maria-developers team mailing list archive

[Patrick Galbraith <patg@xxxxxxxx>]

Antony,

Hi there! How goes? I have a question pertaining to bug 571200. I have 
hand-coded the fix from Federated as shown on 
http://lists.mysql.com/commits/102419. Obviously, FederatedX has changed 
enough that there's a fair amount of work I had to do (see my branch at 
bzr+ssh://bazaar.launchpad.net/~capttofu/maria_bug_571200). Most of it I 
think I have a good solution to. The one thing remaining is how to get 
anything that pertains to getting or setting the "result->data_cursor" 
value, as you'll see is in that patch. Since "result" is now a 
FEDERATEDX_IO_RESULT, it doesn't have that structure member, although 
FEDERATEDX_IO_RESULT for federatedx_io_mysql class is a MYSQL_RES, but 
trying to access results->data_cursor results in errors such as:

ha_federatedx.cc: In member function ‘int 
ha_federatedx::read_next(uchar*, FEDERATEDX_IO_RESULT*)’:
ha_federatedx.cc:2888: error: invalid use of incomplete type ‘struct 
st_federatedx_result’
ha_federatedx.h:127: error: forward declaration of ‘struct 
st_federatedx_result’


So, the solution I think is to have a virtual method in the 
federatedx_io class called "get_data_cursor_pos" and 
"set_data_cursor_pos" that within federatedx_io_mysql class obtains 
result->data_cursor. For other driver classes, it'll require some 
thinking, but for now this would get the mysql driver working.

What are your thoughts on this and what I have thus far?

Thanks, hope all is going well!

regards,

Patrick