Added "Input Encoding" section to mdoc*.1.VERSION_1_4_4

Newline-escaped lines are concatenated.
Added -fign-macro option.

7 files changed, 123 insertions, 30 deletions

diff --git a/Makefile b/Makefile
index d1c01c92..ccd831df 100644
--- a/Makefile
+++ b/Makefile
@@ -1,7 +1,7 @@
 .SUFFIXES:	.html .sgml
 
-VERSION	= 1.4.2
-VDATE	= 8 March 2009
+VERSION	= 1.4.4
+VDATE	= 9 March 2009
 
 BINDIR		= $(PREFIX)/bin
 INCLUDEDIR	= $(PREFIX)/include
diff --git a/mdoc.c b/mdoc.c
index 7525bc41..99262ec6 100644
--- a/mdoc.c
+++ b/mdoc.c
@@ -1,4 +1,4 @@
-/* $Id: mdoc.c,v 1.57 2009/03/08 20:57:35 kristaps Exp $ */
+/* $Id: mdoc.c,v 1.58 2009/03/09 13:04:01 kristaps Exp $ */
 /*
  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -38,6 +38,7 @@ static	int		  mdoc_node_append(struct mdoc *,
 
 static	int		  parsetext(struct mdoc *, int, char *);
 static	int		  parsemacro(struct mdoc *, int, char *);
+static	int		  macrowarn(struct mdoc *, int, const char *);
 
 
 const	char *const __mdoc_macronames[MDOC_MAX] = {		 
@@ -495,6 +496,19 @@ parsetext(struct mdoc *mdoc, int line, char *buf)
 }
 
 
+static int
+macrowarn(struct mdoc *m, int ln, const char *buf)
+{
+	if ( ! (MDOC_IGN_MACRO & m->pflags))
+		return(mdoc_perr(m, ln, 1, "unknown macro: %s%s", 
+				buf, buf[3] ? "..." : ""));
+	return(mdoc_pwarn(m, ln, 1, WARN_SYNTAX,
+				"unknown macro: %s%s",
+				buf, buf[3] ? "..." : ""));
+}
+
+
+
 /*
  * Parse a macro line, that is, a line beginning with the control
  * character.
@@ -525,14 +539,15 @@ parsemacro(struct mdoc *m, int ln, char *buf)
 	mac[i - 1] = 0;
 
 	if (i == 5 || i <= 2) {
-		(void)mdoc_perr(m, ln, 1, "unknown macro: %s%s", 
-				mac, i == 5 ? "..." : "");
-		goto err;
+		if ( ! macrowarn(m, ln, mac))
+			goto err;
+		return(1);
 	} 
 	
 	if (MDOC_MAX == (c = mdoc_tokhash_find(m->htab, mac))) {
-		(void)mdoc_perr(m, ln, 1, "unknown macro: %s", mac);
-		goto err;
+		if ( ! macrowarn(m, ln, mac))
+			goto err;
+		return(1);
 	}
 
 	/* The macro is sane.  Jump to the next word. */
diff --git a/mdoc.h b/mdoc.h
index fd946d6e..9144150e 100644
--- a/mdoc.h
+++ b/mdoc.h
@@ -1,4 +1,4 @@
-/* $Id: mdoc.h,v 1.42 2009/03/08 20:57:35 kristaps Exp $ */
+/* $Id: mdoc.h,v 1.43 2009/03/09 13:04:01 kristaps Exp $ */
 /*
  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -264,6 +264,7 @@ struct	mdoc_node {
 
 #define	MDOC_IGN_SCOPE	 (1 << 0) /* Ignore scope violations. */
 #define	MDOC_IGN_ESCAPE	 (1 << 1) /* Ignore bad escape sequences. */
+#define	MDOC_IGN_MACRO	 (1 << 2) /* Ignore unknown macros. */
 
 /* Call-backs for parse messages. */
 struct	mdoc_cb {
diff --git a/mdoclint.1 b/mdoclint.1
index 14cfe750..1eefbc16 100644
--- a/mdoclint.1
+++ b/mdoclint.1
@@ -1,4 +1,4 @@
-.\" $Id: mdoclint.1,v 1.2 2009/03/08 18:02:36 kristaps Exp $
+.\" $Id: mdoclint.1,v 1.3 2009/03/09 13:04:01 kristaps Exp $
 .\"
 .\" Copyright (c) 2009 Kristaps Dzonsons <kristaps@kth.se>
 .\"
@@ -16,7 +16,7 @@
 .\" TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
 .\" PERFORMANCE OF THIS SOFTWARE.
 .\"
-.Dd $Mdocdate: March 8 2009 $
+.Dd $Mdocdate: March 9 2009 $
 .Dt mdoclint 1
 .Os
 .\" SECTION
@@ -100,6 +100,8 @@ When rewinding the scope of a block macro, forces the compiler to ignore
 scope violations.  This can seriously mangle the resulting tree.
 .It Fl f Ns Ar ign-escape
 Ignore invalid escape sequences.
+.It Fl f Ns Ar ign-macro
+Ignore unknown macros at the start of input lines.
 .El
 .\" PARAGRAPH
 .Pp
@@ -110,6 +112,21 @@ flag, multiple
 options may be grouped and delimited with a comma.  Using
 .Fl f Ns Ar ign-scope,ign-escape ,
 for example, will try to ignore scope and character-escape errors.
+.\" SUB-SECTION
+.Ss Input Encoding
+The
+.Nm
+utility expects its input to be 7-bit ASCII as defined in
+.Xr ascii 7 .
+The only non-graphing characters accepted are spaces,
+.Sq \  ,
+and tabs,
+.Sq \et .
+Tabs are only accepted in literal block-displays and as column
+delimiters.
+.Pp
+Only Unix-style newlines (\en) are accepted; if the newline is escaped,
+the line is concatenated with the next.
 .\" SECTION
 .Sh EXAMPLES
 To validate this manual page:
diff --git a/mdocterm.1 b/mdocterm.1
index 3725440c..d0c82113 100644
--- a/mdocterm.1
+++ b/mdocterm.1
@@ -1,4 +1,4 @@
-.\" $Id: mdocterm.1,v 1.15 2009/03/08 18:02:36 kristaps Exp $
+.\" $Id: mdocterm.1,v 1.16 2009/03/09 13:04:01 kristaps Exp $
 .\"
 .\" Copyright (c) 2009 Kristaps Dzonsons <kristaps@kth.se>
 .\"
@@ -16,7 +16,7 @@
 .\" TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
 .\" PERFORMANCE OF THIS SOFTWARE.
 .\"
-.Dd $Mdocdate: March 8 2009 $
+.Dd $Mdocdate: March 9 2009 $
 .Dt mdocterm 1
 .Os
 .\" SECTION
@@ -106,6 +106,8 @@ When rewinding the scope of a block macro, forces the compiler to ignore
 scope violations.  This can seriously mangle the resulting tree.
 .It Fl f Ns Ar ign-escape
 Ignore invalid escape sequences.
+.It Fl f Ns Ar ign-macro
+Ignore unknown macros at the start of input lines.
 .El
 .\" PARAGRAPH
 .Pp
@@ -117,6 +119,21 @@ options may be grouped and delimited with a comma.  Using
 .Fl f Ns Ar ign-scope,ign-escape ,
 for example, will try to ignore scope and character-escape errors.
 .\" SUB-SECTION
+.Ss Input Encoding
+The
+.Nm
+utility expects its input to be 7-bit ASCII as defined in
+.Xr ascii 7 .
+The only non-graphing characters accepted are spaces,
+.Sq \  ,
+and tabs,
+.Sq \et .
+Tabs are only accepted in literal block-displays and as column
+delimiters.
+.Pp
+Only Unix-style newlines (\en) are accepted; if the newline is escaped,
+the line is concatenated with the next.
+.\" SUB-SECTION
 .Ss Character Escapes
 This section documents the character-escapes accepted by
 .Xr mdocterm 1 .
diff --git a/mdoctree.1 b/mdoctree.1
index 05e93570..0894b2ba 100644
--- a/mdoctree.1
+++ b/mdoctree.1
@@ -1,4 +1,4 @@
-.\" $Id: mdoctree.1,v 1.5 2009/03/08 18:02:36 kristaps Exp $
+.\" $Id: mdoctree.1,v 1.6 2009/03/09 13:04:01 kristaps Exp $
 .\"
 .\" Copyright (c) 2009 Kristaps Dzonsons <kristaps@kth.se>
 .\"
@@ -16,7 +16,7 @@
 .\" TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
 .\" PERFORMANCE OF THIS SOFTWARE.
 .\"
-.Dd $Mdocdate: March 8 2009 $
+.Dd $Mdocdate: March 9 2009 $
 .Dt mdoctree 1
 .Os
 .\" SECTION
@@ -103,6 +103,8 @@ When rewinding the scope of a block macro, forces the compiler to ignore
 scope violations.  This can seriously mangle the resulting tree.
 .It Fl f Ns Ar ign-escape
 Ignore invalid escape sequences.
+.It Fl f Ns Ar ign-macro
+Ignore unknown macros at the start of input lines.
 .El
 .\" PARAGRAPH
 .Pp
@@ -113,6 +115,27 @@ flag, multiple
 options may be grouped and delimited with a comma.  Using
 .Fl f Ns Ar ign-scope,ign-escape ,
 for example, will try to ignore scope and character-escape errors.
+.\" SUB-SECTION
+.Ss Input Encoding
+The
+.Nm
+utility expects its input to be 7-bit ASCII as defined in
+.Xr ascii 7 .
+The only non-graphing characters accepted are spaces,
+.Sq \  ,
+and tabs,
+.Sq \et .
+Tabs are only accepted in literal block-displays and as column
+delimiters.
+.Pp
+Only Unix-style newlines (\en) are accepted; if the newline is escaped,
+the line is concatenated with the next.
+.\" SUB-SECTION
+.Ss Character Escapes
+Since
+.Nm
+doesn't format its output, character escapes are displayed as passed
+into the compiler. 
 .\" SECTION
 .Sh EXAMPLES
 To validate this manual page:
diff --git a/mmain.c b/mmain.c
index 52ff9a0a..ccadcaeb 100644
--- a/mmain.c
+++ b/mmain.c
@@ -1,4 +1,4 @@
-	/* $Id: mmain.c,v 1.9 2009/03/08 19:32:03 kristaps Exp $ */
+	/* $Id: mmain.c,v 1.10 2009/03/09 13:04:01 kristaps Exp $ */
 /*
  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -217,7 +217,8 @@ static int
 optsopt(struct mmain *p, char *arg)
 {
 	char		*v;
-	char		*toks[] = { "ign-scope", "ign-escape", NULL };
+	char		*toks[] = { "ign-scope", "ign-escape", 
+				    "ign-macro", NULL };
 
 	while (*arg) 
 		switch (getsubopt(&arg, toks, &v)) {
@@ -227,8 +228,11 @@ optsopt(struct mmain *p, char *arg)
 		case (1):
 			p->pflags |= MDOC_IGN_ESCAPE;
 			break;
+		case (2):
+			p->pflags |= MDOC_IGN_MACRO;
+			break;
 		default:
-			/* FIXME: report? */
+			warnx("unknown -f argument");
 			return(0);
 		}
 
@@ -258,7 +262,7 @@ optswarn(struct mmain *p, char *arg)
 			p->warn |= MD_WARN_ERR;
 			break;
 		default:
-			/* FIXME: report? */
+			warnx("unknown -W argument");
 			return(0);
 		}
 
@@ -270,10 +274,10 @@ static int
 parse(struct mmain *p)
 {
 	ssize_t		 sz;
-	int		 i, pos, len, lnn;
-	char		*line;
+	int		 j, i, pos, len, lnn;
+	char		*ln;
 
-	for (line = NULL, lnn = 1, len = pos = 0; ; ) {
+	for (ln = NULL, lnn = 1, len = pos = 0; ; ) {
 		if (-1 == (sz = read(p->fdin, p->buf, p->bufsz))) {
 			warn("%s", p->in);
 			return(0);
@@ -283,27 +287,43 @@ parse(struct mmain *p)
 		for (i = 0; i < (int)sz; i++) {
 			if (pos >= len) {
 				len += MD_LINE_SZ;
-				line = realloc(line, (size_t)len);
-				if (NULL == line)
+				ln = realloc(ln, (size_t)len);
+				if (NULL == ln)
 					err(1, "realloc");
 			}
 
 			if ('\n' != p->buf[i]) {
-				line[pos++] = p->buf[i];
+				ln[pos++] = p->buf[i];
 				continue;
 			}
 
-			line[pos] = 0;
-			if ( ! mdoc_parseln(p->mdoc, lnn, line))
-				return(0);
+			/* Check for escaped newline.  */
+
+			if (pos > 0 && '\\' == ln[pos - 1]) {
+				for (j = pos - 1; j >= 0; j--)
+					if ('\\' != ln[j])
+						break;
 
+				if ( ! ((pos - j) % 2)) {
+					pos--;
+					lnn++;
+					continue;
+				}
+			}
+
+			ln[pos] = 0;
+			if ( ! mdoc_parseln(p->mdoc, lnn, ln))
+				return(0);
 			lnn++;
 			pos = 0;
 		}
 	}
 
-	if (line)
-		free(line);
+	if (pos > 0)
+		warnx("%s: file not eof-terminated", p->in);
+
+	if (ln)
+		free(ln);
 	return(mdoc_endparse(p->mdoc));
 }