Provide a new function roff_req_or_macro() to parse and handle a request

or macro, including context-dependent error handling inside tbl(7) code
and inside .ce/.rj blocks.  Use it both in the top level roff(7) parser
and inside conditional blocks.

This fixes an assertion failure triggered by ".if 1 .ce" inside tbl(7)
code, found by tb@ using afl(1).

As a side benefit for readability, only one place remains in the
code that calls the main handler functions for the various roff(7)
requests.  This patch also improves column numbers in some error
messages and various comments.

9 files changed, 161 insertions, 44 deletions

diff --git a/regress/roff/ce/Makefile b/regress/roff/ce/Makefile
index 2024bdce..2528847e 100644
--- a/regress/roff/ce/Makefile
+++ b/regress/roff/ce/Makefile
@@ -1,5 +1,10 @@
-# $OpenBSD: Makefile,v 1.1 2019/01/04 01:06:44 schwarze Exp $
+# $OpenBSD: Makefile,v 1.2 2022/04/30 18:46:16 schwarze Exp $
 
-REGRESS_TARGETS	 = basic
+REGRESS_TARGETS	 = basic break
+
+# mandoc defect:
+# - For now, high level macros break .ce and .rj blocks.
+
+SKIP_GROFF	 = break
 
 .include <bsd.regress.mk>
diff --git a/regress/roff/ce/break.in b/regress/roff/ce/break.in
new file mode 100644
index 00000000..69dc99c4
--- /dev/null
+++ b/regress/roff/ce/break.in
@@ -0,0 +1,33 @@
+.\" $OpenBSD: break.in,v 1.1 2022/04/30 18:46:16 schwarze Exp $
+.TH CE-BREAK 1 "April 30, 2022"
+.SH NAME
+ce-break \- centering requests broken by high-level macros
+.SH DESCRIPTION
+initial
+text
+.ce 3
+Text centered with the .ce request is
+.I not
+filled.
+.rj 4
+Text adjusted to the right margin
+works in just the same way and is
+.I not
+filled either.
+.ce
+High-level macros break centering even in a
+.if 1 .I conditional
+block.
+.PP
+.nf
+Now entering
+explicit no-fill mode.
+.ce 3
+Text is still
+.I not
+filled.
+.PP
+.fi
+final
+text
+in fill mode
diff --git a/regress/roff/ce/break.out_ascii b/regress/roff/ce/break.out_ascii
new file mode 100644
index 00000000..8572cc0a
--- /dev/null
+++ b/regress/roff/ce/break.out_ascii
@@ -0,0 +1,24 @@
+CE-BREAK(1)                 General Commands Manual                CE-BREAK(1)
+
+NNAAMMEE
+       ce-break - centering requests broken by high-level macros
+
+DDEESSCCRRIIPPTTIIOONN
+       initial text
+                        Text centered with the .ce request is
+       _n_o_t filled.
+                                             Text adjusted to the right margin
+                                             works in just the same way and is
+       _n_o_t filled either.
+                     High-level macros break centering even in a
+       _c_o_n_d_i_t_i_o_n_a_l block.
+
+       Now entering
+       explicit no-fill mode.
+                                    Text is still
+       _n_o_t
+       filled.
+
+       final text in fill mode
+
+OpenBSD                         April 30, 2022                     CE-BREAK(1)
diff --git a/regress/tbl/macro/Makefile b/regress/tbl/macro/Makefile
index 70ec1b33..8ea4a0d6 100644
--- a/regress/tbl/macro/Makefile
+++ b/regress/tbl/macro/Makefile
@@ -1,12 +1,15 @@
-# $OpenBSD: Makefile,v 1.2 2016/08/20 14:43:40 schwarze Exp $
+# $OpenBSD: Makefile,v 1.3 2022/04/30 18:46:16 schwarze Exp $
 
-REGRESS_TARGETS	 = man nested column
-LINT_TARGETS	 = man nested
+REGRESS_TARGETS	 = column man nested req
+LINT_TARGETS	 = man nested req
 
-# trivial differences to groff-1.22.3:
+# mandoc defect:
+# mandoc(1) ignores .br inside tbl(7) code.
+
+# trivial differences to groff-1.23.0:
 # .TS in a table causes a blank table line in GNU tbl(1), but not in mandoc.
 # .TS in a column list causes a blank line in mandoc, but not in GNU tbl(1).
 
-SKIP_GROFF	 = nested column
+SKIP_GROFF	 = column nested req
 
 .include <bsd.regress.mk>
diff --git a/regress/tbl/macro/nested.out_lint b/regress/tbl/macro/nested.out_lint
index dd044ba7..9bffbb0c 100644
--- a/regress/tbl/macro/nested.out_lint
+++ b/regress/tbl/macro/nested.out_lint
@@ -1 +1 @@
-mandoc: nested.in:13:4: UNSUPP: ignoring macro in table: TS
+mandoc: nested.in:13:2: UNSUPP: ignoring macro in table: TS
diff --git a/regress/tbl/macro/req.in b/regress/tbl/macro/req.in
new file mode 100644
index 00000000..30e1fe98
--- /dev/null
+++ b/regress/tbl/macro/req.in
@@ -0,0 +1,30 @@
+.\" $OpenBSD: req.in,v 1.1 2022/04/30 18:46:16 schwarze Exp $
+.TH TBL-MACRO-REQ 1 "April 30, 2022"
+.SH NAME
+tbl-macro-req \- requests generating nodes in a table
+.SH DESCRIPTION
+initial text
+.TS
+box tab(:);
+l | l .
+a:b:stray
+_
+c:T{
+d
+e
+T}
+.ce
+f:T{
+g
+.br
+h
+T}
+.if 1 .ce
+i:T{
+j
+.if 1 .br
+k
+T}
+.TE
+.PP
+final text
diff --git a/regress/tbl/macro/req.out_ascii b/regress/tbl/macro/req.out_ascii
new file mode 100644
index 00000000..168b1ea1
--- /dev/null
+++ b/regress/tbl/macro/req.out_ascii
@@ -0,0 +1,18 @@
+TBL-MACRO-REQ(1)            General Commands Manual           TBL-MACRO-REQ(1)
+
+NNAAMMEE
+       tbl-macro-req - requests generating nodes in a table
+
+DDEESSCCRRIIPPTTIIOONN
+       initial text
+
+       +--+-----+
+       |a | b   |
+       +--+-----+
+       |c | d e |
+       |f | g h |
+       |i | j k |
+       +--+-----+
+       final text
+
+OpenBSD                         April 30, 2022                TBL-MACRO-REQ(1)
diff --git a/regress/tbl/macro/req.out_lint b/regress/tbl/macro/req.out_lint
new file mode 100644
index 00000000..26af8f71
--- /dev/null
+++ b/regress/tbl/macro/req.out_lint
@@ -0,0 +1,5 @@
+mandoc: req.in:10:5: ERROR: ignoring extra tbl data cells: stray
+mandoc: req.in:16:2: UNSUPP: ignoring macro in table: ce
+mandoc: req.in:19:2: UNSUPP: ignoring macro in table: br
+mandoc: req.in:22:8: UNSUPP: ignoring macro in table: ce
+mandoc: req.in:25:8: UNSUPP: ignoring macro in table: br
diff --git a/roff.c b/roff.c
index ca45b0f6..323122ba 100644
--- a/roff.c
+++ b/roff.c
@@ -1,4 +1,4 @@
-/* $Id: roff.c,v 1.385 2022/04/30 11:32:42 schwarze Exp $ */
+/* $Id: roff.c,v 1.386 2022/04/30 18:51:36 schwarze Exp $ */
 /*
  * Copyright (c) 2010-2015, 2017-2022 Ingo Schwarze <schwarze@openbsd.org>
  * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
@@ -237,6 +237,7 @@ static	enum roff_tok	 roff_parse(struct roff *, char *, int *,
 static	int		 roff_parsetext(struct roff *, struct buf *,
 				int, int *);
 static	int		 roff_renamed(ROFF_ARGS);
+static	int		 roff_req_or_macro(ROFF_ARGS);
 static	int		 roff_return(ROFF_ARGS);
 static	int		 roff_rm(ROFF_ARGS);
 static	int		 roff_rn(ROFF_ARGS);
@@ -1905,7 +1906,6 @@ roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs, size_t len)
 	/*
 	 * If a scope is open, go to the child handler for that macro,
 	 * as it may want to preprocess before doing anything with it.
-	 * Don't do so if an equation is open.
 	 */
 
 	if (r->last) {
@@ -1913,19 +1913,27 @@ roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs, size_t len)
 		return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
 	}
 
-	/* No scope is open.  This is a new request or macro. */
-
 	r->options &= ~MPARSE_COMMENT;
 	spos = pos;
 	t = roff_parse(r, buf->buf, &pos, ln, ppos);
+	return roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
+}
+
+/*
+ * Handle a new request or macro.
+ * May be called outside any scope or from inside a conditional scope.
+ */
+static int
+roff_req_or_macro(ROFF_ARGS) {
 
-	/* Tables ignore most macros. */
+	/* For now, tables ignore most macros and some request. */
 
-	if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
-	    t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
+	if (r->tbl != NULL && (tok == TOKEN_NONE || tok == ROFF_TS ||
+	    tok == ROFF_br || tok == ROFF_ce || tok == ROFF_rj ||
+	    tok == ROFF_sp)) {
 		mandoc_msg(MANDOCERR_TBLMACRO,
-		    ln, pos, "%s", buf->buf + spos);
-		if (t != TOKEN_NONE)
+		    ln, ppos, "%s", buf->buf + ppos);
+		if (tok != TOKEN_NONE)
 			return ROFF_IGN;
 		while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
 			pos++;
@@ -1938,9 +1946,9 @@ roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs, size_t len)
 
 	/* For now, let high level macros abort .ce mode. */
 
-	if (ctl && roffce_node != NULL &&
-	    (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
-	     t == ROFF_TH || t == ROFF_TS)) {
+	if (roffce_node != NULL &&
+	    (tok == TOKEN_NONE || tok == ROFF_Dd || tok == ROFF_EQ ||
+	     tok == ROFF_TH || tok == ROFF_TS)) {
 		r->man->last = roffce_node;
 		r->man->next = ROFF_NEXT_SIBLING;
 		roffce_lines = 0;
@@ -1952,12 +1960,12 @@ roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs, size_t len)
 	 * Let the standard macro set parsers handle it.
 	 */
 
-	if (t == TOKEN_NONE)
+	if (tok == TOKEN_NONE)
 		return ROFF_CONT;
 
-	/* Execute a roff request or a user defined macro. */
+	/* Execute a roff request or a user-defined macro. */
 
-	return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
+	return (*roffs[tok].proc)(r, tok, buf, ln, ppos, pos, offs);
 }
 
 /*
@@ -2000,8 +2008,10 @@ roff_endparse(struct roff *r)
 }
 
 /*
- * Parse a roff node's type from the input buffer.  This must be in the
- * form of ".foo xxx" in the usual way.
+ * Parse the request or macro name at buf[*pos].
+ * Return ROFF_RENAMED, ROFF_USERDEF, or a ROFF_* token value.
+ * For empty, undefined, mdoc(7), and man(7) macros, return TOKEN_NONE.
+ * As a side effect, set r->current_string to the definition or to NULL.
  */
 static enum roff_tok
 roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
@@ -2393,27 +2403,18 @@ static int
 roff_cond_sub(ROFF_ARGS)
 {
 	struct roffnode	*bl;
-	int		 irc, rr;
+	int		 irc, rr, spos;
 	enum roff_tok	 t;
 
 	rr = 0;  /* If arguments follow "\}", skip them. */
 	irc = roff_cond_checkend(r, tok, buf, ln, ppos, pos, &rr);
+	spos = pos;
 	t = roff_parse(r, buf->buf, &pos, ln, ppos);
 
-	/* For now, let high level macros abort .ce mode. */
-
-	if (roffce_node != NULL &&
-	    (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
-             t == ROFF_TH || t == ROFF_TS)) {
-		r->man->last = roffce_node;
-		r->man->next = ROFF_NEXT_SIBLING;
-		roffce_lines = 0;
-		roffce_node = NULL;
-	}
-
 	/*
-	 * Fully handle known macros when they are structurally
-	 * required or when the conditional evaluated to true.
+	 * Handle requests and macros if the conditional evaluated
+	 * to true or if they are structurally required.
+	 * The .break request is always handled specially.
 	 */
 
 	if (t == ROFF_break) {
@@ -2426,13 +2427,11 @@ roff_cond_sub(ROFF_ARGS)
 					break;
 			}
 		}
-	} else if (t != TOKEN_NONE &&
-	    (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
-		irc |= (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
+	} else if (rr || (t < TOKEN_NONE && roffs[t].flags & ROFFMAC_STRUCT)) {
+		irc |= roff_req_or_macro(r, t, buf, ln, spos, pos, offs);
 		if (irc & ROFF_WHILE)
 			irc &= ~(ROFF_LOOPCONT | ROFF_LOOPEXIT);
-	} else
-		irc |= rr ? ROFF_CONT : ROFF_IGN;
+	}
 	return irc;
 }