===================================================================
RCS file: /cvs/mandoc/mdoc_html.c,v
retrieving revision 1.269
retrieving revision 1.282
diff -u -p -r1.269 -r1.282
--- mandoc/mdoc_html.c	2017/02/05 21:00:43	1.269
+++ mandoc/mdoc_html.c	2017/05/04 22:16:09	1.282
@@ -1,4 +1,4 @@
-/*	$Id: mdoc_html.c,v 1.269 2017/02/05 21:00:43 schwarze Exp $ */
+/*	$Id: mdoc_html.c,v 1.282 2017/05/04 22:16:09 schwarze Exp $ */
 /*
  * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2014, 2015, 2016, 2017 Ingo Schwarze <schwarze@openbsd.org>
@@ -48,7 +48,7 @@ struct	htmlmdoc {
 	void		(*post)(MDOC_ARGS);
 };
 
-static	char		 *make_id(const struct roff_node *);
+static	char		 *cond_id(const struct roff_node *);
 static	void		  print_mdoc_head(MDOC_ARGS);
 static	void		  print_mdoc_node(MDOC_ARGS);
 static	void		  print_mdoc_nodelist(MDOC_ARGS);
@@ -118,8 +118,7 @@ static	int		  mdoc_vt_pre(MDOC_ARGS);
 static	int		  mdoc_xr_pre(MDOC_ARGS);
 static	int		  mdoc_xx_pre(MDOC_ARGS);
 
-static	const struct htmlmdoc mdocs[MDOC_MAX] = {
-	{mdoc_ap_pre, NULL}, /* Ap */
+static	const struct htmlmdoc __mdocs[MDOC_MAX - MDOC_Dd] = {
 	{NULL, NULL}, /* Dd */
 	{NULL, NULL}, /* Dt */
 	{NULL, NULL}, /* Os */
@@ -135,6 +134,7 @@ static	const struct htmlmdoc mdocs[MDOC_MAX] = {
 	{mdoc_it_pre, NULL}, /* It */
 	{mdoc_ad_pre, NULL}, /* Ad */
 	{mdoc_an_pre, NULL}, /* An */
+	{mdoc_ap_pre, NULL}, /* Ap */
 	{mdoc_ar_pre, NULL}, /* Ar */
 	{mdoc_cd_pre, NULL}, /* Cd */
 	{mdoc_cm_pre, NULL}, /* Cm */
@@ -237,12 +237,12 @@ static	const struct htmlmdoc mdocs[MDOC_MAX] = {
 	{mdoc_quote_pre, mdoc_quote_post}, /* En */
 	{mdoc_xx_pre, NULL}, /* Dx */
 	{mdoc__x_pre, mdoc__x_post}, /* %Q */
-	{mdoc_sp_pre, NULL}, /* br */
 	{mdoc_sp_pre, NULL}, /* sp */
 	{mdoc__x_pre, mdoc__x_post}, /* %U */
 	{NULL, NULL}, /* Ta */
 	{mdoc_skip_pre, NULL}, /* ll */
 };
+static	const struct htmlmdoc *const mdocs = __mdocs - MDOC_Dd;
 
 
 /*
@@ -393,7 +393,13 @@ print_mdoc_node(MDOC_ARGS)
 			t = h->tag;
 		}
 		assert(h->tblt == NULL);
-		if (mdocs[n->tok].pre && (n->end == ENDBODY_NOT || n->child))
+		if (n->tok < ROFF_MAX) {
+			roff_html_pre(h, n);
+			break;
+		}
+		assert(n->tok >= MDOC_Dd && n->tok < MDOC_MAX);
+		if (mdocs[n->tok].pre != NULL &&
+		    (n->end == ENDBODY_NOT || n->child != NULL))
 			child = (*mdocs[n->tok].pre)(meta, n, h);
 		break;
 	}
@@ -412,13 +418,13 @@ print_mdoc_node(MDOC_ARGS)
 	case ROFFT_EQN:
 		break;
 	default:
-		if ( ! mdocs[n->tok].post || n->flags & NODE_ENDED)
+		if (n->tok < ROFF_MAX ||
+		    mdocs[n->tok].post == NULL ||
+		    n->flags & NODE_ENDED)
 			break;
 		(*mdocs[n->tok].post)(meta, n, h);
 		if (n->end != ENDBODY_NOT)
 			n->body->flags |= NODE_ENDED;
-		if (n->end == ENDBODY_NOSPACE)
-			h->flags |= HTML_NOSPACE;
 		break;
 	}
 }
@@ -479,25 +485,19 @@ mdoc_root_pre(MDOC_ARGS)
 }
 
 static char *
-make_id(const struct roff_node *n)
+cond_id(const struct roff_node *n)
 {
-	const struct roff_node	*nch;
-	char			*buf, *cp;
-
-	for (nch = n->child; nch != NULL; nch = nch->next)
-		if (nch->type != ROFFT_TEXT)
-			return NULL;
-
-	buf = NULL;
-	deroff(&buf, n);
-
-	/* http://www.w3.org/TR/html5/dom.html#the-id-attribute */
-
-	for (cp = buf; *cp != '\0'; cp++)
-		if (*cp == ' ')
-			*cp = '_';
-
-	return buf;
+	if (n->child != NULL &&
+	    n->child->type == ROFFT_TEXT &&
+	    (n->prev == NULL ||
+	     (n->prev->type == ROFFT_TEXT &&
+	      strcmp(n->prev->string, "|") == 0)) &&
+	    (n->parent->tok == MDOC_It ||
+	     (n->parent->tok == MDOC_Xo &&
+	      n->parent->parent->prev == NULL &&
+	      n->parent->parent->parent->tok == MDOC_It)))
+		return html_make_id(n);
+	return NULL;
 }
 
 static int
@@ -507,8 +507,10 @@ mdoc_sh_pre(MDOC_ARGS)
 
 	switch (n->type) {
 	case ROFFT_HEAD:
-		id = make_id(n);
-		print_otag(h, TAG_H1, "ci", "Sh", id);
+		id = html_make_id(n);
+		print_otag(h, TAG_H1, "cTi", "Sh", id);
+		if (id != NULL)
+			print_otag(h, TAG_A, "chR", "selflink", id);
 		free(id);
 		break;
 	case ROFFT_BODY:
@@ -529,8 +531,10 @@ mdoc_ss_pre(MDOC_ARGS)
 	if (n->type != ROFFT_HEAD)
 		return 1;
 
-	id = make_id(n);
-	print_otag(h, TAG_H2, "ci", "Ss", id);
+	id = html_make_id(n);
+	print_otag(h, TAG_H2, "cTi", "Ss", id);
+	if (id != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
 	free(id);
 	return 1;
 }
@@ -538,9 +542,14 @@ mdoc_ss_pre(MDOC_ARGS)
 static int
 mdoc_fl_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_B, "c", "Fl");
-	print_text(h, "\\-");
+	char	*id;
 
+	if ((id = cond_id(n)) != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
+	print_otag(h, TAG_B, "cTi", "Fl", id);
+	free(id);
+
+	print_text(h, "\\-");
 	if (!(n->child == NULL &&
 	    (n->next == NULL ||
 	     n->next->type == ROFFT_TEXT ||
@@ -553,7 +562,12 @@ mdoc_fl_pre(MDOC_ARGS)
 static int
 mdoc_cm_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_B, "c", "Cm");
+	char	*id;
+
+	if ((id = cond_id(n)) != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
+	print_otag(h, TAG_B, "cTi", "Cm", id);
+	free(id);
 	return 1;
 }
 
@@ -566,7 +580,7 @@ mdoc_nd_pre(MDOC_ARGS)
 	/* XXX: this tag in theory can contain block elements. */
 
 	print_text(h, "\\(em");
-	print_otag(h, TAG_SPAN, "c", "Nd");
+	print_otag(h, TAG_SPAN, "cT", "Nd");
 	return 1;
 }
 
@@ -581,9 +595,7 @@ mdoc_nm_pre(MDOC_ARGS)
 		print_otag(h, TAG_TD, "");
 		/* FALLTHROUGH */
 	case ROFFT_ELEM:
-		print_otag(h, TAG_B, "c", "Nm");
-		if (n->child == NULL && meta->name != NULL)
-			print_text(h, meta->name);
+		print_otag(h, TAG_B, "cT", "Nm");
 		return 1;
 	case ROFFT_BODY:
 		print_otag(h, TAG_TD, "");
@@ -603,7 +615,8 @@ mdoc_nm_pre(MDOC_ARGS)
 		len = html_strlen(meta->name);
 
 	t = print_otag(h, TAG_COLGROUP, "");
-	print_otag(h, TAG_COL, "shw", len);
+	/* Increase width to make even bold text fit. */
+	print_otag(h, TAG_COL, "shw", len + 2);
 	print_otag(h, TAG_COL, "");
 	print_tagq(h, t);
 	print_otag(h, TAG_TR, "");
@@ -617,11 +630,11 @@ mdoc_xr_pre(MDOC_ARGS)
 		return 0;
 
 	if (h->base_man)
-		print_otag(h, TAG_A, "chM", "Xr",
+		print_otag(h, TAG_A, "cThM", "Xr",
 		    n->child->string, n->child->next == NULL ?
 		    NULL : n->child->next->string);
 	else
-		print_otag(h, TAG_A, "c", "Xr");
+		print_otag(h, TAG_A, "cT", "Xr");
 
 	n = n->child;
 	print_text(h, n->string);
@@ -650,7 +663,7 @@ mdoc_ns_pre(MDOC_ARGS)
 static int
 mdoc_ar_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_VAR, "c", "Ar");
+	print_otag(h, TAG_VAR, "cT", "Ar");
 	return 1;
 }
 
@@ -670,7 +683,7 @@ mdoc_it_pre(MDOC_ARGS)
 	enum mdoc_list		 type;
 
 	bl = n->parent;
-	while (bl != NULL && bl->tok != MDOC_Bl)
+	while (bl->tok != MDOC_Bl)
 		bl = bl->parent;
 	type = bl->norm->Bl.type;
 
@@ -755,6 +768,7 @@ mdoc_it_pre(MDOC_ARGS)
 		case ROFFT_HEAD:
 			if (h->style != NULL && !bl->norm->Bl.comp &&
 			    (n->parent->prev == NULL ||
+			     n->parent->prev->body == NULL ||
 			     n->parent->prev->body->child != NULL)) {
 				t = print_otag(h, TAG_DT, "csWl",
 				    cattr, bl->norm->Bl.width);
@@ -898,14 +912,14 @@ mdoc_ex_pre(MDOC_ARGS)
 static int
 mdoc_st_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_SPAN, "c", "St");
+	print_otag(h, TAG_SPAN, "cT", "St");
 	return 1;
 }
 
 static int
 mdoc_em_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_I, "c", "Em");
+	print_otag(h, TAG_I, "cT", "Em");
 	return 1;
 }
 
@@ -928,8 +942,8 @@ mdoc_sx_pre(MDOC_ARGS)
 {
 	char	*id;
 
-	id = make_id(n);
-	print_otag(h, TAG_A, "chR", "Sx", id);
+	id = html_make_id(n);
+	print_otag(h, TAG_A, "cThR", "Sx", id);
 	free(id);
 	return 1;
 }
@@ -996,7 +1010,7 @@ mdoc_bd_pre(MDOC_ARGS)
 		 */
 		switch (nn->tok) {
 		case MDOC_Sm:
-		case MDOC_br:
+		case ROFF_br:
 		case MDOC_sp:
 		case MDOC_Bl:
 		case MDOC_D1:
@@ -1025,7 +1039,7 @@ mdoc_bd_pre(MDOC_ARGS)
 static int
 mdoc_pa_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_I, "c", "Pa");
+	print_otag(h, TAG_I, "cT", "Pa");
 	return 1;
 }
 
@@ -1056,7 +1070,7 @@ mdoc_an_pre(MDOC_ARGS)
 	if (n->sec == SEC_AUTHORS && ! (h->flags & HTML_NOSPLIT))
 		h->flags |= HTML_SPLIT;
 
-	print_otag(h, TAG_SPAN, "c", "An");
+	print_otag(h, TAG_SPAN, "cT", "An");
 	return 1;
 }
 
@@ -1064,28 +1078,49 @@ static int
 mdoc_cd_pre(MDOC_ARGS)
 {
 	synopsis_pre(h, n);
-	print_otag(h, TAG_B, "c", "Cd");
+	print_otag(h, TAG_B, "cT", "Cd");
 	return 1;
 }
 
 static int
 mdoc_dv_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_CODE, "c", "Dv");
+	char	*id;
+
+	if ((id = cond_id(n)) != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
+	print_otag(h, TAG_CODE, "cTi", "Dv", id);
+	free(id);
 	return 1;
 }
 
 static int
 mdoc_ev_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_CODE, "c", "Ev");
+	char	*id;
+
+	if ((id = cond_id(n)) != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
+	print_otag(h, TAG_CODE, "cTi", "Ev", id);
+	free(id);
 	return 1;
 }
 
 static int
 mdoc_er_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_CODE, "c", "Er");
+	char	*id;
+
+	id = n->sec == SEC_ERRORS &&
+	    (n->parent->tok == MDOC_It ||
+	     (n->parent->tok == MDOC_Bq &&
+	      n->parent->parent->parent->tok == MDOC_It)) ?
+	    html_make_id(n) : NULL;
+
+	if (id != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
+	print_otag(h, TAG_CODE, "cTi", "Er", id);
+	free(id);
 	return 1;
 }
 
@@ -1096,12 +1131,12 @@ mdoc_fa_pre(MDOC_ARGS)
 	struct tag		*t;
 
 	if (n->parent->tok != MDOC_Fo) {
-		print_otag(h, TAG_VAR, "c", "Fa");
+		print_otag(h, TAG_VAR, "cT", "Fa");
 		return 1;
 	}
 
 	for (nn = n->child; nn; nn = nn->next) {
-		t = print_otag(h, TAG_VAR, "c", "Fa");
+		t = print_otag(h, TAG_VAR, "cT", "Fa");
 		print_text(h, nn->string);
 		print_tagq(h, t);
 		if (nn->next) {
@@ -1132,11 +1167,11 @@ mdoc_fd_pre(MDOC_ARGS)
 	assert(n->type == ROFFT_TEXT);
 
 	if (strcmp(n->string, "#include")) {
-		print_otag(h, TAG_B, "c", "Fd");
+		print_otag(h, TAG_B, "cT", "Fd");
 		return 1;
 	}
 
-	print_otag(h, TAG_B, "c", "In");
+	print_otag(h, TAG_B, "cT", "In");
 	print_text(h, n->string);
 
 	if (NULL != (n = n->next)) {
@@ -1150,10 +1185,10 @@ mdoc_fd_pre(MDOC_ARGS)
 			cp = strchr(buf, '\0') - 1;
 			if (cp >= buf && (*cp == '>' || *cp == '"'))
 				*cp = '\0';
-			t = print_otag(h, TAG_A, "chI", "In", buf);
+			t = print_otag(h, TAG_A, "cThI", "In", buf);
 			free(buf);
 		} else
-			t = print_otag(h, TAG_A, "c", "In");
+			t = print_otag(h, TAG_A, "cT", "In");
 
 		print_text(h, n->string);
 		print_tagq(h, t);
@@ -1180,7 +1215,7 @@ mdoc_vt_pre(MDOC_ARGS)
 	} else if (n->type == ROFFT_HEAD)
 		return 0;
 
-	print_otag(h, TAG_VAR, "c", "Vt");
+	print_otag(h, TAG_VAR, "cT", "Vt");
 	return 1;
 }
 
@@ -1188,7 +1223,7 @@ static int
 mdoc_ft_pre(MDOC_ARGS)
 {
 	synopsis_pre(h, n);
-	print_otag(h, TAG_VAR, "c", "Ft");
+	print_otag(h, TAG_VAR, "cT", "Ft");
 	return 1;
 }
 
@@ -1209,7 +1244,7 @@ mdoc_fn_pre(MDOC_ARGS)
 
 	ep = strchr(sp, ' ');
 	if (NULL != ep) {
-		t = print_otag(h, TAG_VAR, "c", "Ft");
+		t = print_otag(h, TAG_VAR, "cT", "Ft");
 
 		while (ep) {
 			sz = MIN((int)(ep - sp), BUFSIZ - 1);
@@ -1222,7 +1257,7 @@ mdoc_fn_pre(MDOC_ARGS)
 		print_tagq(h, t);
 	}
 
-	t = print_otag(h, TAG_B, "c", "Fn");
+	t = print_otag(h, TAG_B, "cT", "Fn");
 
 	if (sp)
 		print_text(h, sp);
@@ -1235,10 +1270,10 @@ mdoc_fn_pre(MDOC_ARGS)
 
 	for (n = n->child->next; n; n = n->next) {
 		if (NODE_SYNPRETTY & n->flags)
-			t = print_otag(h, TAG_VAR, "css?", "Fa",
+			t = print_otag(h, TAG_VAR, "cTss?", "Fa",
 			    "white-space", "nowrap");
 		else
-			t = print_otag(h, TAG_VAR, "c", "Fa");
+			t = print_otag(h, TAG_VAR, "cT", "Fa");
 		print_text(h, n->string);
 		print_tagq(h, t);
 		if (n->next) {
@@ -1296,17 +1331,13 @@ mdoc_sp_pre(MDOC_ARGS)
 	struct roffsu	 su;
 
 	SCALE_VS_INIT(&su, 1);
+	if (NULL != (n = n->child)) {
+		if ( ! a2roffsu(n->string, &su, SCALE_VS))
+			su.scale = 1.0;
+		else if (su.scale < 0.0)
+			su.scale = 0.0;
+	}
 
-	if (MDOC_sp == n->tok) {
-		if (NULL != (n = n->child)) {
-			if ( ! a2roffsu(n->string, &su, SCALE_VS))
-				su.scale = 1.0;
-			else if (su.scale < 0.0)
-				su.scale = 0.0;
-		}
-	} else
-		su.scale = 0.0;
-
 	print_otag(h, TAG_DIV, "suh", &su);
 
 	/* So the div isn't empty: */
@@ -1319,19 +1350,25 @@ mdoc_sp_pre(MDOC_ARGS)
 static int
 mdoc_lk_pre(MDOC_ARGS)
 {
-	if (NULL == (n = n->child))
+	struct tag	*t;
+
+	if ((n = n->child) == NULL)
 		return 0;
 
-	assert(n->type == ROFFT_TEXT);
-
-	print_otag(h, TAG_A, "ch", "Lk", n->string);
-
-	if (NULL == n->next)
+	/* Link target and link text. */
+	t = print_otag(h, TAG_A, "cTh", "Lk", n->string);
+	if (n->next == NULL || n->next->flags & NODE_DELIMC)
 		print_text(h, n->string);
-
-	for (n = n->next; n; n = n->next)
+	for (n = n->next; n != NULL && !(n->flags & NODE_DELIMC); n = n->next)
 		print_text(h, n->string);
+	print_tagq(h, t);
 
+	/* Trailing punctuation. */
+	while (n != NULL) {
+		h->flags |= HTML_NOSPACE;
+		print_text(h, n->string);
+		n = n->next;
+	}
 	return 0;
 }
 
@@ -1345,7 +1382,7 @@ mdoc_mt_pre(MDOC_ARGS)
 		assert(n->type == ROFFT_TEXT);
 
 		mandoc_asprintf(&cp, "mailto:%s", n->string);
-		t = print_otag(h, TAG_A, "ch", "Mt", cp);
+		t = print_otag(h, TAG_A, "cTh", "Mt", cp);
 		print_text(h, n->string);
 		print_tagq(h, t);
 		free(cp);
@@ -1373,7 +1410,7 @@ mdoc_fo_pre(MDOC_ARGS)
 		return 0;
 
 	assert(n->child->string);
-	t = print_otag(h, TAG_B, "c", "Fn");
+	t = print_otag(h, TAG_B, "cT", "Fn");
 	print_text(h, n->child->string);
 	print_tagq(h, t);
 	return 0;
@@ -1397,7 +1434,7 @@ mdoc_in_pre(MDOC_ARGS)
 	struct tag	*t;
 
 	synopsis_pre(h, n);
-	print_otag(h, TAG_B, "c", "In");
+	print_otag(h, TAG_B, "cT", "In");
 
 	/*
 	 * The first argument of the `In' gets special treatment as
@@ -1416,9 +1453,9 @@ mdoc_in_pre(MDOC_ARGS)
 		assert(n->type == ROFFT_TEXT);
 
 		if (h->base_includes)
-			t = print_otag(h, TAG_A, "chI", "In", n->string);
+			t = print_otag(h, TAG_A, "cThI", "In", n->string);
 		else
-			t = print_otag(h, TAG_A, "c", "In");
+			t = print_otag(h, TAG_A, "cT", "In");
 		print_text(h, n->string);
 		print_tagq(h, t);
 
@@ -1439,14 +1476,19 @@ mdoc_in_pre(MDOC_ARGS)
 static int
 mdoc_ic_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_B, "c", "Ic");
+	char	*id;
+
+	if ((id = cond_id(n)) != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
+	print_otag(h, TAG_B, "cTi", "Ic", id);
+	free(id);
 	return 1;
 }
 
 static int
 mdoc_va_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_VAR, "c", "Va");
+	print_otag(h, TAG_VAR, "cT", "Va");
 	return 1;
 }
 
@@ -1491,7 +1533,12 @@ mdoc_bf_pre(MDOC_ARGS)
 static int
 mdoc_ms_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_B, "c", "Ms");
+	char *id;
+
+	if ((id = cond_id(n)) != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
+	print_otag(h, TAG_B, "cTi", "Ms", id);
+	free(id);
 	return 1;
 }
 
@@ -1520,28 +1567,38 @@ mdoc_rs_pre(MDOC_ARGS)
 	if (n->prev && SEC_SEE_ALSO == n->sec)
 		print_paragraph(h);
 
-	print_otag(h, TAG_CITE, "c", "Rs");
+	print_otag(h, TAG_CITE, "cT", "Rs");
 	return 1;
 }
 
 static int
 mdoc_no_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_SPAN, "c", "No");
+	char *id;
+
+	if ((id = cond_id(n)) != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
+	print_otag(h, TAG_SPAN, "ci", "No", id);
+	free(id);
 	return 1;
 }
 
 static int
 mdoc_li_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_CODE, "c", "Li");
+	char	*id;
+
+	if ((id = cond_id(n)) != NULL)
+		print_otag(h, TAG_A, "chR", "selflink", id);
+	print_otag(h, TAG_CODE, "ci", "Li", id);
+	free(id);
 	return 1;
 }
 
 static int
 mdoc_sy_pre(MDOC_ARGS)
 {
-	print_otag(h, TAG_B, "c", "Sy");
+	print_otag(h, TAG_B, "cT", "Sy");
 	return 1;
 }
 
@@ -1551,7 +1608,7 @@ mdoc_lb_pre(MDOC_ARGS)
 	if (SEC_LIBRARY == n->sec && NODE_LINE & n->flags && n->prev)
 		print_otag(h, TAG_BR, "");
 
-	print_otag(h, TAG_SPAN, "c", "Lb");
+	print_otag(h, TAG_SPAN, "cT", "Lb");
 	return 1;
 }