From cadf35abc496a470ff6d552570fb876a6efc9a62 Mon Sep 17 00:00:00 2001
From: Sveta Smirnova <sveta.smirnova@percona.com>
Date: Thu, 7 Dec 2023 21:51:47 +0300
Subject: [PATCH] PT-2083 - Pt-archiver --charset option is not working for
 MySQL8.0

- Changed logic, introduced by commit 1d1c13f
- Added test case
---
 bin/pt-archiver                   |  6 +--
 t/pt-archiver/pt-2083.t           | 86 +++++++++++++++++++++++++++++++
 t/pt-archiver/pt-2279.t           |  2 +-
 t/pt-archiver/samples/pt-2083.sql | 18 +++++++
 4 files changed, 108 insertions(+), 4 deletions(-)
 create mode 100644 t/pt-archiver/pt-2083.t
 create mode 100644 t/pt-archiver/samples/pt-2083.sql

diff --git a/bin/pt-archiver b/bin/pt-archiver
index c6992c41..3e14ba7a 100755
--- a/bin/pt-archiver
+++ b/bin/pt-archiver
@@ -6368,11 +6368,11 @@ sub main {
       }
 
       my ($dbh_version) =  $dbh->selectrow_array("SELECT version()");
-      #if ($dbh_version =~ m/^(\d+\.\d+)\.\d+.*/ && $1 ge '8.0' && !$o->get('charset')) {
-      if ($dbh_version =~ m/^(\d+\.\d+)\.\d+.*/ && $1 ge '8.0') {
+      if ( $dbh_version =~ m/^(\d+\.\d+)\.\d+.*/ && $1 ge '8.0' && !$got_charset ) {
          PTDEBUG && _d("MySQL 8.0+ detected and charset was not specified.\n Setting character_set_client = utf8mb4 and --charset=utf8");
          $dbh->do('/*!40101 SET character_set_connection = utf8mb4 */;');
-         $o->set('charset', 'utf8');
+      } elsif ( $got_charset ) {
+         $dbh->do("/*!40101 SET character_set_connection = ${got_charset} */;");
       }
 
       $table->{dbh}  = $dbh;
diff --git a/t/pt-archiver/pt-2083.t b/t/pt-archiver/pt-2083.t
new file mode 100644
index 00000000..3962e1c0
--- /dev/null
+++ b/t/pt-archiver/pt-2083.t
@@ -0,0 +1,86 @@
+#!/usr/bin/env perl
+
+BEGIN {
+   die "The PERCONA_TOOLKIT_BRANCH environment variable is not set.\n"
+      unless $ENV{PERCONA_TOOLKIT_BRANCH} && -d $ENV{PERCONA_TOOLKIT_BRANCH};
+   unshift @INC, "$ENV{PERCONA_TOOLKIT_BRANCH}/lib";
+};
+
+use strict;
+use warnings FATAL => 'all';
+use English qw(-no_match_vars);
+use Test::More;
+
+use charnames ':full';
+
+use PerconaTest;
+use Sandbox;
+require "$trunk/bin/pt-archiver";
+
+my $dp  = new DSNParser(opts=>$dsn_opts);
+my $sb  = new Sandbox(basedir => '/tmp', DSNParser => $dp);
+my $dbh = $sb->get_dbh_for('master');
+
+if ( !$dbh ) {
+   plan skip_all => 'Cannot connect to sandbox master';
+}
+
+my $output;
+my $cnf = "/tmp/12345/my.sandbox.cnf";
+my $cmd = "$trunk/bin/pt-archiver";
+
+$sb->wipe_clean($dbh);
+$sb->create_dbs($dbh, ['test']);
+
+# Test --bulk-insert
+$sb->load_file('master', 't/pt-archiver/samples/pt-2083.sql');
+
+$output = output(
+   sub { pt_archiver::main(qw(--commit-each --where 1=1 --statistics --charset latin1),
+      '--source', "L=1,D=test,t=table_1,F=$cnf",
+      '--dest',   "t=table_1_dest") },
+);
+
+unlike(
+   $output,
+   qr/Character set mismatch/,
+   'No character set mismatch error'
+) or diag($output);
+
+my @copied = $dbh->selectrow_array('SELECT c1 FROM test.table_1_dest');
+
+like(
+   $copied[0],
+   qr/I love MySQL!/,
+   'Rows copied into the table successfully'
+) or diag($copied[0]);
+
+# Test --file
+$sb->load_file('master', 't/pt-archiver/samples/pt-2083.sql');
+
+$output = output(
+   sub { pt_archiver::main(qw(--where 1=1 --statistics --charset latin1),
+      '--source', "L=1,D=test,t=table_1,F=$cnf",
+      '--file',   '/tmp/%Y-%m-%d-%D_%H:%i:%s.%t') },
+);
+
+unlike(
+   $output,
+   qr/Character set mismatch/,
+   'No character set mismatch error'
+) or diag($output);
+
+like(
+   `cat /tmp/*.table_1`,
+   qr/I love MySQL!/,
+   'Rows copied into the file successfully'
+) or diag($output);
+
+# #############################################################################
+# Done.
+# #############################################################################
+diag(`rm -f /tmp/*.table_1`);
+$sb->wipe_clean($dbh);
+ok($sb->ok(), "Sandbox servers") or BAIL_OUT(__FILE__ . " broke the sandbox");
+done_testing;
+exit;
diff --git a/t/pt-archiver/pt-2279.t b/t/pt-archiver/pt-2279.t
index 11c9620d..fc4022dc 100644
--- a/t/pt-archiver/pt-2279.t
+++ b/t/pt-archiver/pt-2279.t
@@ -67,7 +67,7 @@ unlike(
 unlike(
    $output,
    qr/Cannot open :encoding(utf8mb4)/,
-   'Destination file can be read if encodin utf8mb4 is used'
+   'Destination file can be read if encoding utf8mb4 is used'
 ) or diag($output);
 
 # #############################################################################
diff --git a/t/pt-archiver/samples/pt-2083.sql b/t/pt-archiver/samples/pt-2083.sql
new file mode 100644
index 00000000..2aa39a50
--- /dev/null
+++ b/t/pt-archiver/samples/pt-2083.sql
@@ -0,0 +1,18 @@
+DROP TABLE IF EXISTS test.table_1;
+DROP TABLE IF EXISTS test.table_1_dest;
+
+SET NAMES latin1;
+
+CREATE TABLE test.table_1 (
+    id int(11) NOT NULL,
+    c1 varchar(20) DEFAULT NULL,
+    PRIMARY KEY (id)
+) ENGINE=InnoDB DEFAULT CHARSET=latin1;
+
+INSERT INTO test.table_1 VALUES(1, 'I love MySQL!');
+
+CREATE TABLE test.table_1_dest (
+    id int(11) NOT NULL,
+    c1 varchar(20) DEFAULT NULL,
+    PRIMARY KEY (id)
+) ENGINE=InnoDB DEFAULT CHARSET=latin1;