Move main format autodetection from the parser dispatcher to the

roff parser where .Dd and .TH are already detected, anyway.  This
improves robustness because it correctly handles whitespace or an
alternate control character before Dd.  In the parser dispatcher,
provide a fallback looking ahead in the input buffer instead of
always assuming man(7).  This corrects autodetection when Dd is
preceded by other macros or macro-like handled requests like .ll.

Triggered by reports from Daniel Levai about issues on Slackware Linux.

3 files changed, 63 insertions, 31 deletions

diff --git a/libmandoc.h b/libmandoc.h
index 1011cc50..6d629db3 100644
--- a/libmandoc.h
+++ b/libmandoc.h
@@ -1,4 +1,4 @@
-/*	$Id: libmandoc.h,v 1.42 2014/07/09 11:31:43 schwarze Exp $ */
+/*	$Id: libmandoc.h,v 1.43 2014/09/06 22:39:36 schwarze Exp $ */
 /*
  * Copyright (c) 2009, 2010, 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2013, 2014 Ingo Schwarze <schwarze@openbsd.org>
@@ -77,6 +77,7 @@ int		 roff_getreg(const struct roff *, const char *);
 char		*roff_strdup(const struct roff *, const char *);
 int		 roff_getcontrol(const struct roff *,
 			const char *, int *);
+int		 roff_getformat(const struct roff *);
 #if 0
 char		 roff_eqndelim(const struct roff *);
 void		 roff_openeqn(struct roff *, const char *,
diff --git a/read.c b/read.c
index 3b9ce545..26cd6288 100644
--- a/read.c
+++ b/read.c
@@ -1,4 +1,4 @@
-/*	$Id: read.c,v 1.82 2014/09/03 23:21:47 schwarze Exp $ */
+/*	$Id: read.c,v 1.83 2014/09/06 22:39:36 schwarze Exp $ */
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010-2014 Ingo Schwarze <schwarze@openbsd.org>
@@ -51,21 +51,22 @@ struct	buf {
 };
 
 struct	mparse {
-	enum mandoclevel  file_status; /* status of current parse */
-	enum mandoclevel  wlevel; /* ignore messages below this */
-	int		  line; /* line number in the file */
-	int		  options; /* parser options */
 	struct man	 *pman; /* persistent man parser */
 	struct mdoc	 *pmdoc; /* persistent mdoc parser */
 	struct man	 *man; /* man parser */
 	struct mdoc	 *mdoc; /* mdoc parser */
 	struct roff	 *roff; /* roff parser (!NULL) */
 	char		 *sodest; /* filename pointed to by .so */
-	int		  reparse_count; /* finite interp. stack */
-	mandocmsg	  mmsg; /* warning/error message handler */
-	const char	 *file;
-	struct buf	 *secondary;
+	const char	 *file; /* filename of current input file */
+	struct buf	 *primary; /* buffer currently being parsed */
+	struct buf	 *secondary; /* preprocessed copy of input */
 	const char	 *defos; /* default operating system */
+	mandocmsg	  mmsg; /* warning/error message handler */
+	enum mandoclevel  file_status; /* status of current parse */
+	enum mandoclevel  wlevel; /* ignore messages below this */
+	int		  options; /* parser options */
+	int		  reparse_count; /* finite interp. stack */
+	int		  line; /* line number in the file */
 };
 
 static	void	  resize_buf(struct buf *, size_t);
@@ -248,19 +249,10 @@ resize_buf(struct buf *buf, size_t initial)
 static void
 pset(const char *buf, int pos, struct mparse *curp)
 {
+	char		*cp, *ep;
+	int		 format;
 	int		 i;
 
-	/*
-	 * Try to intuit which kind of manual parser should be used.  If
-	 * passed in by command-line (-man, -mdoc), then use that
-	 * explicitly.  If passed as -mandoc, then try to guess from the
-	 * line: either skip dot-lines, use -mdoc when finding `.Dt', or
-	 * default to -man, which is more lenient.
-	 *
-	 * Separate out pmdoc/pman from mdoc/man: the first persists
-	 * through all parsers, while the latter is used per-parse.
-	 */
-
 	if ('.' == buf[0] || '\'' == buf[0]) {
 		for (i = 1; buf[i]; i++)
 			if (' ' != buf[i] && '\t' != buf[i])
@@ -269,15 +261,35 @@ pset(const char *buf, int pos, struct mparse *curp)
 			return;
 	}
 
-	if (MPARSE_MDOC & curp->options) {
-		curp->mdoc = curp->pmdoc;
-		return;
-	} else if (MPARSE_MAN & curp->options) {
-		curp->man = curp->pman;
-		return;
+	/*
+	 * If neither command line arguments -mdoc or -man select
+	 * a parser nor the roff parser found a .Dd or .TH macro
+	 * yet, look ahead in the main input buffer.
+	 */
+
+	if ((format = roff_getformat(curp->roff)) == 0) {
+		cp = curp->primary->buf;
+		ep = cp + curp->primary->sz;
+		while (cp < ep) {
+			if (*cp == '.' || *cp != '\'') {
+				cp++;
+				if (cp[0] == 'D' && cp[1] == 'd') {
+					format = MPARSE_MDOC;
+					break;
+				}
+				if (cp[0] == 'T' && cp[1] == 'H') {
+					format = MPARSE_MAN;
+					break;
+				}
+			}
+			cp = memchr(cp, '\n', ep - cp);
+			if (cp == NULL)
+				break;
+			cp++;
+		}
 	}
 
-	if (pos >= 3 && 0 == memcmp(buf, ".Dd", 3))  {
+	if (format == MPARSE_MDOC) {
 		if (NULL == curp->pmdoc)
 			curp->pmdoc = mdoc_alloc(
 			    curp->roff, curp, curp->defos,
@@ -287,6 +299,8 @@ pset(const char *buf, int pos, struct mparse *curp)
 		return;
 	}
 
+	/* Fall back to man(7) as a last resort. */
+
 	if (NULL == curp->pman)
 		curp->pman = man_alloc(curp->roff, curp,
 		    MPARSE_QUICK & curp->options ? 1 : 0);
@@ -720,6 +734,7 @@ mparse_parse_buffer(struct mparse *curp, struct buf blk, const char *file)
 	/* Line number is per-file. */
 	svfile = curp->file;
 	curp->file = file;
+	curp->primary = &blk;
 	curp->line = 1;
 	recursion_depth++;
 
diff --git a/roff.c b/roff.c
index 33361402..1d8a3991 100644
--- a/roff.c
+++ b/roff.c
@@ -1,4 +1,4 @@
-/*	$Id: roff.c,v 1.226 2014/08/19 16:52:32 schwarze Exp $ */
+/*	$Id: roff.c,v 1.227 2014/09/06 22:39:36 schwarze Exp $ */
 /*
  * Copyright (c) 2010, 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010-2014 Ingo Schwarze <schwarze@openbsd.org>
@@ -122,6 +122,7 @@ struct	roff {
 	int		 options; /* parse options */
 	int		 rstacksz; /* current size limit of rstack */
 	int		 rstackpos; /* position in rstack */
+	int		 format; /* current file in mdoc or man format */
 	char		 control; /* control character */
 };
 
@@ -456,6 +457,7 @@ roff_reset(struct roff *r)
 {
 
 	roff_free1(r);
+	r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
 	r->control = 0;
 }
 
@@ -475,6 +477,7 @@ roff_alloc(struct mparse *parse, int options)
 	r = mandoc_calloc(1, sizeof(struct roff));
 	r->parse = parse;
 	r->options = options;
+	r->format = options & (MPARSE_MDOC | MPARSE_MAN);
 	r->rstackpos = -1;
 
 	roffhash_init();
@@ -1776,10 +1779,13 @@ roff_Dd(ROFF_ARGS)
 {
 	const char *const	*cp;
 
-	if (0 == ((MPARSE_MDOC | MPARSE_QUICK) & r->options))
+	if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
 		for (cp = __mdoc_reserved; *cp; cp++)
 			roff_setstr(r, *cp, NULL, 0);
 
+	if (r->format == 0)
+		r->format = MPARSE_MDOC;
+
 	return(ROFF_CONT);
 }
 
@@ -1788,10 +1794,13 @@ roff_TH(ROFF_ARGS)
 {
 	const char *const	*cp;
 
-	if (0 == (MPARSE_QUICK & r->options))
+	if ((r->options & MPARSE_QUICK) == 0)
 		for (cp = __man_reserved; *cp; cp++)
 			roff_setstr(r, *cp, NULL, 0);
 
+	if (r->format == 0)
+		r->format = MPARSE_MAN;
+
 	return(ROFF_CONT);
 }
 
@@ -2307,6 +2316,13 @@ roff_strdup(const struct roff *r, const char *p)
 	return(res);
 }
 
+int
+roff_getformat(const struct roff *r)
+{
+
+	return(r->format);
+}
+
 /*
  * Find out whether a line is a macro line or not.
  * If it is, adjust the current position and return one; if it isn't,