Merge lp:~percona-toolkit-dev/percona-toolkit/experimental-fix-1042036-remove-latin1-special-case into lp:percona-toolkit/2.1

[Brian Fraser]

Rebased the branch on top of trunk, so now we have a realistic diff.

[Daniel Nichter (daniel-nichter)]

Can we clean up this diff? There's a lot of conflicts that make it hard to review. Maybe we need to re-create the merge proposal? Or maybe the branch needs to pull in latest trunk.

review: Needs Resubmitting

[Daniel Nichter (daniel-nichter)]

Re our phone conversation this morning: this change would be best for 2.2 rather than 2.1 because it might introduce a (backwards-incompatible) change in behavior, which we can't do right in the middle of a stable version. Plus, the problem could never be reproduced on our end, nor has it been reported by anyone other than the original bug reporter. So we're going to hold off on this in 2.1, but 2.2 or perhaps 3.0 will revamp pt-table-sync chunking.

Unmerged revisions

420. By Brian Fraser <email address hidden> on 2012-10-20

Experimental fix to TableChunker: Don't use the built-in latin1 char range, let mysql calculate it

=== modified file 'lib/TableChunker.pm'
--- lib/TableChunker.pm	2012-08-24 22:50:34 +0000
+++ lib/TableChunker.pm	2012-10-20 17:59:23 +0000
@@ -553,38 +553,13 @@
    my $base;
    my @chars;
    PTDEBUG && _d("Table charset:", $args{tbl_struct}->{charset});
-   if ( ($args{tbl_struct}->{charset} || "") eq "latin1" ) {
-      # These are the unique, sorted latin1 character codes according to
-      # MySQL.  You'll notice that many are missing.  That's because MySQL
-      # treats many characters as the same, for example "e" and "é".
-      my @sorted_latin1_chars = (
-          32,  33,  34,  35,  36,  37,  38,  39,  40,  41,  42,  43,  44,  45,
-          46,  47,  48,  49,  50,  51,  52,  53,  54,  55,  56,  57,  58,  59,
-          60,  61,  62,  63,  64,  65,  66,  67,  68,  69,  70,  71,  72,  73,
-          74,  75,  76,  77,  78,  79,  80,  81,  82,  83,  84,  85,  86,  87,
-          88,  89,  90,  91,  92,  93,  94,  95,  96, 123, 124, 125, 126, 161,
-         162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175,
-         176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189,
-         190, 191, 215, 216, 222, 223, 247, 255);
-
-      my ($first_char, $last_char);
-      for my $i ( 0..$#sorted_latin1_chars ) {
-         $first_char = $i and last if $sorted_latin1_chars[$i] >= $min_col_ord;
-      }
-      for my $i ( $first_char..$#sorted_latin1_chars ) {
-         $last_char = $i and last if $sorted_latin1_chars[$i] >= $max_col_ord;
-      };
-
-      @chars = map { chr $_; } @sorted_latin1_chars[$first_char..$last_char];
-      $base  = scalar @chars;
-   }
-   else {
+   {
       # If the table's charset isn't latin1, who knows what charset is being
       # used, what characters it contains, and how those characters are sorted.
       # So we create a character map and let MySQL tell us these things.
 
       # Create a temp table with the same char col def as the original table.
-      my $tmp_tbl    = '__maatkit_char_chunking_map';
+      my $tmp_tbl    = '__percona_char_chunking_map';
       my $tmp_db_tbl = $q->quote($args{db}, $tmp_tbl);
       $sql = "DROP TABLE IF EXISTS $tmp_db_tbl";
       PTDEBUG && _d($dbh, $sql);