Andrei Elkin wrote:
> Rafal, hi.
>
> I am fine with the patch.
> I have noticed there are improvements done since July's version.
>
Thanks for review!
> I am okay too that you have not addressed the issue
> of " extending bitmap library with a function that
> does not exist and needed " (my previous review mail) in this patch.
>
Luckily, the issue has dissapeared in the meantime. The relevant code looks now
like this:
+ bitmap_set_all(table->read_set);
+ bitmap_set_all(table->write_set);
+ if (!get_flags(COMPLETE_ROWS_F))
+ bitmap_intersect(table->write_set,&m_cols);
So I'm doing all bitmap operations using existing bitmap functions. This is
possible since I realized that the extra bits on slave (if any) should be set to
0, not to 1. Therefore a simple bitmap intersect does the trick.
Rafal
