===================================================================
RCS file: /cvs/mandoc/Attic/terminal.c,v
retrieving revision 1.1
retrieving revision 1.10
diff -u -p -r1.1 -r1.10
--- mandoc/Attic/terminal.c	2009/03/19 16:17:27	1.1
+++ mandoc/Attic/terminal.c	2009/03/26 14:44:41	1.10
@@ -1,4 +1,4 @@
-/* $Id: terminal.c,v 1.1 2009/03/19 16:17:27 kristaps Exp $ */
+/* $Id: terminal.c,v 1.10 2009/03/26 14:44:41 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@openbsd.org>
  *
@@ -23,64 +23,51 @@
 #include <string.h>
 
 #include "term.h"
+#include "man.h"
+#include "mdoc.h"
 
-static	struct termp	 *termp_alloc(enum termenc);
-static	void		  termp_free(struct termp *);
-static	void		  termp_body(struct termp *, struct termpair *,
-				const struct mdoc_meta *,
-				const struct mdoc_node *);
-static	void		  termp_head(struct termp *,
-				const struct mdoc_meta *);
-static	void		  termp_foot(struct termp *,
-				const struct mdoc_meta *);
-static	void		  termp_pword(struct termp *, const char *, int);
-static	void		  termp_pescape(struct termp *, 
+extern	int		  man_run(struct termp *, 
+				const struct man *);
+extern	int		  mdoc_run(struct termp *, 
+				const struct mdoc *);
+
+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		  termp_nescape(struct termp *,
+static	void		  term_nescape(struct termp *,
 				const char *, size_t);
-static	void		  termp_chara(struct termp *, char);
-static	void		  termp_stringa(struct termp *, 
+static	void		  term_chara(struct termp *, char);
+static	void		  term_stringa(struct termp *, 
 				const char *, size_t);
-static	void		  sanity(const struct mdoc_node *); /* XXX */
+static	int		  term_isopendelim(const char *, int);
+static	int		  term_isclosedelim(const char *, int);
 
 
 void *
-latin1_alloc(void)
-{
-
-	return(termp_alloc(TERMENC_LATIN1));
-}
-
-
-void *
-utf8_alloc(void)
-{
-
-	return(termp_alloc(TERMENC_UTF8));
-}
-
-
-void *
 ascii_alloc(void)
 {
 
-	return(termp_alloc(TERMENC_ASCII));
+	return(term_alloc(TERMENC_ASCII));
 }
 
 
 int
-terminal_run(void *arg, const struct mdoc *mdoc)
+terminal_run(void *arg, const struct man *man,
+		const struct mdoc *mdoc)
 {
 	struct termp	*p;
 
 	p = (struct termp *)arg;
 
 	if (NULL == p->symtab)
-		p->symtab = ascii2htab();
+		p->symtab = term_ascii2htab();
 
-	termp_head(p, mdoc_meta(mdoc));
-	termp_body(p, NULL, mdoc_meta(mdoc), mdoc_node(mdoc));
-	termp_foot(p, mdoc_meta(mdoc));
+	if (man)
+		return(man_run(p, man));
+	if (mdoc)
+		return(mdoc_run(p, mdoc));
 
 	return(1);
 }
@@ -90,25 +77,25 @@ void
 terminal_free(void *arg)
 {
 
-	termp_free((struct termp *)arg);
+	term_free((struct termp *)arg);
 }
 
 
 static void
-termp_free(struct termp *p)
+term_free(struct termp *p)
 {
 
 	if (p->buf)
 		free(p->buf);
 	if (TERMENC_ASCII == p->enc && p->symtab)
-		asciifree(p->symtab);
+		term_asciifree(p->symtab);
 
 	free(p);
 }
 
 
 static struct termp *
-termp_alloc(enum termenc enc)
+term_alloc(enum termenc enc)
 {
 	struct termp *p;
 
@@ -121,6 +108,62 @@ termp_alloc(enum termenc enc)
 }
 
 
+static int
+term_isclosedelim(const char *p, int len)
+{
+
+	if (1 != len)
+		return(0);
+
+	switch (*p) {
+	case('.'):
+		/* FALLTHROUGH */
+	case(','):
+		/* FALLTHROUGH */
+	case(';'):
+		/* FALLTHROUGH */
+	case(':'):
+		/* FALLTHROUGH */
+	case('?'):
+		/* FALLTHROUGH */
+	case('!'):
+		/* FALLTHROUGH */
+	case(')'):
+		/* FALLTHROUGH */
+	case(']'):
+		/* FALLTHROUGH */
+	case('}'):
+		return(1);
+	default:
+		break;
+	}
+
+	return(0);
+}
+
+
+static int
+term_isopendelim(const char *p, int len)
+{
+
+	if (1 != len)
+		return(0);
+
+	switch (*p) {
+	case('('):
+		/* FALLTHROUGH */
+	case('['):
+		/* FALLTHROUGH */
+	case('{'):
+		return(1);
+	default:
+		break;
+	}
+
+	return(0);
+}
+
+
 /*
  * 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
@@ -156,7 +199,7 @@ termp_alloc(enum termenc enc)
  *  possible).
  */
 void
-flushln(struct termp *p)
+term_flushln(struct termp *p)
 {
 	int		 i, j;
 	size_t		 vsz, vis, maxvis, mmax, bp;
@@ -217,21 +260,12 @@ flushln(struct termp *p)
 				for (j = 0; j < (int)p->offset; j++)
 					putchar(' ');
 				vis = 0;
-			} else if (vis + vsz > bp)
-				warnx("word breaks right margin");
-
-			/* TODO: hyphenate. */
-
-		} else {
-			if (vis && vis + vsz > bp) {
-				putchar('\n');
-				for (j = 0; j < (int)p->rmargin; j++)
-					putchar(' ');
-				vis = p->rmargin - p->offset;
-			} else if (vis + vsz > bp) 
-				warnx("word breaks right margin");
-
-			/* TODO: hyphenate. */
+			} 
+		} else if (vis && vis + vsz > bp) {
+			putchar('\n');
+			for (j = 0; j < (int)p->rmargin; j++)
+				putchar(' ');
+			vis = p->rmargin - p->offset;
 		}
 
 		/* 
@@ -289,7 +323,7 @@ flushln(struct termp *p)
  * assertion.
  */
 void
-newln(struct termp *p)
+term_newln(struct termp *p)
 {
 
 	p->flags |= TERMP_NOSPACE;
@@ -297,7 +331,7 @@ newln(struct termp *p)
 		p->flags &= ~TERMP_NOLPAD;
 		return;
 	}
-	flushln(p);
+	term_flushln(p);
 	p->flags &= ~TERMP_NOLPAD;
 }
 
@@ -309,10 +343,10 @@ newln(struct termp *p)
  * assertion.
  */
 void
-vspace(struct termp *p)
+term_vspace(struct termp *p)
 {
 
-	newln(p);
+	term_newln(p);
 	putchar('\n');
 }
 
@@ -324,24 +358,17 @@ vspace(struct termp *p)
  * the word and put it verbatim into the output buffer.
  */
 void
-word(struct termp *p, const char *word)
+term_word(struct termp *p, const char *word)
 {
 	int 		 i, j, len;
 
+	len = (int)strlen(word);
+
 	if (p->flags & TERMP_LITERAL) {
-		termp_pword(p, word, (int)strlen(word));
+		term_pword(p, word, len);
 		return;
 	}
 
-	if (0 == (len = (int)strlen(word)))
-		errx(1, "blank line not in literal context");
-
-	if (mdoc_isdelim(word)) {
-		if ( ! (p->flags & TERMP_IGNDELIM))
-			p->flags |= TERMP_NOSPACE;
-		p->flags &= ~TERMP_IGNDELIM;
-	}
-
 	/* LINTED */
 	for (j = i = 0; i < len; i++) {
 		if (' ' != word[i]) {
@@ -358,208 +385,30 @@ word(struct termp *p, const char *word)
 		if (0 == j)
 			continue;
 		assert(i >= j);
-		termp_pword(p, &word[i - j], j);
+		term_pword(p, &word[i - j], j);
 		j = 0;
 	}
 	if (j > 0) {
 		assert(i >= j);
-		termp_pword(p, &word[i - j], j);
+		term_pword(p, &word[i - j], j);
 	}
 }
 
 
 /*
- * This is the main function for printing out nodes.  It's constituted
- * of PRE and POST functions, which correspond to prefix and infix
- * processing.  The termpair structure allows data to persist between
- * prefix and postfix invocations.
- */
-static void
-termp_body(struct termp *p, struct termpair *ppair,
-		const struct mdoc_meta *meta,
-		const struct mdoc_node *node)
-{
-	int		 dochild;
-	struct termpair	 pair;
-
-	/* Some quick sanity-checking. */
-
-	sanity(node);
-
-	/* Pre-processing. */
-
-	dochild = 1;
-	pair.ppair = ppair;
-	pair.type = 0;
-	pair.offset = pair.rmargin = 0;
-	pair.flag = 0;
-	pair.count = 0;
-
-	if (MDOC_TEXT != node->type) {
-		if (termacts[node->tok].pre)
-			if ( ! (*termacts[node->tok].pre)(p, &pair, meta, node))
-				dochild = 0;
-	} else /* MDOC_TEXT == node->type */
-		word(p, node->string);
-
-	/* Children. */
-
-	if (TERMPAIR_FLAG & pair.type)
-		p->flags |= pair.flag;
-
-	if (dochild && node->child)
-		termp_body(p, &pair, meta, node->child);
-
-	if (TERMPAIR_FLAG & pair.type)
-		p->flags &= ~pair.flag;
-
-	/* Post-processing. */
-
-	if (MDOC_TEXT != node->type)
-		if (termacts[node->tok].post)
-			(*termacts[node->tok].post)(p, &pair, meta, node);
-
-	/* Siblings. */
-
-	if (node->next)
-		termp_body(p, ppair, meta, node->next);
-}
-
-
-static void
-termp_foot(struct termp *p, const struct mdoc_meta *meta)
-{
-	struct tm	*tm;
-	char		*buf, *os;
-
-	if (NULL == (buf = malloc(p->rmargin)))
-		err(1, "malloc");
-	if (NULL == (os = malloc(p->rmargin)))
-		err(1, "malloc");
-
-	tm = localtime(&meta->date);
-
-#ifdef __OpenBSD__
-	if (NULL == strftime(buf, p->rmargin, "%B %d, %Y", tm))
-#else
-	if (0 == strftime(buf, p->rmargin, "%B %d, %Y", tm))
-#endif
-		err(1, "strftime");
-
-	(void)strlcpy(os, meta->os, p->rmargin);
-
-	/*
-	 * This is /slightly/ different from regular groff output
-	 * because we don't have page numbers.  Print the following:
-	 *
-	 * OS                                            MDOCDATE
-	 */
-
-	vspace(p);
-
-	p->flags |= TERMP_NOSPACE | TERMP_NOBREAK;
-	p->rmargin = p->maxrmargin - strlen(buf);
-	p->offset = 0;
-
-	word(p, os);
-	flushln(p);
-
-	p->flags |= TERMP_NOLPAD | TERMP_NOSPACE;
-	p->offset = p->rmargin;
-	p->rmargin = p->maxrmargin;
-	p->flags &= ~TERMP_NOBREAK;
-
-	word(p, buf);
-	flushln(p);
-
-	free(buf);
-	free(os);
-}
-
-
-static void
-termp_head(struct termp *p, const struct mdoc_meta *meta)
-{
-	char		*buf, *title;
-
-	p->rmargin = p->maxrmargin;
-	p->offset = 0;
-
-	if (NULL == (buf = malloc(p->rmargin)))
-		err(1, "malloc");
-	if (NULL == (title = malloc(p->rmargin)))
-		err(1, "malloc");
-
-	/*
-	 * The header is strange.  It has three components, which are
-	 * really two with the first duplicated.  It goes like this:
-	 *
-	 * IDENTIFIER              TITLE                   IDENTIFIER
-	 *
-	 * The IDENTIFIER is NAME(SECTION), which is the command-name
-	 * (if given, or "unknown" if not) followed by the manual page
-	 * section.  These are given in `Dt'.  The TITLE is a free-form
-	 * string depending on the manual volume.  If not specified, it
-	 * switches on the manual section.
-	 */
-
-	assert(meta->vol);
-	(void)strlcpy(buf, meta->vol, p->rmargin);
-
-	if (meta->arch) {
-		(void)strlcat(buf, " (", p->rmargin);
-		(void)strlcat(buf, meta->arch, p->rmargin);
-		(void)strlcat(buf, ")", p->rmargin);
-	}
-
-	(void)snprintf(title, p->rmargin, "%s(%d)", 
-			meta->title, meta->msec);
-
-	p->offset = 0;
-	p->rmargin = (p->maxrmargin - strlen(buf)) / 2;
-	p->flags |= TERMP_NOBREAK | TERMP_NOSPACE;
-
-	word(p, title);
-	flushln(p);
-
-	p->flags |= TERMP_NOLPAD | TERMP_NOSPACE;
-	p->offset = p->rmargin;
-	p->rmargin = p->maxrmargin - strlen(title);
-
-	word(p, buf);
-	flushln(p);
-
-	p->offset = p->rmargin;
-	p->rmargin = p->maxrmargin;
-	p->flags &= ~TERMP_NOBREAK;
-	p->flags |= TERMP_NOLPAD | TERMP_NOSPACE;
-
-	word(p, title);
-	flushln(p);
-
-	p->rmargin = p->maxrmargin;
-	p->offset = 0;
-	p->flags &= ~TERMP_NOSPACE;
-
-	free(title);
-	free(buf);
-}
-
-
-/*
  * Determine the symbol indicated by an escape sequences, that is, one
  * starting with a backslash.  Once done, we pass this value into the
  * output buffer by way of the symbol table.
  */
 static void
-termp_nescape(struct termp *p, const char *word, size_t len)
+term_nescape(struct termp *p, const char *word, size_t len)
 {
 	const char	*rhs;
 	size_t		 sz;
 
-	if (NULL == (rhs = a2ascii(p->symtab, word, len, &sz))) 
+	if (NULL == (rhs = term_a2ascii(p->symtab, word, len, &sz))) 
 		return;
-	termp_stringa(p, rhs, sz);
+	term_stringa(p, rhs, sz);
 }
 
 
@@ -569,7 +418,7 @@ termp_nescape(struct termp *p, const char *word, size_
  * the escape sequence (we assert upon badly-formed escape sequences).
  */
 static void
-termp_pescape(struct termp *p, const char *word, int *i, int len)
+term_pescape(struct termp *p, const char *word, int *i, int len)
 {
 	int		 j;
 
@@ -581,7 +430,7 @@ termp_pescape(struct termp *p, const char *word, int *
 		if (*i + 1 >= len)
 			return;
 
-		termp_nescape(p, &word[*i], 2);
+		term_nescape(p, &word[*i], 2);
 		(*i)++;
 		return;
 
@@ -596,18 +445,18 @@ termp_pescape(struct termp *p, const char *word, int *
 			if (*i + 1 >= len)
 				return;
 
-			termp_nescape(p, &word[*i], 2);
+			term_nescape(p, &word[*i], 2);
 			(*i)++;
 			return;
 		case ('['):
 			break;
 		default:
-			termp_nescape(p, &word[*i], 1);
+			term_nescape(p, &word[*i], 1);
 			return;
 		}
 
 	} else if ('[' != word[*i]) {
-		termp_nescape(p, &word[*i], 1);
+		term_nescape(p, &word[*i], 1);
 		return;
 	}
 
@@ -618,7 +467,7 @@ termp_pescape(struct termp *p, const char *word, int *
 	if (0 == word[*i])
 		return;
 
-	termp_nescape(p, &word[*i - j], (size_t)j);
+	term_nescape(p, &word[*i - j], (size_t)j);
 }
 
 
@@ -628,14 +477,17 @@ termp_pescape(struct termp *p, const char *word, int *
  * handles word styling.
  */
 static void
-termp_pword(struct termp *p, const char *word, int len)
+term_pword(struct termp *p, const char *word, int len)
 {
 	int		 i;
 
-	if ( ! (TERMP_NOSPACE & p->flags) && 
-			! (TERMP_LITERAL & p->flags))
-		termp_chara(p, ' ');
+	if (term_isclosedelim(word, len))
+		if ( ! (TERMP_IGNDELIM & p->flags))
+			p->flags |= TERMP_NOSPACE;
 
+	if ( ! (TERMP_NOSPACE & p->flags))
+		term_chara(p, ' ');
+
 	if ( ! (p->flags & TERMP_NONOSPACE))
 		p->flags &= ~TERMP_NOSPACE;
 
@@ -646,33 +498,36 @@ termp_pword(struct termp *p, const char *word, int len
 
 	for (i = 0; i < len; i++) {
 		if ('\\' == word[i]) {
-			termp_pescape(p, word, &i, len);
+			term_pescape(p, word, &i, len);
 			continue;
 		}
 
 		if (TERMP_STYLE & p->flags) {
 			if (TERMP_BOLD & p->flags) {
-				termp_chara(p, word[i]);
-				termp_chara(p, 8);
+				term_chara(p, word[i]);
+				term_chara(p, 8);
 			}
 			if (TERMP_UNDER & p->flags) {
-				termp_chara(p, '_');
-				termp_chara(p, 8);
+				term_chara(p, '_');
+				term_chara(p, 8);
 			}
 		}
 
-		termp_chara(p, word[i]);
+		term_chara(p, word[i]);
 	}
+
+	if (term_isopendelim(word, len))
+		p->flags |= TERMP_NOSPACE;
 }
 
 
 /*
- * Like termp_chara() but for arbitrary-length buffers.  Resize the
+ * Like term_chara() but for arbitrary-length buffers.  Resize the
  * buffer by a factor of two (if the buffer is less than that) or the
  * buffer's size.
  */
 static void
-termp_stringa(struct termp *p, const char *c, size_t sz)
+term_stringa(struct termp *p, const char *c, size_t sz)
 {
 	size_t		 s;
 
@@ -701,7 +556,7 @@ termp_stringa(struct termp *p, const char *c, size_t s
  * size.
  */
 static void
-termp_chara(struct termp *p, char c)
+term_chara(struct termp *p, char c)
 {
 	size_t		 s;
 
@@ -717,112 +572,3 @@ termp_chara(struct termp *p, char c)
 	p->buf[(int)(p->col)++] = c;
 }
 
-
-static void
-sanity(const struct mdoc_node *n)
-{
-
-	switch (n->type) {
-	case (MDOC_TEXT):
-		if (n->child) 
-			errx(1, "regular form violated (1)");
-		if (NULL == n->parent) 
-			errx(1, "regular form violated (2)");
-		if (NULL == n->string)
-			errx(1, "regular form violated (3)");
-		switch (n->parent->type) {
-		case (MDOC_TEXT):
-			/* FALLTHROUGH */
-		case (MDOC_ROOT):
-			errx(1, "regular form violated (4)");
-			/* NOTREACHED */
-		default:
-			break;
-		}
-		break;
-	case (MDOC_ELEM):
-		if (NULL == n->parent)
-			errx(1, "regular form violated (5)");
-		switch (n->parent->type) {
-		case (MDOC_TAIL):
-			/* FALLTHROUGH */
-		case (MDOC_BODY):
-			/* FALLTHROUGH */
-		case (MDOC_HEAD):
-			break;
-		default:
-			errx(1, "regular form violated (6)");
-			/* NOTREACHED */
-		}
-		if (n->child) switch (n->child->type) {
-		case (MDOC_TEXT):
-			break;
-		default:
-			errx(1, "regular form violated (7(");
-			/* NOTREACHED */
-		}
-		break;
-	case (MDOC_HEAD):
-		/* FALLTHROUGH */
-	case (MDOC_BODY):
-		/* FALLTHROUGH */
-	case (MDOC_TAIL):
-		if (NULL == n->parent)
-			errx(1, "regular form violated (8)");
-		if (MDOC_BLOCK != n->parent->type)
-			errx(1, "regular form violated (9)");
-		if (n->child) switch (n->child->type) {
-		case (MDOC_BLOCK):
-			/* FALLTHROUGH */
-		case (MDOC_ELEM):
-			/* FALLTHROUGH */
-		case (MDOC_TEXT):
-			break;
-		default:
-			errx(1, "regular form violated (a)");
-			/* NOTREACHED */
-		}
-		break;
-	case (MDOC_BLOCK):
-		if (NULL == n->parent)
-			errx(1, "regular form violated (b)");
-		if (NULL == n->child)
-			errx(1, "regular form violated (c)");
-		switch (n->parent->type) {
-		case (MDOC_ROOT):
-			/* FALLTHROUGH */
-		case (MDOC_HEAD):
-			/* FALLTHROUGH */
-		case (MDOC_BODY):
-			/* FALLTHROUGH */
-		case (MDOC_TAIL):
-			break;
-		default:
-			errx(1, "regular form violated (d)");
-			/* NOTREACHED */
-		}
-		switch (n->child->type) {
-		case (MDOC_ROOT):
-			/* FALLTHROUGH */
-		case (MDOC_ELEM):
-			errx(1, "regular form violated (e)");
-			/* NOTREACHED */
-		default:
-			break;
-		}
-		break;
-	case (MDOC_ROOT):
-		if (n->parent)
-			errx(1, "regular form violated (f)");
-		if (NULL == n->child)
-			errx(1, "regular form violated (10)");
-		switch (n->child->type) {
-		case (MDOC_BLOCK):
-			break;
-		default:
-			errx(1, "regular form violated (11)");
-			/* NOTREACHED */
-		}
-		break;
-	}
-}