Convey locale-related environment variables as backup, and nicer error.

Fixes #74.

man/mosh-server.1

@@ -76,6 +76,12 @@ UDP port number to bind
 .B \-c \fICOLORS\fP
 Number of colors to advertise to applications through TERM (e.g. 8, 256)
 
+.TP
+.B \-l \fINAME=VALUE\fP
+Locale-related environment variable to try as part of a fallback
+environment, if the startup environment does not specify a character
+set of UTF-8.
+
 .SH EXAMPLE
 
 .nf

man/mosh.1

@@ -64,6 +64,18 @@ pass between client and server. By default, \fBmosh\fP uses the ports
 between 60000 and 61000, but allows the user to request a particular
 UDP port instead.
 
+\fBmosh\fP will do its best to arrange a UTF-8 character set locale on
+the client and server. The client must have locale-related environment
+variables that specify UTF-8. \fBmosh\fP will pass these client
+variables to the \fBmosh-server\fP on its command line, but in most
+cases they will not need to be used. \fBmosh-server\fP first attempts
+to use its own locale-related environment variables, which come from
+the system default configuration (sometimes /etc/default/locale) or
+from having been passed over the SSH connection. But if these
+variables don't call for the use of UTF-8, \fBmosh-server\fP will
+apply the locale-related environment variables from the client and try
+again.
+
 .SH OPTIONS
 .TP
 .B \fIcommand\fP

scripts/mosh

@@ -196,6 +196,10 @@ if ( $pid == 0 ) { # child
     push @server, ( '-p', $port_request );
   }
 
+  for ( &locale_vars ) {
+    push @server, ( '-l', $_ );
+  }
+
   if ( scalar @command > 0 ) {
     push @server, '--', @command;
   }
@@ -237,3 +241,17 @@ if ( $pid == 0 ) { # child
 }
 
 sub shell_quote { join ' ', map {(my $a = $_) =~ s/'/'\\''/g; "'$a'"} @_ }
+
+sub locale_vars {
+  my @names = qw[LANG LANGUAGE LC_CTYPE LC_NUMERIC LC_TIME LC_COLLATE LC_MONETARY LC_MESSAGES LC_PAPER LC_NAME LC_ADDRESS LC_TELEPHONE LC_MEASUREMENT LC_IDENTIFICATION LC_ALL];
+
+  my @assignments;
+
+  for ( @names ) {
+    if ( defined $ENV{ $_ } ) {
+      push @assignments, $_ . q{=} . $ENV{ $_ };
+    }
+  }
+
+  return @assignments;
+}

src/examples/benchmark.cc

@@ -61,7 +61,7 @@ int main( void )
 
   /* Adopt native locale */
   set_native_locale();
-  assert_utf8_locale();
+  assert( is_utf8_locale() );
 
   for ( int i = 0; i < ITERATIONS; i++ ) {
     /* type a character */

src/examples/parse.cc

@@ -57,7 +57,7 @@ int main( int argc __attribute__((unused)),
   struct termios saved_termios, raw_termios, child_termios;
 
   set_native_locale();
-  assert_utf8_locale();
+  assert( is_utf8_locale() );
 
   if ( tcgetattr( STDIN_FILENO, &saved_termios ) < 0 ) {
     perror( "tcgetattr" );

src/examples/termemu.cc

@@ -62,7 +62,7 @@ int main( void )
   struct termios saved_termios, raw_termios, child_termios;
 
   set_native_locale();
-  assert_utf8_locale();
+  assert( is_utf8_locale() );
 
   if ( tcgetattr( STDIN_FILENO, &saved_termios ) < 0 ) {
     perror( "tcgetattr" );

src/frontend/mosh-server.cc

@@ -72,7 +72,7 @@ using namespace std;
 
 void print_usage( const char *argv0 )
 {
-  fprintf( stderr, "Usage: %s new [-s] [-v] [-i LOCALADDR] [-p PORT] [-c COLORS] [-- COMMAND...]\n", argv0 );
+  fprintf( stderr, "Usage: %s new [-s] [-v] [-i LOCALADDR] [-p PORT] [-c COLORS] [-l NAME=VALUE] [-- COMMAND...]\n", argv0 );
 }
 
 /* Simple spinloop */
@@ -113,6 +113,7 @@ int main( int argc, char *argv[] )
   int colors = 0;
   bool verbose = false; /* don't close stdin/stdout/stderr */
   /* Will cause mosh-server not to correctly detach on old versions of sshd. */
+  list<string> locale_vars;
 
   /* strip off command */
   for ( int i = 0; i < argc; i++ ) {
@@ -130,7 +131,7 @@ int main( int argc, char *argv[] )
        && (strcmp( argv[ 1 ], "new" ) == 0) ) {
     /* new option syntax */
     int opt;
-    while ( (opt = getopt( argc - 1, argv + 1, "i:p:c:sv" )) != -1 ) {
+    while ( (opt = getopt( argc - 1, argv + 1, "i:p:c:svl:" )) != -1 ) {
       switch ( opt ) {
       case 'i':
 	desired_ip = optarg;
@@ -148,6 +149,9 @@ int main( int argc, char *argv[] )
       case 'v':
 	verbose = true;
 	break;
+      case 'l':
+	locale_vars.push_back( string( optarg ) );
+	break;
       default:
 	print_usage( argv[ 0 ] );
 	/* don't die on unknown options */
@@ -203,7 +207,28 @@ int main( int argc, char *argv[] )
 
   /* Adopt implementation locale */
   set_native_locale();
-  assert_utf8_locale();
+  if ( !is_utf8_locale() ) {
+    /* apply locale-related environment variables from client */
+    clear_locale_variables();
+    for ( list<string>::const_iterator i = locale_vars.begin();
+	  i != locale_vars.end();
+	  i++ ) {
+      char *env_string = strdup( i->c_str() );
+      assert( env_string );
+      if ( 0 != putenv( env_string ) ) {
+	perror( "putenv" );
+      }
+    }
+
+    /* check again */
+    set_native_locale();
+    if ( !is_utf8_locale() ) {
+      fprintf( stderr, "mosh-server needs a UTF-8 native locale to run.\n\n" );
+      fprintf( stderr, "Unfortunately, the locale environment variables currently specify\nthe character set \"%s\".\n\n", locale_charset() );
+      int unused __attribute((unused)) = system( "locale" );
+      exit( 1 );
+    }
+  }
 
   try {
     return run_server( desired_ip, desired_port, command, colors, verbose );

src/frontend/stmclient.cc

@@ -49,7 +49,12 @@
 
 void STMClient::init( void )
 {
-  assert_utf8_locale();
+  if ( !is_utf8_locale() ) {
+    fprintf( stderr, "mosh-client needs a UTF-8 native locale to run.\n\n" );
+    fprintf( stderr, "Unfortunately, the locale environment variables currently specify\nthe character set \"%s\".\n\n", locale_charset() );
+    int unused __attribute((unused)) = system( "locale" );
+    exit( 1 );
+  }
 
   /* Verify terminal configuration */
   if ( tcgetattr( STDIN_FILENO, &saved_termios ) < 0 ) {

src/util/locale_utils.cc

@@ -29,19 +29,50 @@
 
 #include "locale_utils.h"
 
-void assert_utf8_locale( void ) {
+const char *locale_charset( void )
-  /* Verify locale calls for UTF-8 */
+{
-  if ( strcmp( nl_langinfo( CODESET ), "UTF-8" ) != 0 &&
+  static const char ASCII_name[] = "US-ASCII (ANSI_X3.4-1968)";
-       strcmp( nl_langinfo( CODESET ), "utf-8" ) != 0 ) {
+
-    fprintf( stderr, "mosh requires a UTF-8 locale.\n" );
+  /* Produce more pleasant name of US-ASCII */
-    exit( 1 );
+  const char *ret = nl_langinfo( CODESET );
+
+  if ( strcmp( ret, "ANSI_X3.4-1968" ) == 0 ) {
+    ret = ASCII_name;
   }
+
+  return ret;
+}
+
+bool is_utf8_locale( void ) {
+  /* Verify locale calls for UTF-8 */
+  if ( strcmp( locale_charset(), "UTF-8" ) != 0 &&
+       strcmp( locale_charset(), "utf-8" ) != 0 ) {
+    return 0;
+  }
+  return 1;
 }
 
 void set_native_locale( void ) {
   /* Adopt native locale */
   if ( NULL == setlocale( LC_ALL, "" ) ) {
     perror( "setlocale" );
-    exit( 1 );
   }
 }
+
+void clear_locale_variables( void ) {
+  unsetenv( "LANG" );
+  unsetenv( "LANGUAGE" );
+  unsetenv( "LC_CTYPE" );
+  unsetenv( "LC_NUMERIC" );
+  unsetenv( "LC_TIME" );
+  unsetenv( "LC_COLLATE" );
+  unsetenv( "LC_MONETARY" );
+  unsetenv( "LC_MESSAGES" );
+  unsetenv( "LC_PAPER" );
+  unsetenv( "LC_NAME" );
+  unsetenv( "LC_ADDRESS" );
+  unsetenv( "LC_TELEPHONE" );
+  unsetenv( "LC_MEASUREMENT" );
+  unsetenv( "LC_IDENTIFICATION" );
+  unsetenv( "LC_ALL" );
+}

src/util/locale_utils.h

@@ -19,7 +19,9 @@
 #ifndef LOCALE_UTILS_HPP
 #define LOCALE_UTILS_HPP
 
-void assert_utf8_locale( void );
+const char *locale_charset( void );
+bool is_utf8_locale( void );
 void set_native_locale( void );
+void clear_locale_variables( void );
 
 #endif