mandoc: Fix three bugs regarding the interaction of \z and \h: 1.

[schwarze@mandoc.bsd.lv]

Log Message:
-----------
Fix three bugs regarding the interaction of \z and \h:

1. The combination \z\h is a no-op whatever the argument may be.
In the past, the \z only affected the first space character generated
by the \h, which was wrong.

2. For the conbination \zX\h with a positive argument, the first
space resulting from the \h is not printed but consumed by the \z.

3. For the combination \zX\h with a negative argument, application
of the \z needs to be completed before the \h can be started.
In the past, if this combination occurred at the beginning of an
output line, the \h backed up to the beginning of the line and
after that, the \z attempted to back up even further, triggering
an assertion.

Bugs found during an audit of assignments to termp->col that i
started after the bugfix tbl_term.c rev. 1.65.  The assertion
triggered by bug 3 was *not* yet found by afl(1).

Modified Files:
--------------
    mandoc:
        term.c
    mandoc/regress/roff/esc:
        Makefile
        z.in
        z.out_ascii

Added Files:
-----------
    mandoc/regress/roff/esc:
        hneg.in
        hneg.out_ascii

Revision Data
-------------
Index: term.c
===================================================================
RCS file: /home/cvs/mandoc/mandoc/term.c,v
retrieving revision 1.285
retrieving revision 1.286
diff -Lterm.c -Lterm.c -u -p -r1.285 -r1.286
--- term.c
+++ term.c
@@ -628,6 +628,10 @@ term_word(struct termp *p, const char *w
 				encode(p, "utf8", 4);
 			continue;
 		case ESCAPE_HORIZ:
+			if (p->flags & TERMP_BACKAFTER) {
+				p->flags &= ~TERMP_BACKAFTER;
+				continue;
+			}
 			if (*seq == '|') {
 				seq++;
 				uc = -p->col;
@@ -636,12 +640,22 @@ term_word(struct termp *p, const char *w
 			if (a2roffsu(seq, &su, SCALE_EM) == NULL)
 				continue;
 			uc += term_hen(p, &su);
-			if (uc > 0) {
+			if (uc >= 0) {
 				while (uc > 0) {
-					bufferc(p, ASCII_NBRSP);
 					uc -= term_len(p, 1);
+					if (p->flags & TERMP_BACKBEFORE)
+						p->flags &= ~TERMP_BACKBEFORE;
+					else
+						bufferc(p, ASCII_NBRSP);
 				}
-			} else if (p->col > (size_t)(-uc)) {
+				continue;
+			}
+			if (p->flags & TERMP_BACKBEFORE) {
+				p->flags &= ~TERMP_BACKBEFORE;
+				assert(p->col > 0);
+				p->col--;
+			}
+			if (p->col >= (size_t)(-uc)) {
 				p->col += uc;
 			} else {
 				uc += p->col;
--- /dev/null
+++ regress/roff/esc/hneg.out_ascii
@@ -0,0 +1,10 @@
+ESC-HNEG(1)                 General Commands Manual                ESC-HNEG(1)
+
+N\bNA\bAM\bME\bE
+     e\bes\bsc\bc-\b-h\bhn\bne\beg\bg - the roff escape h sequence with a negative argument
+
+D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
+     simple: >d<c
+     after z escape: >d<c
+
+OpenBSD                         April 27, 2022                         OpenBSD
Index: z.out_ascii
===================================================================
RCS file: /home/cvs/mandoc/mandoc/regress/roff/esc/z.out_ascii,v
retrieving revision 1.2
retrieving revision 1.3
diff -Lregress/roff/esc/z.out_ascii -Lregress/roff/esc/z.out_ascii -u -p -r1.2 -r1.3
--- regress/roff/esc/z.out_ascii
+++ regress/roff/esc/z.out_ascii
@@ -10,7 +10,10 @@ D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      single z with font escape: >x\bx\bb\bbo\bol\bld\bd<
      single z with nospace escape: > new line<
      single z with overstrike: >a\bb\bc<
+     single z with h escape: >m<
+     single z with char and h with positive argument: >a  b<
+     single z with char and h with zero argument: >a\bb<
      single z near the end of the line: ><
      double z: >x\b<
 
-OpenBSD                          July 4, 2017                          OpenBSD
+OpenBSD                         April 27, 2022                         OpenBSD
--- /dev/null
+++ regress/roff/esc/hneg.in
@@ -0,0 +1,11 @@
+.\" $OpenBSD: hneg.in,v 1.1 2022/04/27 13:30:19 schwarze Exp $
+.Dd $Mdocdate: April 27 2022 $
+.Dt ESC-HNEG 1
+.Os
+.Sh NAME
+.Nm esc-hneg
+.Nd the roff escape h sequence with a negative argument
+.Sh DESCRIPTION
+simple: >abc\h'-3'd<
+.br
+after z escape: >ab\zc\h'-2'd<
Index: z.in
===================================================================
RCS file: /home/cvs/mandoc/mandoc/regress/roff/esc/z.in,v
retrieving revision 1.2
retrieving revision 1.3
diff -Lregress/roff/esc/z.in -Lregress/roff/esc/z.in -u -p -r1.2 -r1.3
--- regress/roff/esc/z.in
+++ regress/roff/esc/z.in
@@ -1,4 +1,4 @@
-.\" $OpenBSD: z.in,v 1.3 2017/07/04 14:53:27 schwarze Exp $
+.\" $OpenBSD: z.in,v 1.4 2022/04/27 13:30:19 schwarze Exp $
 .Dd $Mdocdate$
 .Dt ESC-Z 1
 .Os
@@ -19,6 +19,12 @@ single z with nospace escape: >\z\c
 new line<
 .br
 single z with overstrike: >\z\o'ab'c<
+.br
+single z with h escape: >\z\h'3'm\z\h'-3'<
+.br
+single z with char and h with positive argument: >\za\h'3'b<
+.br
+single z with char and h with zero argument: >\za\h'0'b<
 .br
 single z near the end of the line: >\z<
 .br
Index: Makefile
===================================================================
RCS file: /home/cvs/mandoc/mandoc/regress/roff/esc/Makefile,v
retrieving revision 1.8
retrieving revision 1.9
diff -Lregress/roff/esc/Makefile -Lregress/roff/esc/Makefile -u -p -r1.8 -r1.9
--- regress/roff/esc/Makefile
+++ regress/roff/esc/Makefile
@@ -1,8 +1,15 @@
-# $OpenBSD: Makefile,v 1.18 2022/04/13 13:11:33 schwarze Exp $
+# $OpenBSD: Makefile,v 1.19 2022/04/27 13:30:19 schwarze Exp $
 
-REGRESS_TARGETS	 = one two multi B bs_man bs_mdoc c c_man E1 e f h l O1 o p w z
+REGRESS_TARGETS	 = one two multi
+REGRESS_TARGETS	+= B bs_man bs_mdoc c c_man E1 e f h hneg l O1 o p w z
 REGRESS_TARGETS	+= ignore invalid unsupp
 HTML_TARGETS	 = f
 LINT_TARGETS	 = B h l O1 w ignore invalid unsupp
+
+# mandoc defect:
+# - \h with a negative argument replaces output characters
+#   instead of overstriking them
+
+SKIP_GROFF	 = hneg
 
 .include <bsd.regress.mk>
--
 To unsubscribe send an email to source+unsubscribe@mandoc.bsd.lv