Autodetect a suitable locale for -Tutf8 mode,

and allow overriding it manually.
Based on a patch from Svyatoslav Mishyn <juef at openmailbox dot org>
tweaked by me.
The idea originally came up in a conversation with Markus Waldeck.

diff --git a/TODO b/TODO
index c22d1978b039749b62a152ddefc32cbdc1f92fa4..9fc7b438525d81f9fd3539fb2fdd7a0390276ded 100644 (file)
--- a/TODO
+++ b/TODO
@@ -1,6 +1,6 @@
 ************************************************************************
 * Official mandoc TODO.
-* $Id: TODO,v 1.219 2016/07/19 21:31:55 schwarze Exp $
+* $Id: TODO,v 1.220 2016/07/31 09:29:13 schwarze Exp $
 ************************************************************************
 
 Many issues are annotated for difficulty as follows:
@@ -528,16 +528,6 @@ are mere guesses, and some may be wrong.
   in dig(1).
   loc **  exist **  algo **  size *  imp **
 
-************************************************************************
-* portability
-************************************************************************
-
-- 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
-  to Markus Waldeck  Sat, 18 Jul 2015 01:55:37 +0200
-  loc *  exist *  algo *  size *  imp *
-
 ************************************************************************
 * warning issues
 ************************************************************************

diff --git a/configure b/configure
index 64a61f2b125969e5b666784849b0ff6e8c437a1a..a2f6174f716f94c3d8dd10311f9571386ed65833 100755 (executable)
--- a/configure
+++ b/configure
@@ -1,6 +1,6 @@
 #!/bin/sh
 #
-# $Id: configure,v 1.47 2016/07/20 14:03:06 schwarze Exp $
+# $Id: configure,v 1.48 2016/07/31 09:29:13 schwarze Exp $
 #
 # Copyright (c) 2014, 2015, 2016 Ingo Schwarze <schwarze@openbsd.org>
 #
@@ -35,6 +35,7 @@ echo "config.log: writing..."
 
 MANPATH_DEFAULT="/usr/share/man:/usr/X11R6/man:/usr/local/man"
 OSNAME=
+UTF8_LOCALE=
 
 CC=`printf "all:\\n\\t@echo \\\$(CC)\\n" | env -i make -f -`
 CFLAGS="-g -W -Wall -Wmissing-prototypes -Wstrict-prototypes -Wwrite-strings"
@@ -176,6 +177,23 @@ runtest() {
        return 1
 }
 
+# Select a UTF-8 locale.
+get_locale() {
+       [ -n "${HAVE_WCHAR}" ] && [ "${HAVE_WCHAR}" -eq 0 ] && return 0
+       ismanual UTF8_LOCALE "$UTF8_LOCALE" && return 0
+       echo "UTF8_LOCALE: testing..." 1>&3
+       UTF8_LOCALE=`locale -a | grep -i '^en_US\.UTF-*8$' | head -n 1`
+       if [ -z "${UTF8_LOCALE}" ]; then
+               UTF8_LOCALE=`locale -a | grep -i '\.UTF-*8' | head -n 1`
+               [ -n "${UTF8_LOCALE}" ] || return 1
+       fi
+       echo "UTF8_LOCALE=${UTF8_LOCALE}" 1>&2
+       echo "UTF8_LOCALE=${UTF8_LOCALE}" 1>&3
+       echo 1>&3
+       return 0;
+}
+
+
 # --- library functions ---
 runtest dirent-namlen  DIRENT_NAMLEN   || true
 runtest EFTYPE         EFTYPE          || true
@@ -199,7 +217,15 @@ runtest strptime   STRPTIME        || true
 runtest strsep         STRSEP          || true
 runtest strtonum       STRTONUM        || true
 runtest vasprintf      VASPRINTF       || true
-runtest wchar          WCHAR           || true
+
+# --- wide character and locale support ---
+if get_locale; then
+       runtest wchar WCHAR -DUTF8_LOCALE=\"${UTF8_LOCALE}\" || true
+else
+       HAVE_WCHAR=0
+       echo "wchar: no (no UTF8_LOCALE)" 1>&2
+       echo "wchar: no (no UTF8_LOCALE)" 1>&3
+fi
 
 # --- nanosleep ---
 if [ -n "${LD_NANOSLEEP}" ]; then
@@ -264,6 +290,7 @@ echo
 echo "#define MAN_CONF_FILE \"/etc/${MANM_MANCONF}\""
 echo "#define MANPATH_DEFAULT \"${MANPATH_DEFAULT}\""
 [ -n "${OSNAME}" ] && echo "#define OSNAME \"${OSNAME}\""
+[ -n "${UTF8_LOCALE}" ] && echo "#define UTF8_LOCALE \"${UTF8_LOCALE}\""
 [ -n "${HOMEBREWDIR}" ] && echo "#define HOMEBREWDIR \"${HOMEBREWDIR}\""
 [ ${HAVE_EFTYPE} -eq 0 ] && echo "#define EFTYPE EINVAL"
 

diff --git a/configure.local.example b/configure.local.example
index 18a84fbdea09b15a1197ad47c5c2286c351a3b4f..c1c7be93df6f98ea77fb4ecd1076a24f02787ac3 100644 (file)
--- a/configure.local.example
+++ b/configure.local.example
@@ -1,4 +1,4 @@
-# $Id: configure.local.example,v 1.17 2016/07/20 14:03:06 schwarze Exp $
+# $Id: configure.local.example,v 1.18 2016/07/31 09:29:13 schwarze Exp $
 #
 # Copyright (c) 2014, 2015, 2016 Ingo Schwarze <schwarze@openbsd.org>
 #
@@ -48,6 +48,13 @@ HAVE_WCHAR=1
 
 HAVE_WCHAR=0
 
+# For -Tutf8 mode, mandoc needs to set an arbitrary locale having
+# a UTF-8 character set.  If autodetection of a suitable locale
+# fails or selects an undesirable locale, you can manually choose
+# the locale for -Tutf8 mode:
+
+UTF8_LOCALE=en_US.UTF-8
+
 # When man(1) or apropos(1) is called without -m and -M options,
 # MANPATH is not set in the environment, and man.conf(5) is not
 # available, manuals are searched for in the following directory

diff --git a/term_ascii.c b/term_ascii.c
index fecdb0a964dc9adf76885e1c4def527f1baca695..df5ff13901c26b3c8bc738deb3475fccd63b617f 100644 (file)
--- a/term_ascii.c
+++ b/term_ascii.c
@@ -1,4 +1,4 @@
-/*     $Id: term_ascii.c,v 1.53 2016/07/08 22:29:05 schwarze Exp $ */
+/*     $Id: term_ascii.c,v 1.54 2016/07/31 09:29:13 schwarze Exp $ */
 /*
  * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2014, 2015 Ingo Schwarze <schwarze@openbsd.org>
@@ -98,7 +98,7 @@ ascii_init(enum termenc enc, const struct manoutput *outopts)
 
                v = TERMENC_LOCALE == enc ?
                    setlocale(LC_CTYPE, "") :
-                   setlocale(LC_CTYPE, "en_US.UTF-8");
+                   setlocale(LC_CTYPE, UTF8_LOCALE);
                if (NULL != v && MB_CUR_MAX > 1) {
                        p->enc = enc;
                        p->advance = locale_advance;

diff --git a/test-wchar.c b/test-wchar.c
index a096705ebcf18349119184ec069480796c796ab1..32962d9fe665fbfa5bc5913932213a6c16732e42 100644 (file)
--- a/test-wchar.c
+++ b/test-wchar.c
@@ -1,4 +1,4 @@
-/*     $Id: test-wchar.c,v 1.3 2015/10/06 18:32:20 schwarze Exp $      */
+/*     $Id: test-wchar.c,v 1.4 2016/07/31 09:29:13 schwarze Exp $      */
 /*
  * Copyright (c) 2014 Ingo Schwarze <schwarze@openbsd.org>
  *
@@ -35,9 +35,9 @@ main(void)
                return 1;
        }
 
-       if (setlocale(LC_CTYPE, "en_US.UTF-8") == NULL) {
-               fputs("setlocale(LC_CTYPE, \"en_US.UTF-8\") failed\n",
-                   stderr);
+       if (setlocale(LC_CTYPE, UTF8_LOCALE) == NULL) {
+               fprintf(stderr, "setlocale(LC_CTYPE, \"%s\") failed\n",
+                   UTF8_LOCALE);
                return 1;
        }