Removed the charset munging and made a note in the pt-archiver docs that the only legal charsets are those known by MySQL

bin/pt-archiver

@@ -2121,12 +2121,6 @@ sub fill_in_dsn {
    $dsn->{D} ||= $db;
 }
 
-my %encoding_aliases = (
-   'utf-8'      => 'utf8',
-   'iso-8859-1' => 'latin1',
-   'latin-1'    => 'latin1',
-);
-
 sub get_dbh {
    my ( $self, $cxn_string, $user, $pass, $opts ) = @_;
    $opts ||= {};
@@ -2135,7 +2129,7 @@ sub get_dbh {
       RaiseError         => 1,
       PrintError         => 0,
       ShowErrorStatement => 1,
-      mysql_enable_utf8 => ($cxn_string =~ m/charset=utf-?8/i ? 1 : 0),
+      mysql_enable_utf8 => ($cxn_string =~ m/charset=utf8/i ? 1 : 0),
    };
    @{$defaults}{ keys %$opts } = values %$opts;
 
@@ -2203,8 +2197,7 @@ sub get_dbh {
            . ": $EVAL_ERROR";
       }
 
-      if ( my ($charset) = $cxn_string =~ m/charset=([-\w]+)/ ) {
+      if ( my ($charset) = $cxn_string =~ m/charset=([\w]+)/ ) {
-         $charset = $encoding_aliases{lc($charset)} || $charset;
          $sql = qq{/*!40101 SET NAMES "$charset"*/};
          PTDEBUG && _d($dbh, ':', $sql);
          eval { $dbh->do($sql) };
@@ -5116,6 +5109,9 @@ STDOUT to utf8, passes the mysql_enable_utf8 option to DBD::mysql, and runs SET
 NAMES UTF8 after connecting to MySQL.  Any other value sets binmode on STDOUT
 without the utf8 layer, and runs SET NAMES after connecting to MySQL.
 
+Note that only charsets as known by MySQL are recognized; So for example,
+"UTF8" will work, but "UTF-8" will not.
+
 See also L<"--[no]check-charset">.
 
 =item --[no]check-charset

lib/DSNParser.pm

@@ -263,18 +263,6 @@ sub fill_in_dsn {
    $dsn->{D} ||= $db;
 }
 
-# MySQL won't resolve iso-8859-1 or latin-1 as latin1, while Perl would, so
-# we hardcode the aliases here. The UTF-8 case is a bit different;
-# MySQL doesn't really support UTF-8 in SET NAMES, instead using
-# their own definition, which is constrained to codepoints 0..0xFFFF, so
-# rightfully calls it something different: utf8. I'm not actually sure
-# if the naming convention is intended or plain lucky on their part, though.
-my %encoding_aliases = (
-   'utf-8'      => 'utf8',
-   'iso-8859-1' => 'latin1',
-   'latin-1'    => 'latin1',
-);
-
 # Actually opens a connection, then sets some things on the connection so it is
 # the way the Maatkit tools will expect.  Tools should NEVER open their own
 # connection or use $dbh->reconnect, or these things will not take place!
@@ -286,7 +274,7 @@ sub get_dbh {
       RaiseError         => 1,
       PrintError         => 0,
       ShowErrorStatement => 1,
-      mysql_enable_utf8 => ($cxn_string =~ m/charset=utf-?8/i ? 1 : 0),
+      mysql_enable_utf8 => ($cxn_string =~ m/charset=utf8/i ? 1 : 0),
    };
    @{$defaults}{ keys %$opts } = values %$opts;
 
@@ -364,8 +352,7 @@ sub get_dbh {
       }
 
       # Set character set and binmode on STDOUT.
-      if ( my ($charset) = $cxn_string =~ m/charset=([-\w]+)/ ) {
+      if ( my ($charset) = $cxn_string =~ m/charset=([\w]+)/ ) {
-         $charset = $encoding_aliases{lc($charset)} || $charset;
          $sql = qq{/*!40101 SET NAMES "$charset"*/};
          PTDEBUG && _d($dbh, ':', $sql);
          eval { $dbh->do($sql) };

t/lib/DSNParser.t

@@ -554,7 +554,7 @@ my ($out, undef) = full_output(sub { $dp->get_dbh(@opts, {}) });
 like(
    $out,
    qr/\QUnknown character set/,
-   "get_dbh dies withg an unknown charset"
+   "get_dbh dies with an unknown charset"
 );
 
 $dp->prop('set-vars', "time_zoen='UTC'");
@@ -563,7 +563,7 @@ $dp->prop('set-vars', "time_zoen='UTC'");
 like(
    $out,
    qr/\QUnknown system variable 'time_zoen'/,
-   "get_dbh dies withg an unknown charset"
+   "get_dbh dies with an unknown system variable"
 );
 
 # #############################################################################

t/pt-archiver/file.t

@@ -22,9 +22,6 @@ my $dbh = $sb->get_dbh_for('master');
 if ( !$dbh ) {
    plan skip_all => 'Cannot connect to sandbox master';
 }
-else {
-   plan tests => 13;
-}
 
 my $output;
 my $rows;
@@ -95,7 +92,7 @@ sub test_charset {
    ) or diag($out);
 }
 
-for my $charset (qw(latin1 iso-8859-1 latin-1 utf8 UTF-8 UTF8 )) {
+for my $charset (qw(latin1 utf8 UTF8 )) {
    test_charset($charset);
 }
 
@@ -112,11 +109,16 @@ my ($out) = full_output( sub {
    },
 );
 
-like($out, qr/\QError setting NAMES to some_charset_that_doesn/, "..but an unknown charset fails");
+like(
+   $out,
+   qr/\QError setting NAMES to some_charset_that_doesn/,
+   "..but an unknown charset fails"
+);
 
 # #############################################################################
 # Done.
 # #############################################################################
 $sb->wipe_clean($dbh);
 ok($sb->ok(), "Sandbox servers") or BAIL_OUT(__FILE__ . " broke the sandbox");
-exit;
+
+done_testing;