Added colour styles (not being used) to struct termp.

Added nroff style-escape encoding.
Removed ANSI schema string tables (there's only ANSI and nroff/backspace).
Pushed styling directly into pword (simpler).

5 files changed, 152 insertions, 108 deletions

diff --git a/Makefile b/Makefile
index dae5bf57..bbc0be5a 100644
--- a/Makefile
+++ b/Makefile
@@ -1,6 +1,6 @@
 .SUFFIXES:	.html .sgml
 
-VERSION	= 1.4.15
+VERSION	= 1.4.16
 VDATE	= 14 March 2009
 
 BINDIR		= $(PREFIX)/bin
diff --git a/mdocterm.1 b/mdocterm.1
index 96ecd2d5..796fb8de 100644
--- a/mdocterm.1
+++ b/mdocterm.1
@@ -1,4 +1,4 @@
-.\" $Id: mdocterm.1,v 1.19 2009/03/13 07:46:10 kristaps Exp $
+.\" $Id: mdocterm.1,v 1.20 2009/03/14 12:35:02 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 13 2009 $
+.Dd $Mdocdate: March 14 2009 $
 .Dt mdocterm 1
 .Os
 .\" SECTION
@@ -28,6 +28,7 @@
 .Nm mdocmterm
 .Op Fl vV
 .Op Fl f Ns Ar option...
+.Op Fl O Ns Ar option...
 .Op Fl W Ns Ar err...
 .Op Ar infile
 .\" SECTION
@@ -50,6 +51,15 @@ Override default compiler behaviour.  See
 .Sx Compiler Options
 for details.
 .\" ITEM
+.It Fl O Ns Ar option...
+Terminal-encoding options.  May be set to 
+.Fl O Ns Ar ansi
+for ANSI/VT100-encoded output (the default) or
+.Fl O Ns Ar nroff
+for nroff-encoded 
+.Qq backspace
+output.
+.\" ITEM
 .It Fl W Ns Ar err...
 Print warning messages.  May be set to 
 .Fl W Ns Ar all
@@ -119,30 +129,11 @@ 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 .
-Note that the \\x, \\(xx and \\[n] forms are described here; the \\*(xx,
-\\*[n] and \\*x forms described in
-.Xr mdoc.samples 7
-are deprecated, but still rendered.  All one- and two-character
-sequences may be used in the n-character sequence \\[n].
-.Pp
+The following table shows all 
+.Xr mdoc 7
+character escapes rendered by
+.Nm .
 Note that the
 .Em Output
 column will render differently whether executed with
@@ -266,7 +257,7 @@ Special symbols:
 the more general syntax.
 .\" SECTION
 .Sh EXAMPLES
-To display this manual page:
+To display this manual page on ANSI-capable terminal:
 .\" PARAGRAPH
 .Pp
 .D1 % mdocmterm \-Wall,error mdocmterm.1 
@@ -274,7 +265,7 @@ To display this manual page:
 .Pp
 To pipe a manual page to the pager:
 .Pp
-.D1 % mdocterm mdocterm.1 2>&1 | less \-R
+.D1 % mdocterm -Onroff mdocterm.1 | less
 .\" SECTION
 .Sh SEE ALSO
 .Xr mdoctree 1 ,
diff --git a/mdocterm.c b/mdocterm.c
index 644bbe35..5308d449 100644
--- a/mdocterm.c
+++ b/mdocterm.c
@@ -1,4 +1,4 @@
-/* $Id: mdocterm.c,v 1.41 2009/03/14 05:36:07 kristaps Exp $ */
+/* $Id: mdocterm.c,v 1.42 2009/03/14 12:35:02 kristaps Exp $ */
 /*
  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -29,11 +29,12 @@
 #include "mmain.h"
 #include "term.h"
 
-struct	termenc {
+struct	termseq {
 	const char	 *enc;
 	int		  sym;
 };
 
+static	int		  option(void *, int, const char *);
 static	void		  body(struct termp *,
 				struct termpair *,
 				const struct mdoc_meta *,
@@ -53,14 +54,13 @@ static	void		  stringa(struct termp *,
 				const char *, size_t);
 static	void		  symbola(struct termp *, enum tsym);
 static	void		  sanity(const struct mdoc_node *);
-static	void		  stylea(struct termp *, enum tstyle);
 
 #ifdef __linux__
 extern	size_t		  strlcat(char *, const char *, size_t);
 extern	size_t		  strlcpy(char *, const char *, size_t);
 #endif
 
-static	struct termenc	  termenc1[] = {
+static	struct termseq	  termenc1[] = {
 	{ "\\",		  TERMSYM_SLASH },
 	{ "\'",		  TERMSYM_RSQUOTE },
 	{ "`",		  TERMSYM_LSQUOTE },
@@ -74,7 +74,7 @@ static	struct termenc	  termenc1[] = {
 	{ NULL,		  0 }
 };
 
-static	struct termenc	  termenc2[] = {
+static	struct termseq	  termenc2[] = {
 	{ "rC", 	  TERMSYM_RBRACE },
 	{ "lC", 	  TERMSYM_LBRACE },
 	{ "rB", 	  TERMSYM_RBRACK },
@@ -126,7 +126,7 @@ static	struct termenc	  termenc2[] = {
 	{ NULL,		  0 }
 };
 
-static	struct termsym	  termsym_ansi[TERMSYM_MAX] = {
+static	struct termsym	  termsym_ascii[TERMSYM_MAX] = {
 	{ "]", 1 },		/* TERMSYM_RBRACK */
 	{ "[", 1 },		/* TERMSYM_LBRACK */
 	{ "<-", 2 },		/* TERMSYM_LARROW */
@@ -170,16 +170,6 @@ static	struct termsym	  termsym_ansi[TERMSYM_MAX] = {
 	{ "}", 1 },		/* TERMSYM_RBRACE */
 };
 
-static	const char	  ansi_clear[]  = { 27, '[', '0', 'm' };
-static	const char	  ansi_bold[]  = { 27, '[', '1', 'm' };
-static	const char	  ansi_under[]  = { 27, '[', '4', 'm' };
-
-static	struct termsym	  termstyle_ansi[] = {
-	{ ansi_clear, 4 },
-	{ ansi_bold, 4 },
-	{ ansi_under, 4 }
-};
-
 
 int
 main(int argc, char *argv[])
@@ -189,22 +179,25 @@ main(int argc, char *argv[])
 	const struct mdoc *mdoc;
 	struct termp	   termp;
 
+	(void)memset(&termp, 0, sizeof(struct termp));
+
+	termp.maxrmargin = termp.rmargin = 78; /* XXX */
+	termp.maxcols = 1024;
+	termp.offset = termp.col = 0;
+	termp.flags = TERMP_NOSPACE;
+	termp.symtab = termsym_ascii;
+	termp.enc = TERMENC_ANSI;
+
 	p = mmain_alloc();
+	c = mmain_getopt(p, argc, argv, "[-Ooption...]", 
+			"O:", &termp, option);
 
-	c = mmain_getopt(p, argc, argv, NULL, NULL, NULL, NULL);
 	if (1 != c)
 		mmain_exit(p, -1 == c ? 1 : 0);
 
 	if (NULL == (mdoc = mmain_mdoc(p)))
 		mmain_exit(p, 1);
 
-	termp.maxrmargin = termp.rmargin = 78; /* XXX */
-	termp.maxcols = 1024;
-	termp.offset = termp.col = 0;
-	termp.flags = TERMP_NOSPACE;
-	termp.symtab = termsym_ansi;
-	termp.styletab = termstyle_ansi;
-
 	if (NULL == (termp.buf = malloc(termp.maxcols)))
 		err(1, "malloc");
 
@@ -219,6 +212,26 @@ main(int argc, char *argv[])
 }
 
 
+int
+option(void *ptr, int c, const char *arg)
+{
+	struct termp	*p;
+
+	p = (struct termp *)ptr;
+
+	if (0 == strcmp(arg, "nroff")) {
+		p->enc = TERMENC_NROFF;
+		return(1);
+	} else if (0 == strcmp(arg, "ansi")) {
+		p->enc = TERMENC_ANSI;
+		return(1);
+	}
+
+	warnx("unknown option: -O%s", arg);
+	return(0);
+}
+
+
 /*
  * Flush a line of text.  A "line" is loosely defined as being something
  * that should be followed by a newline, regardless of whether it's
@@ -288,15 +301,18 @@ flushln(struct termp *p)
 		 * space is printed according to regular spacing rules).
 		 */
 
-		/* FIXME: make non-ANSI friendly. */
-
 		/* LINTED */
 		for (j = i, vsz = 0; j < p->col; j++) {
-			if (isspace((u_char)p->buf[j]))
+			if (isspace((u_char)p->buf[j])) {
 				break;
-			else if (27 == p->buf[j]) {
-				assert(j + 4 <= p->col);
-				j += 3;
+			} else if (27 == p->buf[j]) {
+				assert(TERMENC_ANSI == p->enc);
+				assert(j + 5 <= p->col);
+				j += 4;
+			} else if (8 == p->buf[j]) {
+				assert(TERMENC_NROFF == p->enc);
+				assert(j + 2 <= p->col);
+				j += 1;
 			} else
 				vsz++;
 		}
@@ -656,7 +672,7 @@ header(struct termp *p, const struct mdoc_meta *meta)
 static void
 nescape(struct termp *p, const char *word, size_t len)
 {
-	struct termenc	*enc;
+	struct termseq	*enc;
 
 	switch (len) {
 	case (1):
@@ -763,26 +779,70 @@ pword(struct termp *p, const char *word, size_t len)
 		p->flags &= ~TERMP_NOSPACE;
 
 	/* 
-	 * XXX - if literal and underlining, this will underline the
-	 * spaces between literal words. 
+	 * If ANSI (word-length styling), then apply our style now,
+	 * before the word.
 	 */
 
-	if (p->flags & TERMP_BOLD)
-		stylea(p, TERMSTYLE_BOLD);
-	if (p->flags & TERMP_UNDERLINE)
-		stylea(p, TERMSTYLE_UNDER);
+	if (TERMENC_ANSI == p->enc && TERMP_STYLE & p->flags) {
+		if (TERMP_BOLD & p->flags) {
+			chara(p, 27);
+			stringa(p, "[01m", 4);
+		}
+		if (TERMP_UNDER & p->flags) {
+			chara(p, 27);
+			stringa(p, "[04m", 4);
+		}
+		if (TERMP_RED & p->flags) {
+			chara(p, 27);
+			stringa(p, "[31m", 4);
+		}
+		if (TERMP_GREEN & p->flags) {
+			chara(p, 27);
+			stringa(p, "[32m", 4);
+		}
+		if (TERMP_YELLOW & p->flags) {
+			chara(p, 27);
+			stringa(p, "[33m", 4);
+		}
+		if (TERMP_BLUE & p->flags) {
+			chara(p, 27);
+			stringa(p, "[34m", 4);
+		}
+		if (TERMP_MAGENTA & p->flags) {
+			chara(p, 27);
+			stringa(p, "[35m", 4);
+		}
+		if (TERMP_CYAN & p->flags) {
+			chara(p, 27);
+			stringa(p, "[36m", 4);
+		}
+	}
 
 	for (i = 0; i < len; i++) {
 		if ('\\' == word[i]) {
 			pescape(p, word, &i, len);
 			continue;
 		}
+
+		if (TERMENC_NROFF == p->enc && 
+				TERMP_STYLE & p->flags) {
+			if (TERMP_BOLD & p->flags) {
+				chara(p, word[i]);
+				chara(p, 8);
+			}
+			if (TERMP_UNDER & p->flags) {
+				chara(p, '_');
+				chara(p, 8);
+			}
+		}
+
 		chara(p, word[i]);
 	}
 
-	if (p->flags & TERMP_BOLD ||
-			p->flags & TERMP_UNDERLINE)
-		stylea(p, TERMSTYLE_CLEAR);
+	if (TERMENC_ANSI == p->enc && TERMP_STYLE & p->flags) {
+		chara(p, 27);
+		stringa(p, "[00m", 4);
+	}
 }
 
 
@@ -799,18 +859,6 @@ symbola(struct termp *p, enum tsym sym)
 
 
 /*
- * Add a style to the output line buffer.
- */
-static void
-stylea(struct termp *p, enum tstyle style)
-{
-
-	assert(p->styletab[style].sym);
-	stringa(p, p->styletab[style].sym, p->styletab[style].sz);
-}
-
-
-/*
  * Like chara() but for arbitrary-length buffers.  Resize the buffer by
  * a factor of two (if the buffer is less than that) or the buffer's
  * size.
diff --git a/term.c b/term.c
index 4ba00b1b..06d4bc3f 100644
--- a/term.c
+++ b/term.c
@@ -1,4 +1,4 @@
-/* $Id: term.c,v 1.58 2009/03/13 07:46:10 kristaps Exp $ */
+/* $Id: term.c,v 1.59 2009/03/14 12:35:02 kristaps Exp $ */
 /*
  * Copyright (c) 2009 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -67,28 +67,28 @@
 /* TODO: abstract this into mdocterm.c. */
 
 const	int ttypes[TTYPE_NMAX] = {
-	TERMP_BOLD,		/* TTYPE_PROG */
+	TERMP_BOLD, 		/* TTYPE_PROG */
 	TERMP_BOLD,		/* TTYPE_CMD_FLAG */
-	TERMP_UNDERLINE, 	/* TTYPE_CMD_ARG */
+	TERMP_UNDER, 		/* TTYPE_CMD_ARG */
 	TERMP_BOLD, 		/* TTYPE_SECTION */
 	TERMP_BOLD,		/* TTYPE_FUNC_DECL */
-	TERMP_UNDERLINE,	/* TTYPE_VAR_DECL */
-	TERMP_UNDERLINE,	/* TTYPE_FUNC_TYPE */
+	TERMP_UNDER,		/* TTYPE_VAR_DECL */
+	TERMP_UNDER,		/* TTYPE_FUNC_TYPE */
 	TERMP_BOLD, 		/* TTYPE_FUNC_NAME */
-	TERMP_UNDERLINE, 	/* TTYPE_FUNC_ARG */
-	TERMP_UNDERLINE, 	/* TTYPE_LINK */
+	TERMP_UNDER, 		/* TTYPE_FUNC_ARG */
+	TERMP_UNDER, 		/* TTYPE_LINK */
 	TERMP_BOLD,	 	/* TTYPE_SSECTION */
-	TERMP_UNDERLINE, 	/* TTYPE_FILE */
-	TERMP_UNDERLINE, 	/* TTYPE_EMPH */
+	TERMP_UNDER, 		/* TTYPE_FILE */
+	TERMP_UNDER, 		/* TTYPE_EMPH */
 	TERMP_BOLD,	 	/* TTYPE_CONFIG */
 	TERMP_BOLD,	 	/* TTYPE_CMD */
 	TERMP_BOLD,	 	/* TTYPE_INCLUDE */
 	TERMP_BOLD,	 	/* TTYPE_SYMB */
 	TERMP_BOLD,	 	/* TTYPE_SYMBOL */
 	TERMP_BOLD,	 	/* TTYPE_DIAG */
-	TERMP_UNDERLINE, 	/* TTYPE_LINK_ANCHOR */
+	TERMP_UNDER, 		/* TTYPE_LINK_ANCHOR */
 	TERMP_BOLD,	 	/* TTYPE_LINK_TEXT */
-	TERMP_UNDERLINE	 	/* TTYPE_REF_TITLE */
+	TERMP_UNDER	 	/* TTYPE_REF_TITLE */
 };
 
 static	int		  arg_hasattr(int, const struct mdoc_node *);
diff --git a/term.h b/term.h
index 4e1c557a..db35ed70 100644
--- a/term.h
+++ b/term.h
@@ -1,4 +1,4 @@
-/* $Id: term.h,v 1.21 2009/03/14 05:36:07 kristaps Exp $ */
+/* $Id: term.h,v 1.22 2009/03/14 12:35:02 kristaps Exp $ */
 /*
  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -70,12 +70,9 @@ enum	tsym {
 	TERMSYM_MAX = 		41
 };
 
-
-enum	tstyle {
-	TERMSTYLE_CLEAR	=	0,
-	TERMSTYLE_BOLD =	1,
-	TERMSTYLE_UNDER =	2,
-	TERMSTYLE_MAX =		3
+enum	termenc {
+	TERMENC_ANSI,
+	TERMENC_NROFF
 };
 
 struct	termsym {
@@ -90,18 +87,26 @@ struct	termp {
 	size_t		  offset;
 	size_t		  col;
 	int		  flags;
-#define	TERMP_BOLD	 (1 << 0)	/* Embolden words. */
-#define	TERMP_UNDERLINE	 (1 << 1)	/* Underline words. */
-#define	TERMP_NOSPACE	 (1 << 2)	/* No space before words. */
-#define	TERMP_NOLPAD	 (1 << 3)	/* No leftpad before flush. */
-#define	TERMP_NOBREAK	 (1 << 4)	/* No break after flush. */
-#define	TERMP_LITERAL	 (1 << 5)	/* Literal words. */
-#define	TERMP_IGNDELIM	 (1 << 6)	/* Delims like regulars. */
-#define	TERMP_NONOSPACE	 (1 << 7)	/* No space (no autounset). */
-#define	TERMP_NONOBREAK	 (1 << 8)
+#define	TERMP_NOSPACE	 (1 << 0)	/* No space before words. */
+#define	TERMP_NOLPAD	 (1 << 1)	/* No leftpad before flush. */
+#define	TERMP_NOBREAK	 (1 << 2)	/* No break after flush. */
+#define	TERMP_LITERAL	 (1 << 3)	/* Literal words. */
+#define	TERMP_IGNDELIM	 (1 << 4)	/* Delims like regulars. */
+#define	TERMP_NONOSPACE	 (1 << 5)	/* No space (no autounset). */
+#define	TERMP_NONOBREAK	 (1 << 7)	/* Don't newln NOBREAK. */
+
+#define	TERMP_STYLE	 0xff00		/* Style mask. */
+#define	TERMP_BOLD	 (1 << 8)	/* Styles... */
+#define	TERMP_UNDER	 (1 << 9)
+#define	TERMP_BLUE	 (1 << 10)
+#define	TERMP_RED	 (1 << 11)
+#define	TERMP_YELLOW	 (1 << 12)
+#define	TERMP_MAGENTA	 (1 << 13)
+#define	TERMP_CYAN	 (1 << 14)
+#define	TERMP_GREEN	 (1 << 15)
 	char		 *buf;
 	struct termsym	 *symtab;	/* Special-symbol table. */
-	struct termsym	 *styletab;	/* Style table. */
+	enum termenc	  enc;		/* Encoding. */
 };
 
 struct	termpair {