Bring the handling of defective prologues even closer to groff,

in particular relaxing the distinction between prologue and body
and further improving messages.
* The last .Dd wins and the last .Os wins, even in the body.
* The last .Dt before the first body macro wins.
* Missing title in .Dt defaults to UNTITLED.  Warn about it.
* Missing section in .Dt does not default to 1.  But warn about it.
* Do not warn multiple times about the same mdoc(7) prologue macro.
* Warn about missing .Os.
* Incomplete .TH defaults to empty strings.  Warn about it.

9 files changed, 137 insertions, 142 deletions

diff --git a/man_validate.c b/man_validate.c
index 2091840d..76aebe1c 100644
--- a/man_validate.c
+++ b/man_validate.c
@@ -43,7 +43,6 @@ typedef	int	(*v_check)(CHKARGS);
 static	int	  check_eq0(CHKARGS);
 static	int	  check_eq2(CHKARGS);
 static	int	  check_le1(CHKARGS);
-static	int	  check_ge2(CHKARGS);
 static	int	  check_le5(CHKARGS);
 static	int	  check_par(CHKARGS);
 static	int	  check_part(CHKARGS);
@@ -142,7 +141,7 @@ check_root(CHKARGS)
 		man->meta.hasbody = 1;
 
 	if (NULL == man->meta.title) {
-		mandoc_msg(MANDOCERR_TH_MISSING, man->parse,
+		mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 		    n->line, n->pos, NULL);
 
 		/*
@@ -150,8 +149,8 @@ check_root(CHKARGS)
 		 * implication, date and section also aren't set).
 		 */
 
-		man->meta.title = mandoc_strdup("unknown");
-		man->meta.msec = mandoc_strdup("1");
+		man->meta.title = mandoc_strdup("");
+		man->meta.msec = mandoc_strdup("");
 		man->meta.date = man->quick ? mandoc_strdup("") :
 		    mandoc_normdate(man->parse, NULL, n->line, n->pos);
 	}
@@ -189,7 +188,6 @@ check_##name(CHKARGS) \
 INEQ_DEFINE(0, ==, eq0)
 INEQ_DEFINE(2, ==, eq2)
 INEQ_DEFINE(1, <=, le1)
-INEQ_DEFINE(2, >=, ge2)
 INEQ_DEFINE(5, <=, le5)
 
 static int
@@ -324,7 +322,6 @@ post_TH(CHKARGS)
 	struct man_node	*nb;
 	const char	*p;
 
-	check_ge2(man, n);
 	check_le5(man, n);
 
 	free(man->meta.title);
@@ -354,8 +351,11 @@ post_TH(CHKARGS)
 			}
 		}
 		man->meta.title = mandoc_strdup(n->string);
-	} else
+	} else {
 		man->meta.title = mandoc_strdup("");
+		mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
+		    nb->line, nb->pos, "TH");
+	}
 
 	/* TITLE ->MSEC<- DATE SOURCE VOL */
 
@@ -363,8 +363,11 @@ post_TH(CHKARGS)
 		n = n->next;
 	if (n && n->string)
 		man->meta.msec = mandoc_strdup(n->string);
-	else
+	else {
 		man->meta.msec = mandoc_strdup("");
+		mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
+		    nb->line, nb->pos, "TH %s", man->meta.title);
+	}
 
 	/* TITLE MSEC ->DATE<- SOURCE VOL */
 
diff --git a/mandoc.h b/mandoc.h
index 136f1070..1ef88d35 100644
--- a/mandoc.h
+++ b/mandoc.h
@@ -49,16 +49,19 @@ enum	mandocerr {
 	MANDOCERR_WARNING, /* ===== start of warnings ===== */
 
 	/* related to the prologue */
-	MANDOCERR_TH_MISSING, /* missing .TH macro, using "unknown 1" */
+	MANDOCERR_DT_NOTITLE, /* missing manual title, using UNTITLED: line */
+	MANDOCERR_TH_NOTITLE, /* missing manual title, using "": [macro] */
 	MANDOCERR_TITLE_CASE, /* lower case character in document title */
+	MANDOCERR_MSEC_MISSING, /* missing manual section, using "": macro */
 	MANDOCERR_MSEC_BAD, /* unknown manual section: Dt ... section */
 	MANDOCERR_ARCH_BAD, /* unknown manual volume or arch: Dt ... volume */
 	MANDOCERR_DATE_MISSING, /* missing date, using today's date */
 	MANDOCERR_DATE_BAD, /* cannot parse date, using it verbatim: date */
-	MANDOCERR_PROLOG_ORDER, /* prologue macros out of order: macro */
+	MANDOCERR_OS_MISSING, /* missing Os macro, using "" */
 	MANDOCERR_PROLOG_REP, /* duplicate prologue macro: macro */
-	MANDOCERR_PROLOG_BAD, /* incomplete prologue, terminated by: macro */
-	MANDOCERR_PROLOG_ONLY, /* skipping prologue macro in body: macro */
+	MANDOCERR_PROLOG_LATE, /* late prologue macro: macro */
+	MANDOCERR_DT_LATE, /* skipping late title macro: Dt args */
+	MANDOCERR_PROLOG_ORDER, /* prologue macros out of order: macros */
 
 	/* related to document structure */
 	MANDOCERR_SO, /* .so is fragile, better use ln(1): so path */
diff --git a/mandocdb.c b/mandocdb.c
index a2f1c884..4dcb5ac6 100644
--- a/mandocdb.c
+++ b/mandocdb.c
@@ -1198,8 +1198,9 @@ mpages_merge(struct mchars *mc, struct mparse *mp)
 			goto nextpage;
 		} else if (NULL != mdoc) {
 			mpage->form = FORM_SRC;
-			mpage->sec =
-			    mandoc_strdup(mdoc_meta(mdoc)->msec);
+			mpage->sec = mdoc_meta(mdoc)->msec;
+			mpage->sec = mandoc_strdup(
+			    NULL == mpage->sec ? "" : mpage->sec);
 			mpage->arch = mdoc_meta(mdoc)->arch;
 			mpage->arch = mandoc_strdup(
 			    NULL == mpage->arch ? "" : mpage->arch);
diff --git a/mdoc.c b/mdoc.c
index ba5bad7c..fe63c930 100644
--- a/mdoc.c
+++ b/mdoc.c
@@ -123,20 +123,13 @@ mdoc_free1(struct mdoc *mdoc)
 
 	if (mdoc->first)
 		mdoc_node_delete(mdoc, mdoc->first);
-	if (mdoc->meta.title)
-		free(mdoc->meta.title);
-	if (mdoc->meta.os)
-		free(mdoc->meta.os);
-	if (mdoc->meta.name)
-		free(mdoc->meta.name);
-	if (mdoc->meta.arch)
-		free(mdoc->meta.arch);
-	if (mdoc->meta.vol)
-		free(mdoc->meta.vol);
-	if (mdoc->meta.msec)
-		free(mdoc->meta.msec);
-	if (mdoc->meta.date)
-		free(mdoc->meta.date);
+	free(mdoc->meta.msec);
+	free(mdoc->meta.vol);
+	free(mdoc->meta.arch);
+	free(mdoc->meta.date);
+	free(mdoc->meta.title);
+	free(mdoc->meta.os);
+	free(mdoc->meta.name);
 }
 
 /*
@@ -271,32 +264,22 @@ mdoc_macro(MACRO_PROT_ARGS)
 {
 	assert(tok < MDOC_MAX);
 
-	/* If we're in the body, deny prologue calls. */
-
-	if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
-	    MDOC_PBODY & mdoc->flags) {
-		mandoc_vmsg(MANDOCERR_PROLOG_ONLY, mdoc->parse,
-		    line, ppos, "%s", mdoc_macronames[tok]);
-		return(1);
-	}
-
-	/* If we're in the prologue, deny "body" macros.  */
-
-	if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
-	     ! (MDOC_PBODY & mdoc->flags)) {
-		mandoc_vmsg(MANDOCERR_PROLOG_BAD, mdoc->parse,
-		    line, ppos, "%s", mdoc_macronames[tok]);
-		if (NULL == mdoc->meta.msec)
-			mdoc->meta.msec = mandoc_strdup("1");
-		if (NULL == mdoc->meta.title)
-			mdoc->meta.title = mandoc_strdup("UNKNOWN");
+	if (mdoc->flags & MDOC_PBODY) {
+		if (tok == MDOC_Dt) {
+			mandoc_vmsg(MANDOCERR_DT_LATE,
+			    mdoc->parse, line, ppos,
+			    "Dt %s", buf + *pos);
+			return(1);
+		}
+	} else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) {
+		if (mdoc->meta.title == NULL) {
+			mandoc_vmsg(MANDOCERR_DT_NOTITLE,
+			    mdoc->parse, line, ppos, "%s %s",
+			    mdoc_macronames[tok], buf + *pos);
+			mdoc->meta.title = mandoc_strdup("UNTITLED");
+		}
 		if (NULL == mdoc->meta.vol)
 			mdoc->meta.vol = mandoc_strdup("LOCAL");
-		if (NULL == mdoc->meta.os)
-			mdoc->meta.os = mandoc_strdup("LOCAL");
-		if (NULL == mdoc->meta.date)
-			mdoc->meta.date = mandoc_normdate
-				(mdoc->parse, NULL, line, ppos);
 		mdoc->flags |= MDOC_PBODY;
 	}
 
diff --git a/mdoc_html.c b/mdoc_html.c
index 43ff1d27..328cb9bc 100644
--- a/mdoc_html.c
+++ b/mdoc_html.c
@@ -380,8 +380,9 @@ print_mdoc_head(MDOC_ARGS)
 
 	print_gen_head(h);
 	bufinit(h);
-	bufcat_fmt(h, "%s(%s)", meta->title, meta->msec);
-
+	bufcat(h, meta->title);
+	if (meta->msec)
+		bufcat_fmt(h, "(%s)", meta->msec);
 	if (meta->arch)
 		bufcat_fmt(h, " (%s)", meta->arch);
 
@@ -525,7 +526,11 @@ mdoc_root_pre(MDOC_ARGS)
 		mandoc_asprintf(&volume, "%s (%s)",
 		    meta->vol, meta->arch);
 
-	mandoc_asprintf(&title, "%s(%s)", meta->title, meta->msec);
+	if (NULL == meta->msec)
+		title = mandoc_strdup(meta->title);
+	else
+		mandoc_asprintf(&title, "%s(%s)",
+		    meta->title, meta->msec);
 
 	PAIR_SUMMARY_INIT(&tag[0], "Document Header");
 	PAIR_CLASS_INIT(&tag[1], "head");
diff --git a/mdoc_man.c b/mdoc_man.c
index 504790b9..070372cf 100644
--- a/mdoc_man.c
+++ b/mdoc_man.c
@@ -548,8 +548,9 @@ man_mdoc(void *arg, const struct mdoc *mdoc)
 	n = mdoc_node(mdoc);
 
 	printf(".TH \"%s\" \"%s\" \"%s\" \"%s\" \"%s\"\n",
-	    meta->title, meta->msec, meta->date,
-	    meta->os, meta->vol);
+	    meta->title,
+	    (meta->msec == NULL ? "" : meta->msec),
+	    meta->date, meta->os, meta->vol);
 
 	/* Disable hyphenation and if nroff, disable justification. */
 	printf(".nh\n.if n .ad l");
diff --git a/mdoc_term.c b/mdoc_term.c
index ea34908c..9cb6e8e0 100644
--- a/mdoc_term.c
+++ b/mdoc_term.c
@@ -476,7 +476,11 @@ print_mdoc_head(struct termp *p, const void *arg)
 		    meta->vol, meta->arch);
 	vollen = term_strlen(p, volume);
 
-	mandoc_asprintf(&title, "%s(%s)", meta->title, meta->msec);
+	if (NULL == meta->msec)
+		title = mandoc_strdup(meta->title);
+	else
+		mandoc_asprintf(&title, "%s(%s)",
+		    meta->title, meta->msec);
 	titlen = term_strlen(p, title);
 
 	p->flags |= TERMP_NOBREAK | TERMP_NOSPACE;
diff --git a/mdoc_validate.c b/mdoc_validate.c
index 47dd6c39..a7fd30bf 100644
--- a/mdoc_validate.c
+++ b/mdoc_validate.c
@@ -114,7 +114,6 @@ static	int	 post_nm(POST_ARGS);
 static	int	 post_ns(POST_ARGS);
 static	int	 post_os(POST_ARGS);
 static	int	 post_par(POST_ARGS);
-static	int	 post_prol(POST_ARGS);
 static	int	 post_root(POST_ARGS);
 static	int	 post_rs(POST_ARGS);
 static	int	 post_sh(POST_ARGS);
@@ -843,14 +842,12 @@ static int
 pre_dt(PRE_ARGS)
 {
 
-	if (NULL == mdoc->meta.date || mdoc->meta.os)
-		mandoc_msg(MANDOCERR_PROLOG_ORDER, mdoc->parse,
-		    n->line, n->pos, "Dt");
-
-	if (mdoc->meta.title)
+	if (mdoc->meta.title != NULL)
 		mandoc_msg(MANDOCERR_PROLOG_REP, mdoc->parse,
 		    n->line, n->pos, "Dt");
-
+	else if (mdoc->meta.os != NULL)
+		mandoc_msg(MANDOCERR_PROLOG_ORDER, mdoc->parse,
+		    n->line, n->pos, "Dt after Os");
 	return(1);
 }
 
@@ -858,14 +855,12 @@ static int
 pre_os(PRE_ARGS)
 {
 
-	if (NULL == mdoc->meta.title || NULL == mdoc->meta.date)
-		mandoc_msg(MANDOCERR_PROLOG_ORDER, mdoc->parse,
-		    n->line, n->pos, "Os");
-
-	if (mdoc->meta.os)
+	if (mdoc->meta.os != NULL)
 		mandoc_msg(MANDOCERR_PROLOG_REP, mdoc->parse,
 		    n->line, n->pos, "Os");
-
+	else if (mdoc->flags & MDOC_PBODY)
+		mandoc_msg(MANDOCERR_PROLOG_LATE, mdoc->parse,
+		    n->line, n->pos, "Os");
 	return(1);
 }
 
@@ -873,14 +868,18 @@ static int
 pre_dd(PRE_ARGS)
 {
 
-	if (mdoc->meta.title || mdoc->meta.os)
-		mandoc_msg(MANDOCERR_PROLOG_ORDER, mdoc->parse,
-		    n->line, n->pos, "Dd");
-
-	if (mdoc->meta.date)
+	if (mdoc->meta.date != NULL)
 		mandoc_msg(MANDOCERR_PROLOG_REP, mdoc->parse,
 		    n->line, n->pos, "Dd");
-
+	else if (mdoc->flags & MDOC_PBODY)
+		mandoc_msg(MANDOCERR_PROLOG_LATE, mdoc->parse,
+		    n->line, n->pos, "Dd");
+	else if (mdoc->meta.title != NULL)
+		mandoc_msg(MANDOCERR_PROLOG_ORDER, mdoc->parse,
+		    n->line, n->pos, "Dd after Dt");
+	else if (mdoc->meta.os != NULL)
+		mandoc_msg(MANDOCERR_PROLOG_ORDER, mdoc->parse,
+		    n->line, n->pos, "Dd after Os");
 	return(1);
 }
 
@@ -1627,20 +1626,24 @@ post_root(POST_ARGS)
 
 	/* Add missing prologue data. */
 
-	if ( ! (MDOC_PBODY & mdoc->flags)) {
-		mandoc_msg(MANDOCERR_PROLOG_BAD, mdoc->parse, 0, 0, "EOF");
-		if (mdoc->meta.date == NULL)
-			mdoc->meta.date = mdoc->quick ?
-			    mandoc_strdup("") :
-			    mandoc_normdate(mdoc->parse, NULL, 0, 0);
-		if (mdoc->meta.title == NULL)
-			mdoc->meta.title = mandoc_strdup("UNKNOWN");
-		if (mdoc->meta.vol == NULL)
-			mdoc->meta.vol = mandoc_strdup("LOCAL");
-		if (mdoc->meta.arch == NULL)
-			mdoc->meta.msec = mandoc_strdup("1");
-		if (mdoc->meta.os == NULL)
-			mdoc->meta.os = mandoc_strdup("UNKNOWN");
+	if (mdoc->meta.date == NULL)
+		mdoc->meta.date = mdoc->quick ?
+		    mandoc_strdup("") :
+		    mandoc_normdate(mdoc->parse, NULL, 0, 0);
+
+	if (mdoc->meta.title == NULL) {
+		mandoc_msg(MANDOCERR_DT_NOTITLE,
+		    mdoc->parse, 0, 0, "EOF");
+		mdoc->meta.title = mandoc_strdup("UNTITLED");
+	}
+
+	if (mdoc->meta.vol == NULL)
+		mdoc->meta.vol = mandoc_strdup("LOCAL");
+
+	if (mdoc->meta.os == NULL) {
+		mandoc_msg(MANDOCERR_OS_MISSING,
+		    mdoc->parse, 0, 0, NULL);
+		mdoc->meta.os = mandoc_strdup("");
 	}
 
 	n = mdoc->first;
@@ -1977,7 +1980,10 @@ post_sh_head(POST_ARGS)
 
 	/* Check particular section/manual conventions. */
 
-	assert(mdoc->meta.msec);
+	if (mdoc->meta.msec == NULL) {
+		free(secname);
+		return(1);
+	}
 
 	goodsec = NULL;
 	switch (sec) {
@@ -2156,7 +2162,7 @@ post_dd(POST_ARGS)
 	if (NULL == n->child || '\0' == n->child->string[0]) {
 		mdoc->meta.date = mdoc->quick ? mandoc_strdup("") :
 		    mandoc_normdate(mdoc->parse, NULL, n->line, n->pos);
-		return(post_prol(mdoc));
+		goto out;
 	}
 
 	datestr = NULL;
@@ -2168,7 +2174,9 @@ post_dd(POST_ARGS)
 		    datestr, n->line, n->pos);
 		free(datestr);
 	}
-	return(post_prol(mdoc));
+out:
+	mdoc_node_delete(mdoc, n);
+	return(1);
 }
 
 static int
@@ -2180,14 +2188,15 @@ post_dt(POST_ARGS)
 
 	n = mdoc->last;
 
-	if (mdoc->meta.title)
-		free(mdoc->meta.title);
-	if (mdoc->meta.vol)
-		free(mdoc->meta.vol);
-	if (mdoc->meta.arch)
-		free(mdoc->meta.arch);
+	free(mdoc->meta.title);
+	free(mdoc->meta.msec);
+	free(mdoc->meta.vol);
+	free(mdoc->meta.arch);
 
-	mdoc->meta.title = mdoc->meta.vol = mdoc->meta.arch = NULL;
+	mdoc->meta.title = NULL;
+	mdoc->meta.msec = NULL;
+	mdoc->meta.vol = NULL;
+	mdoc->meta.arch = NULL;
 
 	/* First check that all characters are uppercase. */
 
@@ -2202,32 +2211,27 @@ post_dt(POST_ARGS)
 			break;
 		}
 
-	/* Handles: `.Dt'
-	 * title = unknown, volume = local, msec = 0, arch = NULL
-	 */
+	/* No argument: msec and arch remain NULL. */
 
 	if (NULL == (nn = n->child)) {
-		/* XXX: make these macro values. */
-		/* FIXME: warn about missing values. */
-		mdoc->meta.title = mandoc_strdup("UNKNOWN");
+		mandoc_msg(MANDOCERR_DT_NOTITLE,
+		    mdoc->parse, n->line, n->pos, "Dt");
+		mdoc->meta.title = mandoc_strdup("UNTITLED");
 		mdoc->meta.vol = mandoc_strdup("LOCAL");
-		mdoc->meta.msec = mandoc_strdup("1");
-		return(post_prol(mdoc));
+		goto out;
 	}
 
-	/* Handles: `.Dt TITLE'
-	 * title = TITLE, volume = local, msec = 0, arch = NULL
-	 */
+	/* One argument: msec and arch remain NULL. */
 
 	mdoc->meta.title = mandoc_strdup(
-	    '\0' == nn->string[0] ? "UNKNOWN" : nn->string);
+	    '\0' == nn->string[0] ? "UNTITLED" : nn->string);
 
 	if (NULL == (nn = nn->next)) {
-		/* FIXME: warn about missing msec. */
-		/* XXX: make this a macro value. */
+		mandoc_vmsg(MANDOCERR_MSEC_MISSING,
+		    mdoc->parse, n->line, n->pos,
+		    "Dt %s", mdoc->meta.title);
 		mdoc->meta.vol = mandoc_strdup("LOCAL");
-		mdoc->meta.msec = mandoc_strdup("1");
-		return(post_prol(mdoc));
+		goto out;
 	}
 
 	/* Handles: `.Dt TITLE SEC'
@@ -2249,7 +2253,7 @@ post_dt(POST_ARGS)
 	}
 
 	if (NULL == (nn = nn->next))
-		return(post_prol(mdoc));
+		goto out;
 
 	/* Handles: `.Dt TITLE SEC VOL'
 	 * title = TITLE,
@@ -2275,23 +2279,8 @@ post_dt(POST_ARGS)
 
 	/* Ignore any subsequent parameters... */
 	/* FIXME: warn about subsequent parameters. */
-
-	return(post_prol(mdoc));
-}
-
-static int
-post_prol(POST_ARGS)
-{
-	/*
-	 * Remove prologue macros from the document after they're
-	 * processed.  The final document uses mdoc_meta for these
-	 * values and discards the originals.
-	 */
-
-	mdoc_node_delete(mdoc, mdoc->last);
-	if (mdoc->meta.title && mdoc->meta.date && mdoc->meta.os)
-		mdoc->flags |= MDOC_PBODY;
-
+out:
+	mdoc_node_delete(mdoc, n);
 	return(1);
 }
 
@@ -2337,11 +2326,11 @@ post_os(POST_ARGS)
 	mdoc->meta.os = NULL;
 	mdoc_deroff(&mdoc->meta.os, n);
 	if (mdoc->meta.os)
-		return(post_prol(mdoc));
+		goto out;
 
 	if (mdoc->defos) {
 		mdoc->meta.os = mandoc_strdup(mdoc->defos);
-		return(post_prol(mdoc));
+		goto out;
 	}
 
 #ifdef OSNAME
@@ -2358,7 +2347,10 @@ post_os(POST_ARGS)
 	}
 	mdoc->meta.os = mandoc_strdup(defbuf);
 #endif /*!OSNAME*/
-	return(post_prol(mdoc));
+
+out:
+	mdoc_node_delete(mdoc, n);
+	return(1);
 }
 
 /*
diff --git a/read.c b/read.c
index 07329dbe..4ad08236 100644
--- a/read.c
+++ b/read.c
@@ -93,16 +93,19 @@ static	const char * const	mandocerrs[MANDOCERR_MAX] = {
 	"generic warning",
 
 	/* related to the prologue */
-	"missing .TH macro, using \"unknown 1\"",
+	"missing manual title, using UNTITLED",
+	"missing manual title, using \"\"",
 	"lower case character in document title",
+	"missing manual section, using \"\"",
 	"unknown manual section",
 	"unknown manual volume or arch",
 	"missing date, using today's date",
 	"cannot parse date, using it verbatim",
-	"prologue macros out of order",
+	"missing Os macro, using \"\"",
 	"duplicate prologue macro",
-	"incomplete prologue, terminated by",
-	"skipping prologue macro in body",
+	"late prologue macro",
+	"skipping late title macro",
+	"prologue macros out of order",
 
 	/* related to document structure */
 	".so is fragile, better use ln(1)",