Move mdoc_isdelim() into mandoc.h as mandoc_isdelim(). This allows the

removal of manual delimiter checks in html.c and term.c.  Finally, add
the escaped period as a closing delimiter, removing a TODO to this
effect.

9 files changed, 88 insertions, 150 deletions

diff --git a/TODO b/TODO
index a916a1ee..b2abfc0e 100644
--- a/TODO
+++ b/TODO
@@ -10,15 +10,6 @@
 - .TP before .SH is still FATAL in man(7)
   reported by brad@  Sat, 15 Jan 2011 15:54:54 -0500
 
-- The mdoc(7) parser should handle `\.' as punctuation:
-    .Ql .if ${VAR} op something \.
-  should produce
-    `.if ${VAR} op something'.
-  not
-    `.if ${VAR} op something .'
-  as seen in make(1).
-  To escape the dot, one would use `\&.'.
-
 - the roff parser doesn't tolerate additional characters between
   a macro and the \} terminating a conditional block, e.g.
   .if n \{
diff --git a/html.c b/html.c
index 5c042581..381a60ed 100644
--- a/html.c
+++ b/html.c
@@ -514,29 +514,9 @@ void
 print_text(struct html *h, const char *word)
 {
 
-	if (word[0] && '\0' == word[1])
-		switch (word[0]) {
-		case('.'):
-			/* FALLTHROUGH */
-		case(','):
-			/* FALLTHROUGH */
-		case(';'):
-			/* FALLTHROUGH */
-		case(':'):
-			/* FALLTHROUGH */
-		case('?'):
-			/* FALLTHROUGH */
-		case('!'):
-			/* FALLTHROUGH */
-		case(')'):
-			/* FALLTHROUGH */
-		case(']'):
-			if ( ! (HTML_IGNDELIM & h->flags))
-				h->flags |= HTML_NOSPACE;
-			break;
-		default:
-			break;
-		}
+	if (DELIM_CLOSE == mandoc_isdelim(word))
+		if ( ! (HTML_IGNDELIM & h->flags))
+			h->flags |= HTML_NOSPACE;
 
 	if ( ! (HTML_NOSPACE & h->flags)) {
 		/* Manage keeps! */
@@ -566,20 +546,8 @@ print_text(struct html *h, const char *word)
 
 	h->flags &= ~HTML_IGNDELIM;
 
-	/* 
-	 * Note that we don't process the pipe: the parser sees it as
-	 * punctuation, but we don't in terms of typography.
-	 */
-	if (word[0] && '\0' == word[1])
-		switch (word[0]) {
-		case('('):
-			/* FALLTHROUGH */
-		case('['):
-			h->flags |= HTML_NOSPACE;
-			break;
-		default:
-			break;
-		}
+	if (DELIM_OPEN == mandoc_isdelim(word))
+		h->flags |= HTML_NOSPACE;
 }
 
 
diff --git a/libmdoc.h b/libmdoc.h
index 947254fb..372f5063 100644
--- a/libmdoc.h
+++ b/libmdoc.h
@@ -81,13 +81,6 @@ enum	margverr {
 	ARGV_WORD
 };
 
-enum	mdelim {
-	DELIM_NONE = 0,
-	DELIM_OPEN,
-	DELIM_MIDDLE,
-	DELIM_CLOSE
-};
-
 extern	const struct mdoc_macro *const mdoc_macros;
 
 __BEGIN_DECLS
@@ -114,8 +107,6 @@ int		  mdoc_endbody_alloc(struct mdoc *m, int line, int pos,
 void		  mdoc_node_delete(struct mdoc *, struct mdoc_node *);
 void		  mdoc_hash_init(void);
 enum mdoct	  mdoc_hash_find(const char *);
-#define	DELIMSZ	  6 /* maximum size of a delimiter string */
-enum mdelim	  mdoc_isdelim(const char *);
 size_t		  mdoc_isescape(const char *);
 enum	mdoc_sec  mdoc_str2sec(const char *);
 time_t		  mdoc_atotime(const char *);
diff --git a/mandoc.c b/mandoc.c
index e4c5ca6f..aa934217 100644
--- a/mandoc.c
+++ b/mandoc.c
@@ -363,7 +363,6 @@ mandoc_getarg(char **cpp, mandocmsg msg, void *data, int ln, int *pos)
 	return(start);
 }
 
-
 static int
 a2time(time_t *t, const char *fmt, const char *p)
 {
@@ -381,7 +380,6 @@ a2time(time_t *t, const char *fmt, const char *p)
 	return(0);
 }
 
-
 static char *
 time2a(time_t t)
 {
@@ -417,7 +415,6 @@ fail:
 	return(NULL);
 }
 
-
 char *
 mandoc_normdate(char *in, mandocmsg msg, void *data, int ln, int pos)
 {
@@ -439,7 +436,6 @@ mandoc_normdate(char *in, mandocmsg msg, void *data, int ln, int pos)
 	return(out ? out : mandoc_strdup(in));
 }
 
-
 int
 mandoc_eos(const char *p, size_t sz, int enclosed)
 {
@@ -483,7 +479,6 @@ mandoc_eos(const char *p, size_t sz, int enclosed)
 	return(found && !enclosed);
 }
 
-
 int
 mandoc_hyph(const char *start, const char *c)
 {
@@ -510,3 +505,54 @@ mandoc_hyph(const char *start, const char *c)
 
 	return(1);
 }
+
+/*
+ * Check if a string is a punctuation delimiter.  This only applies to
+ * mdoc(7) documents, but as it's used in both front-ends and back-ends,
+ * it needs to go here (instead of, say, in libmdoc.h).
+ */
+enum mdelim
+mandoc_isdelim(const char *p)
+{
+
+	if ('\0' == p[0])
+		return(DELIM_NONE);
+
+	if ('\0' == p[1])
+		switch (p[0]) {
+		case('('):
+			/* FALLTHROUGH */
+		case('['):
+			return(DELIM_OPEN);
+		case('|'):
+			return(DELIM_MIDDLE);
+		case('.'):
+			/* FALLTHROUGH */
+		case(','):
+			/* FALLTHROUGH */
+		case(';'):
+			/* FALLTHROUGH */
+		case(':'):
+			/* FALLTHROUGH */
+		case('?'):
+			/* FALLTHROUGH */
+		case('!'):
+			/* FALLTHROUGH */
+		case(')'):
+			/* FALLTHROUGH */
+		case(']'):
+			return(DELIM_CLOSE);
+		default:
+			return(DELIM_NONE);
+		}
+
+	if ('\\' != p[0])
+		return(DELIM_NONE);
+
+	if (0 == strcmp(p, "\\."))
+		return(DELIM_CLOSE);
+	if (0 == strcmp(p, "\\*(Ba"))
+		return(DELIM_MIDDLE);
+
+	return(DELIM_NONE);
+}
diff --git a/mandoc.h b/mandoc.h
index 71cca22c..dfe0c546 100644
--- a/mandoc.h
+++ b/mandoc.h
@@ -312,6 +312,22 @@ struct	regset {
 	struct reg	  regs[REG__MAX];
 };
 
+/*
+ * A punctuation delimiter, used only in mdoc(7) documents, is opening,
+ * closing, or "middle mark" punctuation.  These govern spacing.
+ * Opening punctuation (e.g., the opening parenthesis) suppresses the
+ * following space; closing punctuation (e.g., the closing parenthesis)
+ * suppresses the leading space; middle punctuation (e.g., the vertical
+ * bar) can do either.  The middle punctuation delimiter bends the rules
+ * depending on usage.
+ */
+enum	mdelim {
+	DELIM_NONE = 0,
+	DELIM_OPEN,
+	DELIM_MIDDLE,
+	DELIM_CLOSE
+};
+
 typedef	int	(*mandocmsg)(enum mandocerr, void *,
 			int, int, const char *);
 
@@ -320,6 +336,8 @@ __BEGIN_DECLS
 void		 *mandoc_calloc(size_t, size_t);
 void		 *mandoc_malloc(size_t);
 void		 *mandoc_realloc(void *, size_t);
+#define	DELIMSZ	  6 /* hint: max possible size of a delimiter */
+enum mdelim	  mandoc_isdelim(const char *);
 
 __END_DECLS
 
diff --git a/mdoc_argv.c b/mdoc_argv.c
index 12c15966..89f30fd7 100644
--- a/mdoc_argv.c
+++ b/mdoc_argv.c
@@ -587,7 +587,7 @@ args_checkpunct(const char *p)
 		return(0);
 
 	buf[j] = '\0';
-	if (DELIM_CLOSE != mdoc_isdelim(buf))
+	if (DELIM_CLOSE != mandoc_isdelim(buf))
 		return(0);
 
 	while (' ' == p[i])
@@ -604,7 +604,7 @@ args_checkpunct(const char *p)
 			return(0);
 
 		buf[j] = '\0';
-		d = mdoc_isdelim(buf);
+		d = mandoc_isdelim(buf);
 		if (DELIM_NONE == d || DELIM_OPEN == d)
 			return(0);
 
diff --git a/mdoc_macro.c b/mdoc_macro.c
index ac5b2c24..74e80d19 100644
--- a/mdoc_macro.c
+++ b/mdoc_macro.c
@@ -599,7 +599,7 @@ append_delims(struct mdoc *m, int line, int *pos, char *buf)
 		else if (ARGS_EOLN == ac)
 			break;
 
-		assert(DELIM_NONE != mdoc_isdelim(p));
+		assert(DELIM_NONE != mandoc_isdelim(p));
 		if ( ! mdoc_word_alloc(m, line, la, p))
 			return(0);
 
@@ -870,7 +870,7 @@ in_line(MACRO_PROT_ARGS)
 		 * the word. 
 		 */
 
-		d = ARGS_QWORD == ac ? DELIM_NONE : mdoc_isdelim(p);
+		d = ARGS_QWORD == ac ? DELIM_NONE : mandoc_isdelim(p);
 
 		if (DELIM_NONE != d) {
 			/*
@@ -1061,7 +1061,7 @@ blk_full(MACRO_PROT_ARGS)
 				ARGS_PHRASE != ac &&
 				ARGS_PPHRASE != ac &&
 				ARGS_QWORD != ac &&
-				DELIM_OPEN == mdoc_isdelim(p)) {
+				DELIM_OPEN == mandoc_isdelim(p)) {
 			if ( ! mdoc_word_alloc(m, line, la, p))
 				return(0);
 			continue;
@@ -1224,7 +1224,7 @@ blk_part_imp(MACRO_PROT_ARGS)
 			break;
 
 		if (NULL == body && ARGS_QWORD != ac &&
-		    DELIM_OPEN == mdoc_isdelim(p)) {
+		    DELIM_OPEN == mandoc_isdelim(p)) {
 			if ( ! mdoc_word_alloc(m, line, la, p))
 				return(0);
 			continue;
@@ -1355,7 +1355,7 @@ blk_part_exp(MACRO_PROT_ARGS)
 		/* Flush out leading punctuation. */
 
 		if (NULL == head && ARGS_QWORD != ac &&
-		    DELIM_OPEN == mdoc_isdelim(p)) {
+		    DELIM_OPEN == mandoc_isdelim(p)) {
 			assert(NULL == body);
 			if ( ! mdoc_word_alloc(m, line, la, p))
 				return(0);
@@ -1502,7 +1502,7 @@ in_line_argn(MACRO_PROT_ARGS)
 
 		if ( ! (MDOC_IGNDELIM & mdoc_macros[tok].flags) && 
 				ARGS_QWORD != ac &&
-				0 == j && DELIM_OPEN == mdoc_isdelim(p)) {
+				0 == j && DELIM_OPEN == mandoc_isdelim(p)) {
 			if ( ! mdoc_word_alloc(m, line, la, p))
 				return(0);
 			continue;
@@ -1531,7 +1531,7 @@ in_line_argn(MACRO_PROT_ARGS)
 		if ( ! (MDOC_IGNDELIM & mdoc_macros[tok].flags) &&
 				ARGS_QWORD != ac &&
 				! flushed &&
-				DELIM_NONE != mdoc_isdelim(p)) {
+				DELIM_NONE != mandoc_isdelim(p)) {
 			if ( ! rew_elem(m, tok))
 				return(0);
 			flushed = 1;
diff --git a/mdoc_strings.c b/mdoc_strings.c
index cd4654b4..8f2dddbc 100644
--- a/mdoc_strings.c
+++ b/mdoc_strings.c
@@ -54,50 +54,6 @@ static	const char * const secnames[SEC__MAX] = {
 	NULL
 };
 
-enum mdelim
-mdoc_isdelim(const char *p)
-{
-
-	if ('\0' == p[0])
-		return(DELIM_NONE);
-
-	if ('\0' == p[1])
-		switch (p[0]) {
-		case('('):
-			/* FALLTHROUGH */
-		case('['):
-			return(DELIM_OPEN);
-		case('|'):
-			return(DELIM_MIDDLE);
-		case('.'):
-			/* FALLTHROUGH */
-		case(','):
-			/* FALLTHROUGH */
-		case(';'):
-			/* FALLTHROUGH */
-		case(':'):
-			/* FALLTHROUGH */
-		case('?'):
-			/* FALLTHROUGH */
-		case('!'):
-			/* FALLTHROUGH */
-		case(')'):
-			/* FALLTHROUGH */
-		case(']'):
-			return(DELIM_CLOSE);
-		default:
-			return(DELIM_NONE);
-		}
-
-	/*
-	 * XXX; account for groff bubu where the \*(Ba reserved string
-	 * is treated in exactly the same way as the vertical bar.  This
-	 * is the only function that checks for this.
-	 */
-	return(strcmp(p, "\\*(Ba") ? DELIM_NONE : DELIM_MIDDLE);
-}
-
-
 enum mdoc_sec 
 mdoc_str2sec(const char *p)
 {
diff --git a/term.c b/term.c
index a106224b..2e24eb3d 100644
--- a/term.c
+++ b/term.c
@@ -463,29 +463,9 @@ term_word(struct termp *p, const char *word)
 
 	sv = word;
 
-	if (word[0] && '\0' == word[1])
-		switch (word[0]) {
-		case('.'):
-			/* FALLTHROUGH */
-		case(','):
-			/* FALLTHROUGH */
-		case(';'):
-			/* FALLTHROUGH */
-		case(':'):
-			/* FALLTHROUGH */
-		case('?'):
-			/* FALLTHROUGH */
-		case('!'):
-			/* FALLTHROUGH */
-		case(')'):
-			/* FALLTHROUGH */
-		case(']'):
-			if ( ! (TERMP_IGNDELIM & p->flags))
-				p->flags |= TERMP_NOSPACE;
-			break;
-		default:
-			break;
-		}
+	if (DELIM_CLOSE == mandoc_isdelim(word))
+		if ( ! (TERMP_IGNDELIM & p->flags))
+			p->flags |= TERMP_NOSPACE;
 
 	if ( ! (TERMP_NOSPACE & p->flags)) {
 		if ( ! (TERMP_KEEP & p->flags)) {
@@ -548,20 +528,8 @@ term_word(struct termp *p, const char *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('['):
-			p->flags |= TERMP_NOSPACE;
-			break;
-		default:
-			break;
-		}
+	if (DELIM_OPEN == mandoc_isdelim(sv))
+		p->flags |= TERMP_NOSPACE;
 }