In ./configure, select a RE syntax for word boundaries supported by libc;

issue reported by Svyatoslav Mishyn, Peter Bray, and Daniel Levai.

5 files changed, 67 insertions, 8 deletions

diff --git a/TODO b/TODO
index 7be67398..f851e9fd 100644
--- a/TODO
+++ b/TODO
@@ -1,6 +1,6 @@
 ************************************************************************
 * Official mandoc TODO.
-* $Id: TODO,v 1.214 2015/09/29 18:45:36 schwarze Exp $
+* $Id: TODO,v 1.215 2015/11/06 21:19:09 schwarze Exp $
 ************************************************************************
 
 Many issues are annotated for difficulty as follows:
@@ -527,12 +527,6 @@ are mere guesses, and some may be wrong.
 * portability
 ************************************************************************
 
-- word boundaries in regular expressions for whatis(1)
-  set up config tests to use [[:<:]], \<, or nothing
-  Svyatoslav Mishyn  Wed, 17 Dec 2014 11:07:10 +0200
-  reminded by Peter Bray Fri, 03 Apr 2015 23:02:16 +1100
-  loc *  exist *  algo *  size *  imp *
-
 - systems having UTF-8 but not en_US.UTF-8
   call locale(1) from ./configure, select a UTF-8-locale,
   and use that for test-wchar.c and term_ascii.c
diff --git a/configure b/configure
index 28435810..882b8f4b 100755
--- a/configure
+++ b/configure
@@ -53,6 +53,8 @@ HAVE_MMAP=
 HAVE_PLEDGE=
 HAVE_PROGNAME=
 HAVE_REALLOCARRAY=
+HAVE_REWB_BSD=
+HAVE_REWB_SYSV=
 HAVE_STRCASESTR=
 HAVE_STRINGLIST=
 HAVE_STRLCAT=
@@ -184,6 +186,8 @@ runtest mmap		MMAP		|| true
 runtest pledge		PLEDGE		|| true
 runtest progname	PROGNAME	|| true
 runtest reallocarray	REALLOCARRAY	|| true
+runtest rewb-bsd	REWB_BSD	|| true
+runtest rewb-sysv	REWB_SYSV	|| true
 runtest strcasestr	STRCASESTR	|| true
 runtest stringlist	STRINGLIST	|| true
 runtest strlcat		STRLCAT		|| true
@@ -306,6 +310,8 @@ cat << __HEREDOC__
 #define HAVE_PLEDGE ${HAVE_PLEDGE}
 #define HAVE_PROGNAME ${HAVE_PROGNAME}
 #define HAVE_REALLOCARRAY ${HAVE_REALLOCARRAY}
+#define HAVE_REWB_BSD ${HAVE_REWB_BSD}
+#define HAVE_REWB_SYSV ${HAVE_REWB_SYSV}
 #define HAVE_STRCASESTR ${HAVE_STRCASESTR}
 #define HAVE_STRINGLIST ${HAVE_STRINGLIST}
 #define HAVE_STRLCAT ${HAVE_STRLCAT}
diff --git a/mansearch.c b/mansearch.c
index 41fdb499..756a73f9 100644
--- a/mansearch.c
+++ b/mansearch.c
@@ -1,4 +1,4 @@
-/*	$Id: mansearch.c,v 1.60 2015/10/13 15:53:05 schwarze Exp $ */
+/*	$Id: mansearch.c,v 1.61 2015/11/06 21:19:09 schwarze Exp $ */
 /*
  * Copyright (c) 2012 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2013, 2014, 2015 Ingo Schwarze <schwarze@openbsd.org>
@@ -766,7 +766,14 @@ exprterm(const struct mansearch *search, char *buf, int cs)
 	if (search->argmode == ARG_WORD) {
 		e->bits = TYPE_Nm;
 		e->substr = NULL;
+#if HAVE_REWB_BSD
 		mandoc_asprintf(&val, "[[:<:]]%s[[:>:]]", buf);
+#elif HAVE_REWB_SYSV
+		mandoc_asprintf(&val, "\\<%s\\>", buf);
+#else
+		mandoc_asprintf(&val,
+		    "(^|[^a-zA-Z01-9_])%s([^a-zA-Z01-9_]|$)", buf);
+#endif
 		cs = 0;
 	} else if ((val = strpbrk(buf, "=~")) == NULL) {
 		e->bits = TYPE_Nm | TYPE_Nd;
diff --git a/test-rewb-bsd.c b/test-rewb-bsd.c
new file mode 100644
index 00000000..88d3d357
--- /dev/null
+++ b/test-rewb-bsd.c
@@ -0,0 +1,26 @@
+#include <sys/types.h>
+#include <regex.h>
+
+int
+main(void)
+{
+	regex_t	 re;
+
+	if (regcomp(&re, "[[:<:]]word[[:>:]]", REG_EXTENDED | REG_NOSUB))
+		return 1;
+	if (regexec(&re, "the word is here", 0, NULL, 0))
+		return 2;
+	if (regexec(&re, "same word", 0, NULL, 0))
+		return 3;
+	if (regexec(&re, "word again", 0, NULL, 0))
+		return 4;
+	if (regexec(&re, "word", 0, NULL, 0))
+		return 5;
+	if (regexec(&re, "wordy", 0, NULL, 0) != REG_NOMATCH)
+		return 6;
+	if (regexec(&re, "sword", 0, NULL, 0) != REG_NOMATCH)
+		return 7;
+	if (regexec(&re, "reworded", 0, NULL, 0) != REG_NOMATCH)
+		return 8;
+	return 0;
+}
diff --git a/test-rewb-sysv.c b/test-rewb-sysv.c
new file mode 100644
index 00000000..cb35c544
--- /dev/null
+++ b/test-rewb-sysv.c
@@ -0,0 +1,26 @@
+#include <sys/types.h>
+#include <regex.h>
+
+int
+main(void)
+{
+	regex_t	 re;
+
+	if (regcomp(&re, "\\<word\\>", REG_EXTENDED | REG_NOSUB))
+		return 1;
+	if (regexec(&re, "the word is here", 0, NULL, 0))
+		return 2;
+	if (regexec(&re, "same word", 0, NULL, 0))
+		return 3;
+	if (regexec(&re, "word again", 0, NULL, 0))
+		return 4;
+	if (regexec(&re, "word", 0, NULL, 0))
+		return 5;
+	if (regexec(&re, "wordy", 0, NULL, 0) != REG_NOMATCH)
+		return 6;
+	if (regexec(&re, "sword", 0, NULL, 0) != REG_NOMATCH)
+		return 7;
+	if (regexec(&re, "reworded", 0, NULL, 0) != REG_NOMATCH)
+		return 8;
+	return 0;
+}