Re: bk commit into 5.1 tree (mattiasj:1.2518) BUG#33479

[Sergei Golubchik]

Hi!

On Apr 25, Mattias Jonsson wrote:

>>> First, I just realized that the use of ha_data will eventually
>>> collide for partitioning, when another storage engines start to use
>>> it. So in the future, I believe that partitioning should have an own
>>> ha_partition_data (HA_PARTITION_DATA*) in TABLE_SHARE, since it is
>>> not mutual exclusive using partitioning and an other storage engine
>>> using the new ha_data.
>
> I believe that the table_share is shared by all instances of a table,
> and that must mean that the partition storage engine shares this
> structure with the used partitions storage engines.

It is assumed that every table (TABLE_SHARE) can be only in a one
storage engine. But for partitioning it is not true already. Besides,
technically different partitions could be in different storage engines -
it's not possible now, but there's nothing in the concept of
"partitioning" that prevents it. So creating ha_partition_data won't
help, unless you plan to create ha_innodb_data, ha_myisam_data, etc :)

We'll need to find a different solution for partitioning :(

Regards / Mit vielen Grüssen,
Sergei

-- 
   __  ___     ___ ____  __
  /  |/  /_ __/ __/ __ \/ /   Sergei Golubchik <serg@stripped>
 / /|_/ / // /\ \/ /_/ / /__  Principal Software Developer/Server Architect
/_/  /_/\_, /___/\___\_\___/  MySQL GmbH, Dachauer Str. 37, D-80335 München
       <___/                  Geschäftsführer: Kaj Arnö - HRB
München 162140