Please consider the following: 1. Allow multiple files with round robin writes. I have have situations in the past where splitting the backup across two+ devices that have space allowed backups to be taken when no single device had enough space. This option may also increase backup speed. 2.

I will implement this in mine Google Summer of Code project.

If you are mentioning bug numbers, please include BUG#26489

This is a feature that we would really like to have. It is troublesome to have to restart the slave when a new database is added to a server that needs to be replicated.

I tend to like the filtering done on the slave, as I may have two slaves that are applying binlogs for different

Dynamic row format has been implemented as a patch for 5.0.45. With the new HP_DATASPACE structure, BLOB should be easy to add. See http://code.google.com/p/mysql-heap-dynamic-rows

I think functional indexes, particularly partial indexes, is of great value. We are a blog service provider and have a database of several terabytes. For many times we found we want to create a partial index over just a few rows, such as an index on only recommended blogs. However, MySQL

Also, "alter table add index" rewrites the entire table, including all other indexes on the table. Can this be avoided as part of this work?

Hm. Why do not I see what other people are saying...

Get this done already. Please! For one thing, it's important to get rid of InnoDB's internal parser. It's bloat and causes inconsistency bugs from time to time.

Just have checked import library stuff. Turns out that if a symbol marked __declspec(dllexport) is present, the linker will create the import library mysqld.lib anyway.Seems like CMake support is not absolutely necessary.