Re: bk commit into 5.1 tree (aelkin:1.2659) BUG#32971

[Mats Kindahl]

Andrei Elkin wrote:
> Mats, hej.
>
> Thanks for your tidy comments!
>
> As you said on #rep the loop problem actually ceased in the latest
> patch.
> Many thanks to you and Zhen Xing for noting it out!
>
> BTW I changed my test to check three errors per event. This is enough
> to check the loop.
> Actually it should something like a rule of thumb: the minimum test of
> a loop should include at least its three passes.
>   

Actually not a bad idea: every sequence should at least handle 
beginning, middle, and end, so it should catch most bad cases. I think 
it would be very uncommon to have something that is dependent on other 
abstractions (e.g., "second from beginning" or "prime number of elements").

Just my few cents,
Mats Kindahl

-- 
Mats Kindahl
Lead Software Developer
Replication Team
MySQL AB, www.mysql.com