Fix for bug#20648 We introduce a new field method for knowing "real size", and we now in archive null unused bits of a row to null before writing.

sql/field.cc:
  data_length for field needs to be calculated directly.
sql/field.h:
  The new method data_length() returns the "real" length of the field.
sql/ha_archive.cc:
  Before a write_row() archive nulls space beyond the size of the row in varchars to make sure that compression only sees NULL.

1 files changed, 22 insertions, 0 deletions

diff --git a/sql/ha_archive.cc b/sql/ha_archive.cc
index 3885defb4d5..0fb0bc9f791 100644
--- a/sql/ha_archive.cc
+++ b/sql/ha_archive.cc
@@ -710,6 +710,28 @@ int ha_archive::write_row(byte *buf)
     if (init_archive_writer())
       DBUG_RETURN(HA_ERR_CRASHED_ON_USAGE);
 
+  /*
+    Varchar structures are constant in size but are not cleaned up request
+    to request. The following sets all unused space to null to improve
+    compression.
+  */
+  for (Field **field=table->field ; *field ; field++)
+  {
+    DBUG_PRINT("archive",("Pack is %d\n", (*field)->pack_length()));
+    DBUG_PRINT("archive",("MyPack is %d\n", (*field)->data_length((char*) buf + (*field)->offset())));
+    if ((*field)->real_type() == MYSQL_TYPE_VARCHAR) 
+    {
+      uint actual_length= (*field)->data_length((char*) buf + (*field)->offset());
+      uint offset= (*field)->offset() + actual_length + 
+        (actual_length > 255 ? 2 : 1);
+      DBUG_PRINT("archive",("Offset is %d -> %d\n", actual_length, offset));
+      /*
+      if ((*field)->pack_length() + (*field)->offset() != offset)
+        bzero(buf + offset, (size_t)((*field)->pack_length() + (actual_length > 255 ? 2 : 1) - (*field)->data_length));
+    */
+    }
+  }
+
   share->rows_recorded++;
   rc= real_write_row(buf, share->archive_write);
   pthread_mutex_unlock(&share->mutex);