 {
        int              i, j;

-       size_t           vsz, vis, maxvis, mmax, bp;
+       size_t           vbl, vsz, vis, maxvis, mmax, bp;
+       static int       overstep = 0;

 
        /*
         * First, establish the maximum columns of "visible" content.
 
        /*
         * First, establish the maximum columns of "visible" content.


@@ -215,10 +227,11 @@ term_flushln(struct termp *p)
         */
 
        assert(p->offset < p->rmargin);
         */
 
        assert(p->offset < p->rmargin);

-       maxvis = p->rmargin - p->offset;
-       mmax = p->maxrmargin - p->offset;
+       maxvis = p->rmargin - p->offset - overstep;
+       mmax = p->maxrmargin - p->offset - overstep;

        bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
        vis = 0;
        bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
        vis = 0;

+       overstep = 0;

 
        /*
         * If in the standard case (left-justified), then begin with our
 
        /*
         * If in the standard case (left-justified), then begin with our


@@ -241,44 +254,46 @@ term_flushln(struct termp *p)
 
                /* LINTED */
                for (j = i, vsz = 0; j < (int)p->col; j++) {
 
                /* LINTED */
                for (j = i, vsz = 0; j < (int)p->col; j++) {

-                       if (' ' == p->buf[j])
+                       if (' ' == p->buf[j]) 

                                break;
                        else if (8 == p->buf[j])
                                break;
                        else if (8 == p->buf[j])

-                               j += 1;
+                               vsz--;

                        else
                                vsz++;
                }
 
                /*
                        else
                                vsz++;
                }
 
                /*

-                * Do line-breaking.  If we're greater than our
-                * break-point and already in-line, break to the next
-                * line and start writing.  If we're at the line start,
-                * then write out the word (TODO: hyphenate) and break
-                * in a subsequent loop invocation.
+                * Choose the number of blanks to prepend: no blank at the
+                * beginning of a line, one between words -- but do not
+                * actually write them yet.

                 */
                 */

+               vbl = (size_t)(0 == vis ? 0 : 1);

 
 

-               if ( ! (TERMP_NOBREAK & p->flags)) {
-                       if (vis && vis + vsz > bp) {
-                               putchar('\n');
+               /*
+                * 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 (vis && vis + vbl + vsz > bp) {
+                       putchar('\n');
+                       if (TERMP_NOBREAK & p->flags) {
+                               for (j = 0; j < (int)p->rmargin; j++)
+                                       putchar(' ');
+                               vis = p->rmargin - p->offset;
+                       } else {

                                for (j = 0; j < (int)p->offset; j++)
                                        putchar(' ');
                                vis = 0;
                                for (j = 0; j < (int)p->offset; j++)
                                        putchar(' ');
                                vis = 0;

-                       } 
-               } else if (vis && vis + vsz > bp) {
-                       putchar('\n');
-                       for (j = 0; j < (int)p->rmargin; j++)
+                       }
+               } else {
+                       for (j = 0; j < (int)vbl; j++)

                                putchar(' ');
                                putchar(' ');

-                       vis = p->rmargin - p->offset;
+                       vis += vbl;

                }
 
                /*
                }
 
                /*

-                * Prepend a space if we're not already at the beginning
-                * of the line, then the word.
+                * Finally, write out the word.

                 */
                 */

-
-               if (0 < vis++)
-                       putchar(' ');
-

                for ( ; i < (int)p->col; i++) {
                        if (' ' == p->buf[i])
                                break;
                for ( ; i < (int)p->col; i++) {
                        if (' ' == p->buf[i])
                                break;


@@ -286,35 +301,42 @@ term_flushln(struct termp *p)
                }
                vis += vsz;
        }
                }
                vis += vsz;
        }

+       p->col = 0;

 
 

-       /*
-        * If we've overstepped our maximum visible no-break space, then
-        * cause a newline and offset at the right margin.
-        */
-
-       if ((TERMP_NOBREAK & p->flags) && vis > maxvis) {
-               if ( ! (TERMP_NONOBREAK & p->flags)) {
-                       putchar('\n');
-                       for (i = 0; i < (int)p->rmargin; i++)
-                               putchar(' ');
-               }
-               p->col = 0;
+       if ( ! (TERMP_NOBREAK & p->flags)) {
+               putchar('\n');

                return;
        }
 
                return;
        }
 

-       /*
-        * If we're not to right-marginalise it (newline), then instead
-        * pad to the right margin and stay off.
-        */
+       if (TERMP_HANG & p->flags) {

 
 

-       if (p->flags & TERMP_NOBREAK) {
-               if ( ! (TERMP_NONOBREAK & p->flags))
-                       for ( ; vis <= maxvis; vis++)
-                               putchar(' ');
-       } else
-               putchar('\n');
+               /* We need one blank after the tag. */
+               overstep = vis - maxvis + 1;

 
 

-       p->col = 0;
+               /*
+                * Behave exactly the same way as groff:
+                * If we have overstepped the margin, temporarily move it
+                * to the right and flag the rest of the line to be shorter.
+                * If we landed right at the margin, be happy.
+                * If we are one step before the margin, temporarily move it
+                * one step LEFT and flag the rest of the line to be longer.
+                */
+               if (overstep >= -1)
+                       maxvis += overstep;
+               else
+                       overstep = 0;
+
+       } else if (TERMP_DANGLE & p->flags)
+               return;
+
+       if (maxvis > vis + ((TERMP_TWOSPACE & p->flags) ? 1 : 0))  /* pad */
+               for ( ; vis < maxvis; vis++)
+                       putchar(' ');
+       else {  /* break */
+               putchar('\n');
+               for (i = 0; i < (int)p->rmargin; i++)
+                       putchar(' ');
+       }

 }
 
 
 }
 
 


@@ -352,50 +374,6 @@ term_vspace(struct termp *p)
 }
 
 
 }
 
 

-/*
- * Break apart a word into "pwords" (partial-words, usually from
- * breaking up a phrase into individual words) and, eventually, put them
- * into the output buffer.  If we're a literal word, then don't break up
- * the word and put it verbatim into the output buffer.
- */
-void
-term_word(struct termp *p, const char *word)
-{
-       int              i, j, len;
-
-       len = (int)strlen(word);
-
-       if (p->flags & TERMP_LITERAL) {
-               term_pword(p, word, len);
-               return;
-       }
-
-       /* LINTED */
-       for (j = i = 0; i < len; i++) {
-               if (' ' != word[i]) {
-                       j++;
-                       continue;
-               } 
-               
-               /* Escaped spaces don't delimit... */
-               if (i && ' ' == word[i] && '\\' == word[i - 1]) {
-                       j++;
-                       continue;
-               }
-
-               if (0 == j)
-                       continue;
-               assert(i >= j);
-               term_pword(p, &word[i - j], j);
-               j = 0;
-       }
-       if (j > 0) {
-               assert(i >= j);
-               term_pword(p, &word[i - j], j);
-       }
-}
-
-

 /*
  * Determine the symbol indicated by an escape sequences, that is, one
  * starting with a backslash.  Once done, we pass this value into the
 /*
  * Determine the symbol indicated by an escape sequences, that is, one
  * starting with a backslash.  Once done, we pass this value into the


@@ -408,7 +386,9 @@ term_nescape(struct termp *p, const char *word, size_t len)
        size_t           sz;
        int              i;
 
        size_t           sz;
        int              i;
 

-       if ((rhs = term_a2ascii(p->symtab, word, len, &sz))) 
+       rhs = term_a2ascii(p->symtab, word, len, &sz);
+
+       if (rhs)

                for (i = 0; i < (int)sz; i++) 
                        term_encodea(p, rhs[i]);
 }
                for (i = 0; i < (int)sz; i++) 
                        term_encodea(p, rhs[i]);
 }


@@ -420,48 +400,61 @@ term_nescape(struct termp *p, const char *word, size_t len)
  * the escape sequence (we assert upon badly-formed escape sequences).
  */
 static void
  * the escape sequence (we assert upon badly-formed escape sequences).
  */
 static void

-term_pescape(struct termp *p, const char *word, int *i, int len)
+term_pescape(struct termp *p, const char **word)

 {
        int              j;
 {
        int              j;

+       const char      *wp;

 
 

-       if (++(*i) >= len)
+       wp = *word;
+
+       if (0 == *(++wp)) {
+               *word = wp;

                return;
                return;

+       }

 
 

-       if ('(' == word[*i]) {
-               (*i)++;
-               if (*i + 1 >= len)
+       if ('(' == *wp) {
+               wp++;
+               if (0 == *wp || 0 == *(wp + 1)) {
+                       *word = 0 == *wp ? wp : wp + 1;

                        return;
                        return;

+               }

 
 

-               term_nescape(p, &word[*i], 2);
-               (*i)++;
+               term_nescape(p, wp, 2);
+               *word = ++wp;

                return;
 
                return;
 

-       } else if ('*' == word[*i]) { 
-               (*i)++;
-               if (*i >= len)
+       } else if ('*' == *wp) {
+               if (0 == *(++wp)) {
+                       *word = wp;

                        return;
                        return;

+               }

 
 

-               switch (word[*i]) {
+               switch (*wp) {

                case ('('):
                case ('('):

-                       (*i)++;
-                       if (*i + 1 >= len)
+                       wp++;
+                       if (0 == *wp || 0 == *(wp + 1)) {
+                               *word = 0 == *wp ? wp : wp + 1;

                                return;
                                return;

+                       }

 
 

-                       term_nescape(p, &word[*i], 2);
-                       (*i)++;
+                       term_nescape(p, wp, 2);
+                       *word = ++wp;

                        return;
                case ('['):
                        break;
                default:
                        return;
                case ('['):
                        break;
                default:

-                       term_nescape(p, &word[*i], 1);
+                       term_nescape(p, wp, 1);
+                       *word = wp;

                        return;
                }
        
                        return;
                }
        

-       } else if ('f' == word[*i]) {
-               (*i)++;
-               if (*i >= len)
+       } else if ('f' == *wp) {
+               if (0 == *(++wp)) {
+                       *word = wp;

                        return;
                        return;

-               switch (word[*i]) {
+               }
+
+               switch (*wp) {

                case ('B'):
                        p->flags |= TERMP_BOLD;
                        break;
                case ('B'):
                        p->flags |= TERMP_BOLD;
                        break;


@@ -476,21 +469,27 @@ term_pescape(struct termp *p, const char *word, int *i, int len)
                default:
                        break;
                }
                default:
                        break;
                }

+
+               *word = wp;

                return;
 
                return;
 

-       } else if ('[' != word[*i]) {
-               term_nescape(p, &word[*i], 1);
+       } else if ('[' != *wp) {
+               term_nescape(p, wp, 1);
+               *word = wp;

                return;
        }
 
                return;
        }
 

-       (*i)++;
-       for (j = 0; word[*i] && ']' != word[*i]; (*i)++, j++)
+       wp++;
+       for (j = 0; *wp && ']' != *wp; wp++, j++)

                /* Loop... */ ;
 
                /* Loop... */ ;
 

-       if (0 == word[*i])
+       if (0 == *wp) {
+               *word = wp;

                return;
                return;

+       }

 
 

-       term_nescape(p, &word[*i - j], (size_t)j);
+       term_nescape(p, wp - j, (size_t)j);
+       *word = wp;

 }
 
 
 }
 
 


@@ -499,12 +498,12 @@ term_pescape(struct termp *p, const char *word, int *i, int len)
  * phrase that cannot be broken down (such as a literal string).  This
  * handles word styling.
  */
  * phrase that cannot be broken down (such as a literal string).  This
  * handles word styling.
  */

-static void
-term_pword(struct termp *p, const char *word, int len)
+void
+term_word(struct termp *p, const char *word)

 {
 {

-       int              i;
+       const char       *sv;

 
 

-       if (term_isclosedelim(word, len))
+       if (term_isclosedelim(word))

                if ( ! (TERMP_IGNDELIM & p->flags))
                        p->flags |= TERMP_NOSPACE;
 
                if ( ! (TERMP_IGNDELIM & p->flags))
                        p->flags |= TERMP_NOSPACE;
 


@@ -519,13 +518,13 @@ term_pword(struct termp *p, const char *word, int len)
         * before the word.
         */
 
         * before the word.
         */
 

-       for (i = 0; i < len; i++) 
-               if ('\\' == word[i]) 
-                       term_pescape(p, word, &i, len);
+       for (sv = word; *word; word++)
+               if ('\\' != *word)
+                       term_encodea(p, *word);

                else
                else

-                       term_encodea(p, word[i]);
+                       term_pescape(p, &word);

 
 

-       if (term_isopendelim(word, len))
+       if (term_isopendelim(sv))

                p->flags |= TERMP_NOSPACE;
 }
 
                p->flags |= TERMP_NOSPACE;
 }
 


@@ -556,8 +555,8 @@ term_chara(struct termp *p, char c)
 static void
 term_encodea(struct termp *p, char c)
 {
 static void
 term_encodea(struct termp *p, char c)
 {

-
-       if (TERMP_STYLE & p->flags) {
+       
+       if (' ' != c && TERMP_STYLE & p->flags) {

                if (TERMP_BOLD & p->flags) {
                        term_chara(p, c);
                        term_chara(p, 8);
                if (TERMP_BOLD & p->flags) {
                        term_chara(p, c);
                        term_chara(p, 8);